v6.2.4 release notes

This feature works for gs >= 9.24, but we don't want to change behavior in
a backport, so we're disabling it unconditionally.

.coveragerc

@@ -0,0 +1,16 @@
+# Coverage isn't really compatible with subprocesses so results are unreliable
+
+[run]
+branch = True
+#concurrency = multiprocessing
+source = ocrmypdf/
+
+[report]
+exclude_lines =
+	pragma: no cover
+	def __repr__
+	raise AssertionError
+	raise NotImplementedError
+	if 0:
+	if False:
+	if __name__ == .__main__.:

.docker/Dockerfile

@@ -0,0 +1,50 @@
+# OCRmyPDF
+#
+FROM      ubuntu:18.04
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  ocrmypdf \
+  unpaper \
+  tesseract-ocr-eng \
+  tesseract-ocr-fra \
+  tesseract-ocr-deu \
+  tesseract-ocr-spa \
+  tesseract-ocr-por \
+  tesseract-ocr-chi-sim \
+  python3-venv \
+  python3-pip
+
+ENV LANG=C.UTF-8
+
+RUN python3 -m venv --system-site-packages /appenv
+
+# This installs the latest binary wheel instead of the code in the current
+# folder. Installing from source will fail, apparently because cffi needs
+# build-essentials (gcc) to do a source installation 
+# (i.e. "pip install ."). It's unclear to me why this is the case.
+RUN . /appenv/bin/activate; \
+  pip install --upgrade pip \
+  && pip install ocrmypdf[fitz]
+
+# Now copy the application in, mainly to get the test suite.
+# Do this now to make the best use of Docker cache.
+COPY . /application
+RUN . /appenv/bin/activate; \
+  pip install -r /application/test_requirements.txt
+
+# Remove the junk, including the source version of application since it was
+# already installed
+RUN rm -rf /tmp/* /var/tmp/* /root/* /application/ocrmypdf \
+  && apt-get autoremove -y \
+  && apt-get autoclean -y
+
+RUN useradd docker \
+  && mkdir /home/docker \
+  && chown docker:docker /home/docker
+
+USER docker
+WORKDIR /home/docker
+
+# Must use array form of ENTRYPOINT
+# Non-array form does not append other arguments, because that is "intuitive"
+ENTRYPOINT ["/application/.docker/docker-wrapper.sh"]

.docker/Dockerfile.polyglot

@@ -0,0 +1,17 @@
+# OCRmyPDF polyglot
+#
+FROM      jbarlow83/ocrmypdf:latest
+
+USER root
+
+# Update system and install our dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  tesseract-ocr-all 
+
+RUN apt-get autoremove -y && apt-get clean -y
+
+USER docker
+
+# Must use array form of ENTRYPOINT
+# Non-array form does not append other arguments, because that is "intuitive"
+ENTRYPOINT ["/application/.docker/docker-wrapper.sh"]

.dockerignore

@@ -1,21 +1,25 @@
+*.ipynb
+*.pdf
+*.pyc
+*.rst
+*.sublime*
+**/*.pyc
+.*/
+!.docker/
+.ruffus_history.sqlite
 bin/
 build/
+docs/
 dist/
+htmlcov/
 include/
 lib/
+MANIFEST.in
 ocrmypdf.egg-info/
 staging/
-.git/
-.ruffus_history.sqlite
-MANIFEST.in
-*.sublime*
-*.pdf
-*.rst
-*.pyc
-*/*.pyc
-*/*/*.pyc
-*/*/*/*.pyc
-*/*/*/*/*.pyc
-*/*/*/*/*/*.pyc
-*/*/*/*/*/*/*.pyc
-*/*/*/*/*/*/*/*.pyc
+tests/cache/
+tests/output/
+tests/resources/private/
+tmp/
+venv*/
+wheelhouse/

.git_archival.txt

@@ -0,0 +1 @@
+ref-names: $Format:%D$

.gitattributes

@@ -5,4 +5,9 @@
 # (binary is a macro for -text -diff)
 *.jar	binary
 *.pdf	binary
-*.PDF	binary
+*.PDF	binary
+*.png	binary
+*.jpg	binary
+*.bin   binary
+
+.git_archival.txt  export-subst

.github/ISSUE_TEMPLATE

@@ -0,0 +1,5 @@
+Please include the command line and a test file with your issue report.
+
+If possible, please use a test file that we can include in future test cases (no personal information, no copyrighted material).
+
+If you wish to encrypt a test file for the OCRmyPDF maintainer only, see the [Wiki](https://github.com/jbarlow83/OCRmyPDF/wiki).

.gitignore

@@ -1,17 +1,47 @@
-tmp/
-log/
+# Development environment
 *.pyc
-tests/output/
-.ruffus_history.sqlite
 *.sublime-*
-/*.pdf
+venv*/
+.venv/
+pyvenv.cfg
+tasks.py
+.bash_history
+.ruffus_history.sqlite
+.idea/
+.pytest_cache/
+.pylintrc
+
+# Package building
+*.egg-info/
+.cache/
+.eggs/
 build/
 dist/
-*.egg-info/
-venv/
-*/test/output
-bin/
-include/
-lib/
-pip-selfcheck.json
-pyvenv.cfg
+wheelhouse/
+
+# Automatically generated files
+ocrmypdf/lib/_*.py
+ocrmypdf/version.py
+docs/_build/
+docs/_static/
+docs/_templates/
+docs/Makefile
+
+# Code coverage
+.coverage
+htmlcov/
+
+# Testing
+log/
+/*.pdf
+/*.qdf
+*.ipynb
+.ipynb_checkpoints/
+tests/output/
+tests/resources/private/
+tmp/
+pdfbox-app*.jar
+.vscode/
+IDEAS
+_Dockerfile.local
+/scratch.py

.travis.yml

@@ -0,0 +1,98 @@
+dist: trusty
+language: python
+cache:
+  pip: true
+  directories:
+  - $HOME/Library/Caches/Homebrew
+
+addons:
+  apt:
+    update: true
+    sources:
+    - sourceline: 'ppa:alex-p/tesseract-ocr'
+    - sourceline: 'ppa:heyarje/libav-11'
+    - sourceline: 'ppa:vshn/ghostscript'
+    packages:
+    - ghostscript
+    - libavcodec56
+    - libavformat56
+    - libavutil54
+    - libffi-dev
+    - poppler-utils
+    - qpdf
+    - tesseract-ocr
+    - tesseract-ocr-deu
+    - tesseract-ocr-eng
+    - tesseract-ocr-fra
+
+matrix:
+  include:
+    - os: linux
+      sudo: required
+      language: python
+      python: "3.5"
+      env: EXTRAS=
+    - os: linux
+      sudo: required
+      language: python
+      python: "3.6"
+      env: EXTRAS=
+    - os: linux
+      sudo: required
+      language: python
+      python: "3.6"
+      env: EXTRAS=[fitz]
+    - os: linux
+      sudo: required
+      language: python
+      python: "3.7-dev"
+    - os: osx
+      osx_image: xcode8
+      language: generic
+      env: EXTRAS=
+    - os: osx
+      osx_image: xcode8
+      language: generic
+      env: EXTRAS=[fitz]
+
+before_cache:
+- rm -f $HOME/.cache/pip/log/debug.log
+
+before_install: |
+  if [[ "$TRAVIS_OS_NAME" == "linux" ]]; then
+    pip install --upgrade pip
+    mkdir -p packages
+    wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O packages/unpaper_6.1-1.deb
+    sudo dpkg -i packages/unpaper_6.1-1.deb
+  elif [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
+    brew update && brew bundle --file=.travis/Brewfile
+    pip3 install --upgrade pip
+    pip3 install wheel
+  fi
+
+install:
+- pip3 install pycparser # py3.7 workaround for https://github.com/eliben/pycparser/issues/251
+- pip3 install ".$EXTRAS"
+- pip3 install -r test_requirements.txt
+
+script:
+- tesseract --version
+- qpdf --version
+- pytest -n auto
+
+deploy:
+  # release for main pypi
+  # 3.6 w/o fitz is considered the build leader and does the deploy, otherwise there is
+  # a race and all versions will try to deploy
+  # OTOH if we ever need separate binary wheels then each version needs its
+  # own deploy
+- provider: pypi
+  user: ocrmypdf-travis
+  password:
+    secure: "DTFOmmNL6olA0+yXvp4u9jXZlZeqrJsJ0526jzqf4a3gZ6jnGTq5UI6WzRsslSyoMMfXKtHQebqHM6ogSgCZinyZ3ufHJo8fn9brxbEc2gsiWkbj5o3bGwdWMT1vNNE7XW0VCpw87rZ1EEwjl4FJHFudMlPR1yfU5+uq0k0PACo="
+  distributions: "sdist bdist_wheel"
+  on:
+    branch: master
+    tags: true
+    condition: $TRAVIS_PYTHON_VERSION == "3.6" && $TRAVIS_OS_NAME == "linux" && $EXTRAS == ""
+  skip_upload_docs: true

.travis/Brewfile

@@ -0,0 +1,13 @@
+brew 'python'
+
+brew 'ghostscript'
+brew 'jbig2dec'
+brew 'leptonica'
+brew 'libffi'
+brew 'libtiff'
+brew 'libxml2'
+brew 'mupdf-tools'
+brew 'openjpeg'
+brew 'qpdf'
+brew 'tesseract'
+brew 'unpaper'

Dockerfile

@@ -1,87 +0,0 @@
-# OCRmyPDF
-#
-# VERSION               3.0.0
-FROM      debian:stretch
-MAINTAINER James R. Barlow <jim@purplerock.ca>
-
-# Add unprivileged user
-RUN useradd docker \
-  && mkdir /home/docker \
-  && chown docker:docker /home/docker
-
-# Update system and install our dependencies
-RUN apt-get update && apt-get install -y --no-install-recommends \
-  locales \
-  ghostscript \
-  tesseract-ocr \
-  tesseract-ocr-deu tesseract-ocr-spa tesseract-ocr-eng tesseract-ocr-fra \
-  qpdf \
-  poppler-utils \
-  python3 \
-  python3-pip \
-  python3-venv \
-  python3-reportlab \
-  python3-pil
-
-# Enforce UTF-8
-# Borrowed from https://index.docker.io/u/crosbymichael/python/ 
-RUN dpkg-reconfigure locales && \
-  locale-gen C.UTF-8 && \
-  /usr/sbin/update-locale LANG=C.UTF-8
-ENV LC_ALL C.UTF-8
-
-# Build unpaper 6.1
-RUN apt-get install -y \
-  wget \
-  gcc \
-  libavformat-dev \
-  libavcodec-dev \
-  libavutil-dev \
-  autoconf \
-  automake \
-  make \
-  pkg-config \
-  xsltproc
-
-WORKDIR /root
-RUN wget https://github.com/Flameeyes/unpaper/archive/unpaper-6.1.tar.gz
-RUN tar xf unpaper-6.1.tar.gz
-WORKDIR /root/unpaper-unpaper-6.1
-RUN autoreconf -i
-RUN ./configure CFLAGS="-O2 -march=native -pipe -flto"
-RUN make -j install
-
-RUN apt-get remove -y \
-  gcc \
-  autoconf \
-  automake \
-  pkg-config \
-  xsltproc \
-  make
-RUN apt-get autoremove -y && apt-get clean -y
-RUN rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
-
-# Set up a Python virtualenv and take all of the system packages, so we can
-# rely on the platform packages rather than importing GCC and compiling them
-RUN pyvenv /appenv \
-  && pyvenv --system-site-packages /appenv
-
-COPY . /application/
-
-# Install application and dependencies
-# In this arrangement Pillow and reportlab will be provided by the system
-RUN . /appenv/bin/activate; \
-  pip install --upgrade pip \
-  && pip install --no-cache-dir /application \
-  && pip install --no-cache-dir -r /application/test_requirements.txt
-
-USER docker
-WORKDIR /home/docker
-
-ENV DEFAULT_RUFFUS_HISTORY_FILE=/tmp/.{basename}.ruffus_history.sqlite
-ENV OCRMYPDF_TEST_OUTPUT=/tmp/test-output
-ENV OCRMYPDF_IN_DOCKER=1
-
-# Must use array form of ENTRYPOINT
-# Non-array form does not append other arguments, because that is "intuitive"
-ENTRYPOINT ["/application/docker-wrapper.sh"]

LICENSE

@@ -0,0 +1,674 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU General Public License is a free, copyleft license for
+software and other kinds of works.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+the GNU General Public License is intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.  We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors.  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights.  Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received.  You must make sure that they, too, receive
+or can get the source code.  And you must show them these terms so they
+know their rights.
+
+  Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+  For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software.  For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+  Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so.  This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software.  The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable.  Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products.  If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+  Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary.  To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Use with the GNU Affero General Public License.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU Affero General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+    <program>  Copyright (C) <year>  <name of author>
+    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU GPL, see
+<https://www.gnu.org/licenses/>.
+
+  The GNU General Public License does not permit incorporating your program
+into proprietary programs.  If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.  But first, please read
+<https://www.gnu.org/licenses/why-not-lgpl.html>.

LICENSE.rst

@@ -1,20 +0,0 @@
-Copyright (c) 2013-2015, The OCRmyPDF Authors
-
-Permission is hereby granted, free of charge, to any person obtaining a
-copy of this software and associated documentation files (the
-"Software"), to deal in the Software without restriction, including
-without limitation the rights to use, copy, modify, merge, publish,
-distribute, sublicense, and/or sell copies of the Software, and to
-permit persons to whom the Software is furnished to do so, subject to
-the following conditions:
-
-The above copyright notice and this permission notice shall be included
-in all copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
-OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
-IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
-CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
-TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
-SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

MANIFEST.in

@@ -1,3 +1,49 @@
-recursive-exclude tests/output *
+# requirements
 include requirements.txt
-include test_requirements.txt
+include test_requirements.txt
+include dev_requirements.txt
+
+# git
+include .git_archival.txt
+
+# docker
+include .dockerignore
+recursive-include .docker *
+
+# tests
+include .coveragerc
+recursive-include tests *.jpg
+recursive-include tests *.png
+recursive-include tests *.pdf
+recursive-include tests *.py
+recursive-include tests *.rst
+recursive-include tests *.txt
+recursive-include tests/cache *
+recursive-exclude tests/output *
+recursive-exclude tests/output_pageinfo *
+recursive-exclude tests/resources/private *
+
+# documentation
+include LICENSE
+include *.rst
+recursive-exclude .github * 
+recursive-include docs *.py
+recursive-include docs *.rst
+recursive-include docs *.svg
+recursive-exclude docs/_build *
+
+
+# support files
+recursive-include src/ocrmypdf/data *
+include *.py
+exclude tasks.py
+recursive-exclude .travis *
+exclude .travis*
+
+
+# code
+recursive-include src/ocrmypdf *.py
+exclude ocrmypdf/lib/_leptonica.py
+
+
+

OCRmyPDF.sh

@@ -1,6 +0,0 @@
-#!/bin/sh
-##############################################################################
-# Copyright (c) 2013-14: fritz-hh from Github (https://github.com/fritz-hh)
-##############################################################################
-
-python3 -m ocrmypdf.main "$@"

README.rst

@@ -1,32 +1,52 @@
 OCRmyPDF
 ========
 
+.. image:: https://travis-ci.org/jbarlow83/OCRmyPDF.svg?branch=master
+    :target: https://travis-ci.org/jbarlow83/OCRmyPDF
+
+.. image:: https://img.shields.io/pypi/v/ocrmypdf.svg
+    :target: https://pypi.org/project/ocrmypdf/
+
+.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg
+    :alt: homebrew
+    :target: http://brewformulas.org/Ocrmypdf
+
+
 OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to
-be searched.
+be searched or copy-pasted.
+
+.. code-block:: bash
+
+   ocrmypdf                      # it's a scriptable command line program
+      -l eng+fra                 # it supports multiple languages
+      --rotate-pages             # it can fix pages that are misrotated
+      --deskew                   # it can deskew crooked PDFs!
+      --title "My PDF"           # it can change output metadata
+      --jobs 4                   # it uses multiple cores by default
+      --output-type pdfa         # it produces PDF/A by default
+      input_scanned.pdf          # takes PDF input (or images)
+      output_searchable.pdf      # produces validated PDF output
+
 
 Main features
 -------------
 
 -  Generates a searchable
-   `PDF/A <https://en.wikipedia.org/?title=PDF/A>`__ file from a regular PDF
-   only containing images
--  Places OCRed text accurately below the image to ease copy / paste
+   `PDF/A <https://en.wikipedia.org/?title=PDF/A>`_ file from a regular PDF
+-  Places OCR text accurately below the image to ease copy / paste
 -  Keeps the exact resolution of the original embedded images
-
-   -  or if requested oversamples the images before OCRing so as to get
-      better results
-
--  When possible, copies input images directly to output without transcoding them,
-   to preserve image quality
+-  When possible, inserts OCR information as a "lossless" operation without rendering vector information
 -  Keeps file size about the same
 -  If requested deskews and/or cleans the image before performing OCR
 -  Validates input and output files
 -  Provides debug mode to enable easy verification of the OCR results
--  Processes several pages in parallel when more than one CPU core is
+-  Processes pages in parallel when more than one CPU core is
    available
--  Uses Tesseract OCR engine
+-  Uses `Tesseract OCR <https://github.com/tesseract-ocr/tesseract>`_ engine
+-  Supports more than `100 languages <https://github.com/tesseract-ocr/tessdata>`_ recognized by Tesseract
+-  Battle-tested on thousands of PDFs, a test suite and continuous integration
 
-For details: please consult the `release notes <RELEASE_NOTES.rst>`__
+For details: please consult the `documentation <https://ocrmypdf.readthedocs.io/en/latest/>`_.
 
 Motivation
 ----------
@@ -34,209 +54,97 @@ Motivation
 I searched the web for a free command line tool to OCR PDF files on
 Linux/UNIX: I found many, but none of them were really satisfying.
 
--  Either they produced PDF files with misplaced text under the image (making copy/paste impossible) 
--  Or they did not display correctly some escaped HTML characters located in the hOCR file produced by the OCR engine 
+-  Either they produced PDF files with misplaced text under the image (making copy/paste impossible)
+-  Or they did not handle accents and multilingual characters
 -  Or they changed the resolution of the embedded images
--  Or they generated PDF files having a ridiculous big size
+-  Or they generated ridiculously large PDF files
 -  Or they crashed when trying to OCR some of my PDF files
 -  Or they did not produce valid PDF files (even though they were readable with my current PDF reader)
 -  On top of that none of them produced PDF/A files (format dedicated for long time storage)
 
-... so I decided to develop my own tool (using various existing scripts
-as an inspiration)
+...so I decided to develop my own tool (using various existing scripts
+as an inspiration).
 
 Installation
 ------------
 
-Download OCRmyPDF here: https://github.com/jbarlow83/OCRmyPDF/releases
+Linux, UNIX, and macOS are supported. Windows is not directly supported but there is a Docker image available that runs on Windows.
 
-You can install it to a Python virtual environment or system-wide. 
+Users of Debian 9 or later or Ubuntu 16.10 or later may simply
 
-Installing the Docker container
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: bash
 
-For many users, installing the Docker container will be easier than installing all of OCRmyPDF's dependencies. For Windows, it is the only option.
+   apt-get install ocrmypdf
 
-If you have `Docker <https://docs.docker.com/>`__ installed on your system, you can install
-a Docker container of the latest release.
+and macOS users with Homebrew may simply
 
-Follow the Docker installation instructions for your platform.  If you can run this command
-successfully, your system is ready to download and execute the image::
+.. code-block:: bash
 
-   docker run hello-world
-   
-OCRmyPDF will use all available CPU cores.  By default, the VirtualBox machine instance on Windows and OS X has only a single CPU core enabled. Use the VirtualBox Manager to determine the name of your Docker container host, and then follow these optional steps to enable multiple CPUs::
+   brew install ocrmypdf
 
-   # Optional
-   docker-machine stop "yourVM"
-   VBoxManage modifyvm "yourVM" --cpus 2  # or whatever number of core is desired
-   docker-machine start "yourVM"
-   eval $(docker-machine env "yourVM")
+For everyone else, `see our documentation <https://ocrmypdf.readthedocs.io/en/latest/installation.html>`_ for installation steps.
 
-Assuming you have a Docker engine running somewhere, you can run these commands to download
-the image::
+Languages
+---------
 
-   docker pull jbarlow83/ocrmypdf
+OCRmyPDF uses Tesseract for OCR, and relies on its language packs. For Linux users,
+you can often find packages that provide language packs:
 
-Then tag it to give a more convenient name, just ocrmypdf::
+.. code-block:: bash
 
-   docker tag jbarlow83/ocrmypdf ocrmypdf
-  
-You can then run using the command::
+   # Display a list of all Tesseract language packs
+   apt-cache search tesseract-ocr
 
-   docker run ocrmypdf --help
-  
-To execute the OCRmyPDF on a local file, you must `provide a writable volume to the Docker image <https://docs.docker.com/userguide/dockervolumes/>`__, such as this in this template::
+   # Debian/Ubuntu users
+   apt-get install tesseract-ocr-chi-sim  # Example: Install Chinese Simplified language back
 
-   docker run -v "$(pwd):/home/docker" <other docker arguments>   ocrmypdf <your arguments to ocrmypdf>
+You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple
+languages can be requested.
 
-In this worked example, the current working directory contains an input file called `test.pdf` and the output will go to `output.pdf`:: 
+Documentation and support
+-------------------------
 
-   docker run -v "$(pwd):/home/docker"   ocrmypdf --skip-text test.pdf output.pdf
+Once ocrmypdf is installed, the built-in help which explains the command syntax and options can be accessed via:
 
-Note that `ocrmypdf` has its own separate -v argument to control debug verbosity. All Docker arguments should before the `ocrmypdf` container name and all arguments to `ocrmypdf` should be listed after.
-
-Installing on Mac OS X Yosemite
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-If it's not already present, `install Homebrew <http://brew.sh/>`__
-
-Update Homebrew::
-
-   brew update
-   
-Install or upgrade the required Homebrew packages, if any are missing::
-
-   brew install libpng openjpeg jbig2dec     # image libraries
-   brew install qpdf
-   brew install ghostscript
-   brew install python3
-   brew install libxml2
-   brew install leptonica
-   brew install tesseract
-   
-It is also recommended that install Pillow and confirm it can read and write JPEG and PNG files::
-
-   pip3 install --upgrade pip
-   pip3 install --upgrade pillow
-
-To test that your Python imaging library (Pillow) can access JPEG and PNG files, try this command::
-
-   python3 -c "from PIL import Image; im = Image.new('1', (1, 1)); im.save('test.png'); im.save('test.jpg')"
-
-If you have trouble getting Pillow to access JPEG and PNG files, `review the installation instructions <https://pillow.readthedocs.org/installation.html>`__.
-
-You can then install OCRmyPDF from PyPI::
-
-   pip3 install ocrmypdf
-
-The command line program should now be available::
+.. code-block:: bash
 
    ocrmypdf --help
 
-Installing on Ubuntu 14.04 LTS
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Our `documentation is served on Read the Docs <https://ocrmypdf.readthedocs.io/en/latest/index.html>`_.
 
-Installing on Ubuntu 14.04 LTS (trusty) is more difficult than other options, because of certain bugs in package installation.
+If you detect an issue, please:
 
-Update apt-get::
-
-   sudo apt-get update
-   sudo apt-get upgrade
-   
-Install system dependencies::
-
-   sudo apt-get install \
-      zlib1g-dev \
-      libjpeg-dev \
-      ghostscript \
-      tesseract-ocr \
-      qpdf \
-      unpaper \
-      python3-pip \
-      python3-pil \
-      python3-pytest \
-      python3-reportlab
-
-If you wish install OCRmyPDF to the system Python, then install as follows (note this installs new packages
-into your system Python, which could interfere with other programs)::
-
-   sudo pip3 install ocrmypdf
-   
-If you wish to install OCRmyPDF to a virtual environment to isolate system Python from modified, you can
-follow these steps.  This includes a workaround `for a known, unresolved issue in Ubuntu 14.04's ensurepip
-package <http://www.thefourtheye.in/2014/12/Python-venv-problem-with-ensurepip-in-Ubuntu.html>`__::
-
-   sudo apt-get install python3-venv
-   python3 -m venv venv-ocrmypdf --without-pip
-   source venv-ocrmypdf/bin/activate
-   wget -O - -o /dev/null https://bootstrap.pypa.io/get-pip.py | python
-   deactivate
-   pyvenv --system-site-packages venv-ocrmypdf
-   source venv-ocrmypdf/bin/activate
-   pip install ocrmypdf
-
-Ubuntu 14.04 only installs `unpaper` version 0.4.2, which is not supported by OCRmyPDF because it is produces invalid output. This program is an optional dependency, and provides page deskewing and cleaning. See `Dockerfile <Dockerfile>`__ for an example of how to building unpaper 6.1 from source. If you choose to install unpaper later, OCRmyPDF will use the foremost version on the system PATH.
-
-      
-Installing HEAD revision from sources
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-To install the HEAD revision from sources in development mode::
-
-   git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
-   cd OCRmyPDF
-   pip3 install -e .
-   
-On certain Linux/UNIX platforms such as Ubuntu, you may need to use 
-run the install command as superuser::
-
-   sudo pip3 install -e .
-   
-Note that this will alter your system's Python distribution. If you prefer 
-to not install as superuser, you can install the package in a Python virtual environment::
-
-   git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
-   pyvenv venv
-   source venv/bin/activate
-   cd OCRmyPDF
-   pip3 install -e .
-
-If your platform does not have ``pip3``, make sure that Python 3.4+ and the `pip` 
-package are installed.
-
-To run the program::
-   
-   ocrmypdf --help
-
-If not yet installed, the script will notify you about dependencies that
-need to be installed. The script requires specific versions of the
-dependencies. Older version than the ones mentioned in the release notes
-are likely not to be compatible to OCRmyPDF.
-
-Support
--------
-
-In case you detect an issue, please:
-
--  Check if your issue is already known
+-  Check whether your issue is already known
 -  If no problem report exists on github, please create one here:
-   https://github.com/fritz-hh/OCRmyPDF/issues
+   https://github.com/jbarlow83/OCRmyPDF/issues
 -  Describe your problem thoroughly
 -  Append the console output of the script when running the debug mode
-   (-v 1 option)
+   (``-v 1`` option)
 -  If possible provide your input PDF file as well as the content of the
-   temporary folder (using a file sharing service like
-   www.file-upload.net)
+   temporary folder (using a file sharing service like Dropbox)
+
+Requirements
+------------
+
+Runs on CPython 3.5, 3.6 and 3.7. Requires external program installations of Ghostscript, Tesseract OCR, QPDF, and Leptonica. ocrmypdf is pure Python, but uses CFFI to portably generate library bindings.
 
 Press & Media
 -------------
 
--  `c't 1-2014, page 59 <http://www.heise.de/ct/inhalt/2014/1/58/>`__:
+-  `c't 1-2014, page 59 <http://heise.de/-2279695>`_:
    Detailed presentation of OCRmyPDF v1.0 in the leading German IT
    magazine c't
 -  `heise Open Source, 09/2014: Texterkennung mit
-   OCRmyPDF <http://www.heise.de/-2356670>`__
+   OCRmyPDF <http://heise.de/-2356670>`_
+
+License
+-------
+
+The OCRmyPDF software is licensed under the GNU GPLv3. Certain files are covered by other licenses, as noted in their source files.
+
+The license for each test file varies, and is noted in tests/resources/README.rst. The documentation is licensed under Creative Commons Attribution-ShareAlike 4.0 (CC-BY-SA 4.0).
+
+OCRmyPDF versions prior to 6.0 were licensed under the MIT License.
 
 Disclaimer
 ----------

RELEASE_NOTES.rst

@@ -1,518 +0,0 @@
-RELEASE NOTES
-=============
-
-Please always read this file before installing the package
-
-Download software here: https://github.com/jbarlow83/OCRmyPDF/tags
-
-v3.0:
-=====
-
-New features
-------------
-
--  Easier installation with a Docker container or Python's ``pip`` package manager 
--  Eliminated many external dependencies, so it's easier to setup
--  Now installs ``ocrmypdf`` to ``/usr/local/bin`` or equivalent for system-wide
-   access and easier typing
--  Improved command line syntax and usage help (``--help``)
--  Tesseract 3.03+ PDF page rendering can be used instead for better positioning
-   of recognized text (``--pdf-renderer tesseract``)
--  PDF metadata (title, author, keywords) are now transferred to the 
-   output PDF
--  PDF metadata can also be set from the command line (``--title``, etc.)
--  Automatic repairs malformed input PDFs if possible
--  Added test cases to confirm everything is working
--  Added option to skip extremely large pages that take too long to OCR and are 
-   often not OCRable (e.g. large scanned maps or diagrams); other pages are still
-   processed (``--skip-big``)
--  Added option to kill Tesseract OCR process if it seems to be taking too long on
-   a page, while still processing other pages (``--tesseract-timeout``)
--  Less common colorspaces (CMYK, palette) are now supported by conversion to RGB
--  Multiple images on the same PDF page are now supported
-
-Changes
--------
-
--  New, robust rewrite in Python 3.4+ with ruffus_ pipelines
--  Now uses Ghostscript 9.14's improved color conversion model to preserve PDF colors
--  All "tasks" in the pipeline can be executed in parallel on any
-   available CPUs, increasing performance
--  The ``-o DPI`` argument has been phased out, in favor of ``--oversample DPI``, in
-   case we need ``-o OUTPUTFILE`` in the future
--  Removed several dependencies, so it's easier to install.  We no 
-   longer use:
-   
-   - GNU parallel_
-   - ImageMagick_
-   - Python 2.7
-   - Poppler
-   - MuPDF_ tools
-   - shell scripts
-   - Java and JHOVE_
-   - libxml2
-
--  Some new external dependencies are required or optional, compared to v2.x:
-
-   - Ghostscript 9.14+
-   - qpdf_ 5.0.0+
-   - Unpaper_ 6.1 (optional)
-   - some automatically managed Python packages
-  
-.. _ruffus: http://www.ruffus.org.uk/index.html
-.. _parallel: https://www.gnu.org/software/parallel/
-.. _ImageMagick: http://www.imagemagick.org/script/index.php
-.. _MuPDF: http://mupdf.com/docs/
-.. _qpdf: http://qpdf.sourceforge.net/
-.. _Unpaper: https://github.com/Flameeyes/unpaper
-.. _JHOVE: http://jhove.sourceforge.net/
-
-Release candidates
-------------------
-
--  rc9:
-
-   - fix issue #118: report error if ghostscript iccprofiles are missing
-   - fixed another issue related to #111: PDF rasterized to palette file
-   - add support image files with a palette
-   - don't try to validate PDF file after an exception occurs
-
--  rc8:
-
-   - fix issue #111: exception thrown if PDF is missing DocumentInfo dictionary
-
--  rc7:
-
-   - fix error when installing direct from pip, "no such file 'requirements.txt'"
-
--  rc6:
-
-   - dropped libxml2 (Python lxml) since Python 3's internal XML parser is sufficient
-   - set up Docker container
-   - fix Unicode errors if recognized text contains Unicode characters and system locale is not UTF-8
-
--  rc5:
-
-   - dropped Java and JHOVE in favour of qpdf
-   - improved command line error output
-   - additional tests and bug fixes
-   - tested on Ubuntu 14.04 LTS
-
--  rc4:
-
-   - dropped MuPDF in favour of qpdf
-   - fixed some installer issues and errors in installation instructions
-   - improve performance: run Ghostscript with multithreaded rendering
-   - improve performance: use multiple cores by default
-   - bug fix: checking for wrong exception on process timeout 
-
--  rc3: skipping version number intentionally to avoid confusion with Tesseract
--  rc2: first release for public testing to test-PyPI, Github
--  rc1: testing release process
-
-Compatibility notes
--------------------
-
--  ``./OCRmyPDF.sh`` script is still available for now
--  Stacking the verbosity option like ``-vvv`` is no longer supported
-
--  The configuration file ``config.sh`` has been removed.  Instead, you can
-   feed a file to the arguments for common settings:
-
-::
-
-   ocrmypdf input.pdf output.pdf @settings.txt
-
-where ``settings.txt`` contains *one argument per line*, for example:
-
-::
-
-   -l 
-   deu 
-   --author 
-   A. Merkel 
-   --pdf-renderer 
-   tesseract
-
-
-Fixes
------
-
--  Handling of filenames containing spaces: fixed
-
-Notes and known issues
-----------------------
-
--  Some dependencies may work with lower versions than tested, so try
-   overriding dependencies if they are "in the way" to see if they work.
-
--  ``--pdf-renderer tesseract`` will output files with an incorrect page size in Tesseract 3.03,
-   due to a bug in Tesseract.
-
--  PDF files containing "inline images" are not supported and won't be for the 3.0 release. Scanned
-   images almost never contain inline images.
-
-
-v2.2-stable (2014-09-29):
-=========================
-
-New features
-------------
-
-- None
-
-Changes
--------
-
-- Update to jhove v1.11
-- Request the python library reportlab v3.0 or newer (So that we could remove a patch to the previous version of reportlab leading to issues for some users)
-
-Fixes
------
-
-- Fix bug on Mac OS X (resolution of simlink to OCRmyPDF.sh script) (thanks to jbarlow83)
-- Check if the input pdf file exists before to continue
-
-Tested with
------------
-
-- Operating system: FreeBSD 9.2
-- Dependencies:
-
-   - parallel 20140822
-   - poppler-utils 0.24.5
-   - ImageMagick 6.8.9-4 2014-09-17
-   - Unpaper 0.3
-   - tesseract 3.02.02
-   - Python 2.7.8
-   - ghostcript (gs): 9.06
-   - java: openjdk version "1.7.0_65"
-
-
-v2.1-stable (2014-09-20):
-=========================
-
-New features
-------------
-
--  None
-
-Changes
--------
-
--  None
-
-Fixes
------
-
--  Allow execution via simlink
--  Add support for tesseract 3.03
--  Add support for newer version of reportlab
--  Lowered minimum version of gnu parallel
--  Various typo
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  parallel 20130222
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v2.0-stable (2014-01-25):
-=========================
-
-New features
-------------
-
--  Check if the language(s) passed using the -l option is supported by
-   tesseract (fixes #60)
-
-Changes
--------
-
--  Allow OCRmyPDF to be used with tesseract 3.02.01, even though OCR
-   might fail for few PDF file (see issue #28). Rationale: For some
-   linux distribution, no newer version than tesseract 3.02.01 is
-   available
-
-Fixes
------
-
--  More robust algorithm for checking the version of the installed
-   tesseract package
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  parallel 20130222
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v2.0-rc2 (2014-01-16):
-======================
-
-New features
-------------
-
--  None
-
-Changes
--------
-
--  Size reduction of final PDF file: (fixes #50)
--  Support for monochrome (Black&White) images (massive size reduction
-   in final PDF: >80%)
--  Reduced size of grayscale images (by 13% on test PDF file)
--  Preventing fi, fl ligatures does not require anymore to pass an
-   additional config file to tesseract using the -C option (fixes #58)
--  Location of temporary folder according to content of environment
-   variable TMPDIR.
--  Dependency to pdftk removed
--  Check for compatible versions of dependencies: (fixes #51)
--  parallel and tesseract
--  python libraries reportlab and lxml
-
-Fixes
------
-
--  Improved portability with various shells (dash, bash, tcsh) and OS
-   (FreeBSD, MAC OSX, Linux) (fixes #59)
--  Corrected bug in case the input PDF file contains a space character
-   (fixes #48)
--  Prevent spurious error message in case there is no image in a PDF
-   page
--  Prevent collision of temporary folder names (fixes #57)
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  parallel 20130222
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v2.0-rc1 (2014-01-07):
-======================
-
-New features
-------------
-
--  Huge performance improvement on machines having multiple CPU/cores
-   (processing of several pages concurrently) (fixes #18)
--  By default prevent from processing a PDF file already containing
-   fonts (i.e. text)(it can be overridden with the -f flag) (fixes #16)
--  Warn if the resolution is too low to get reasonable OCR results
-   (fixes #37)
--  New option (-o) to perform automatic oversampling if the image
-   resolution is too low. This can improve OCR results.
--  Warn if using a tesseract version older than v3.02.02 (as older
-   versions are known to produce invalid output) (fixes #41)
--  Echo version of the installed dependencies (e.g. tesseract) in debug
-   mode in order to ease support (fixes #35)
--  Echo the arguments passed to the script in debug mode to ease support
-
-Changes
--------
-
--  In debug mode: The debug page is now placed after the respective
-   "normal" page
--  Reduced disk space usage in temporary folder if -d (deskew) or -c
-   (cleanup) options are not selected
--  New file src/config.sh containing various configuration parameters
--  Documentation of the tesseract config file "tess-cfg/no\_ligature"
-   improved
--  Improved consistency of the temporary file names
-
-Fixes
------
-
--  Improved robustness:
--  in case vertical resolution differs from horizontal resolution (fixes
-   #38)
--  in case a PDF page contains more than one image (fixes #36)
--  Fix a problem occurring if python 3 is the standard interpreter
-   (fixes #33)
--  Fix a problem occurring if the input PDF file contains special
-   characters like "#" (fixes #34)
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  parallel 20130222
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  pdftk 1.45
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v1.1-stable (2014-01-06):
-=========================
-
-New features
-------------
-
--  N/A
-
-Changes
--------
-
--  N/A
-
-Fixes
------
-
--  Fixed syntax error (bashism) leading to an error message on certain
-   systems (fixes #42)
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  pdftk 1.45
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v1.0-stable (2013-05-06):
-=========================
-
-New features
-------------
-
--  In debug mode: compute and echo time required for processing (fixes
-   #26)
-
-Changes
--------
-
--  Removed feature to add metadata in final pdf file (because it lead to
-   to final PDF file that does not comply to the PDF/A-1 format)
--  Removed feature to set same owner & permissions in final PDF file
-   than in input file
--  Removed many unused jhove files (e.g. documentation, \*.java and
-   \*.class files)
-
-Fixes
------
-
--  Correction to handle correctly path and input PDF files having spaces
-   (fixes #31)
--  Resolutions (x/y) that are nearly equal are now supported (fixes #25)
--  Fix compatibility issue with Ubuntu server 12.04 / Ubuntu server
-   10.04 / Linux Mint 13 Maya and probably other Linux distributions
-   (fixes #27)
--  Commit missing jhove files (\*.jar mainly) due to wrong .gitignore
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  pdftk 1.45
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v1.0-rc2 (2013-04-29):
-======================
-
-New features
-------------
-
--  Keep temporary files if debug mode is set (fixes #22)
--  Set same owner & permissions in final PDF file than in input file
-   (fixes #9)
--  Added metadata in final pdf file (fixes #4)
-
-Changes
--------
-
--  N/A
-
-Fixes
------
-
--  Fixed wrong image cropping when deskew option is activated
--  Exit with error message if page size is not found in hocr file (fixes
-   #21)
--  Various minor fixes in log messages
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  pdftk 1.45
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"
-
-v1.0-rc1 (2013-04-26):
-======================
-
-New features
-------------
-
--  First release candidate
-
-Changes
--------
-
--  N/A
-
-Fixes
------
-
--  N/A
-
-Tested with
------------
-
--  Operating system: FreeBSD 9.1
--  Dependencies:
--  poppler-utils 0.22.2
--  ImageMagick 6.8.0-7 2013-03-30
--  Unpaper 0.3
--  tesseract 3.02.02
--  Python 2.7.3
--  pdftk 1.45
--  ghoscript (gs): 9.06
--  java: openjdk version "1.7.0\_17"

dev_requirements.txt

@@ -0,0 +1,4 @@
+check-manifest >= 0.35
+twine >= 1.8.1
+coverage >= 4.4
+GitPython == 2.1.3

docs/advanced.rst

@@ -0,0 +1,139 @@
+Advanced features
+=================
+
+Control of OCR options
+----------------------
+
+OCRmyPDF provides many features to control the behavior of the OCR engine, Tesseract.
+
+When OCR is skipped
+"""""""""""""""""""
+
+If a page in a PDF seems to have text, by default OCRmyPDF will exit without modifying the PDF. This is to ensure that PDFs that were previously OCRed or were "born digital" rather than scanned are not processed. 
+
+If ``--skip-text`` is issued, then no OCR will be performed on pages that already have text. The page will be copied to the output. This may be useful for documents that contain both "born digital" and scanned content, or to use OCRmyPDF to normalize and convert to PDF/A regardless of their contents.
+
+If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information.
+
+
+Time and image size limits
+""""""""""""""""""""""""""
+
+By default, OCRmyPDF permits tesseract to run for three minutes (180 seconds) per page. This is usually more than enough time to find all text on a reasonably sized page with modern hardware. 
+
+If a page is skipped, it will be inserted without OCR. If preprocessing was requested, the preprocessed image layer will be inserted.
+
+If you want to adjust the amount of time spent on OCR, change ``--tesseract-timeout``.  You can also automatically skip images that exceed a certain number of megapixels with ``--skip-big``. (A 300 DPI, 8.5×11" page is 8.4 megapixels.)
+
+.. code-block:: bash
+
+	# Allow 300 seconds for OCR; skip any page larger than 50 megapixels
+	ocrmypdf --tesseract-timeout 300 --skip-big 50 bigfile.pdf output.pdf
+
+Overriding default tesseract
+""""""""""""""""""""""""""""
+
+OCRmyPDF checks the system ``PATH`` for the ``tesseract`` binary.  
+
+Some relevant environment variables that influence Tesseract's behavior include:
+
+.. envvar:: TESSDATA_PREFIX
+
+	Overrides the path to Tesseract's data files. This can allow simultaneous installation of the "best" and "fast" training data sets. OCRmyPDF does not manage this environment variable.
+
+.. envvar:: OMP_THREAD_LIMIT
+
+	Controls the number of threads Tesseract will use. OCRmyPDF will manage this environment if it is not already set. (Currently, it will set it to 1 because this gives the best results in testing.)
+
+For example, if you are testing tesseract 4.00 and don't wish to use an existing tesseract 3.04 installation, you can launch OCRmyPDF as follows:
+
+.. code-block:: bash
+
+	env \
+		PATH=/home/user/src/tesseract4/api:$PATH \
+		TESSDATA_PREFIX=/home/user/src/tesseract4 \
+		ocrmypdf --tesseract-oem 2 input.pdf output.pdf
+
+In this example ``TESSDATA_PREFIX`` directs Tesseract 4.0 to use LSTM training data. ``--tesseract-oem 1`` requests tesseract 4.0's new LSTM engine. (Tesseract 4.0 only.)
+
+
+Overriding other support programs
+"""""""""""""""""""""""""""""""""
+
+In addition to tesseract, OCRmyPDF uses the following external binaries:
+
+* ``gs`` (Ghostscript)
+* ``unpaper``
+* ``qpdf``
+
+In each case OCRmyPDF will search the ``PATH`` environment variable to locate the binaries.
+
+
+Changing tesseract configuration variables
+""""""""""""""""""""""""""""""""""""""""""
+
+You can override tesseract's default `control parameters <https://github.com/tesseract-ocr/tesseract/wiki/ControlParams>`_ with a configuration file.
+
+As an example, this configuration will disable Tesseract's dictionary for current language. Normally the dictionary is helpful for interpolating words that are unclear, but it may interfere with OCR if the document does not contain many words (for example, a list of part numbers).
+
+Create a file named "no-dict.cfg" with these contents:
+
+::
+
+	load_system_dawg 0
+	language_model_penalty_non_dict_word 0
+	language_model_penalty_non_freq_dict_word 0
+
+then run ocrmypdf as follows (along with any other desired arguments):
+
+.. code-block:: bash
+
+	ocrmypdf --tesseract-config no-dict.cfg input.pdf output.pdf
+
+.. warning::
+
+	Some combinations of control parameters will break Tesseract or break assumptions that OCRmyPDF makes about Tesseract's output.
+
+
+Changing the PDF renderer
+-------------------------
+
+rasterizing
+  Converting a PDF to an image for display.
+
+rendering
+  Creating a new PDF from other data (such as an existing PDF).
+
+
+OCRmyPDF has three PDF renderers: ``sandwich``, ``hocr``, ``tesseract``. The renderer may be selected using ``--pdf-renderer``. The default is ``auto`` which lets OCRmyPDF select the renderer to use. Currently, ``auto`` selects ``sandwich`` for Tesseract 3.05.01 or newer, or ``hocr`` for older versions of Tesseract.
+
+The ``sandwich`` renderer
+"""""""""""""""""""""""""
+
+The ``sandwich`` renderer uses Tesseract's new text-only PDF feature, which produces a PDF page that lays out the OCR in invisible text. This page is then "sandwiched" onto the original PDF page, allowing lossless application of OCR even to PDF pages that contain other vector objects.
+
+Currently this is the best renderer for most uses, however it is implemented in Tesseract so OCRmyPDF cannot influence it. Currently some problematic PDF viewers like Mozilla PDF.js and macOS Preview have problems with segmenting its text output, and mightrunseveralwordstogether.
+
+When image preprocessing features like ``--deskew`` are used, the original PDF will be rendered as a full page and the OCR layer will be placed on top.
+
+This renderer requires Tesseract 3.05.01 or newer.
+
+The ``hocr`` renderer
+"""""""""""""""""""""
+
+The ``hocr`` renderer works with older versions of Tesseract. The image layer is copied from the original PDF page if possible, avoiding potentially lossy transcoding or loss of other PDF information. If preprocessing is specified, then the image layer is a new PDF.
+
+Unlike ``sandwich`` this renderer is implemented within OCRmyPDF; anyone looking to customize how OCR is presented should look here. A major disadvantage of this renderer is it not capable of correctly handling text outside the Latin alphabet. Pull requests to improve the situation are welcome.
+
+Currently, this renderer has the best compatibility with Mozilla's PDF.js viewer.
+
+This works in all versions of Tesseract.
+
+The ``tesseract`` renderer
+""""""""""""""""""""""""""
+
+The ``tesseract`` renderer creates a PDF with the image and text layers precomposed, meaning that it always transcodes, loses image quality and rasterizes any vector objects. It does a better job on non-Latin text and document structure than ``hocr``.
+
+If a PDF created with this renderer using Tesseract versions older than 3.05.00 is then passed through Ghostscript's pdfwrite feature, the OCR text *may* be corrupted. The ``--output-type=pdfa`` argument will produce a warning in this situation.
+
+*This renderer is deprecated and will be removed whenever support for older versions of Tesseract is dropped.*

docs/batch.rst

@@ -0,0 +1,214 @@
+Batch processing
+================
+
+This article provides information about running OCRmyPDF on multiple files or configuring it as a service triggered by file system events.
+
+Batch jobs
+----------
+
+Consider using the excellent `GNU Parallel <https://www.gnu.org/software/parallel/>`_ to apply OCRmyPDF to multiple files at once.
+
+Both ``parallel`` and ``ocrmypdf`` will try to use all available processors. To maximize parallelism without overloading your system with processes, consider using ``parallel -j 2`` to limit parallel to running two jobs at once.
+
+This command will run all ocrmypdf all files named ``*.pdf`` in the current directory and write them to the previous created ``output/`` folder. It will not search subdirectories.
+
+The ``--tag`` argument tells parallel to print the filename as a prefix whenever a message is printed, so that one can trace any errors to the file that produced them.
+
+.. code-block:: bash
+
+	parallel --tag -j 2 ocrmypdf '{}' 'output/{}' ::: *.pdf
+
+OCRmyPDF automatically repairs PDFs before parsing and gathering information from them.  If you are already repairing PDFs with ``qpdf`` prior to attempting OCR, or you can use ``--skip-repair`` to skip this step.  It may improve performance for large files, since repairing PDFs is single-threaded.
+
+Directory trees
+---------------
+
+This will walk through a directory tree and run OCR on all files in place, printing the output in a way that makes 
+
+.. code-block:: bash
+
+	find . --printf '%p' -name '*.pdf' -exec ocrmypdf '{}' '{}' \;
+
+This only runs one ``ocrmypdf`` process at a time. This variation uses ``find`` to create a directory list and ``parallel`` to parallelize runs of ``ocrmypdf``, again updating files in place.
+
+.. code-block:: bash
+
+	find . -name '*.pdf' | parallel --tag -j 2 ocrmypdf '{}' '{}'
+
+
+Sample script
+"""""""""""""
+
+This user contributed script also provides an example of batch processing.
+
+.. code-block:: python
+
+	#!/usr/bin/env python3
+	# Walk through directory tree, replacing all files with OCR'd version
+	# Contributed by DeliciousPickle@github
+
+	import logging
+	import os
+	import subprocess
+	import sys
+
+	script_dir = os.path.dirname(os.path.realpath(__file__))
+	print(script_dir + '/ocr-tree.py: Start')
+
+	if len(sys.argv) > 1:
+	    start_dir = sys.argv[1]
+	else:
+	    start_dir = '.'
+
+	if len(sys.argv) > 2:
+	    log_file = sys.argv[2]
+	else:
+	    log_file = script_dir + '/ocr-tree.log'
+
+	logging.basicConfig(
+			level=logging.INFO, format='%(asctime)s %(message)s', 
+			filename=log_file, filemode='w')
+
+	for dir_name, subdirs, file_list in os.walk(start_dir):
+	    logging.info('\n')
+	    logging.info(dir_name + '\n')
+	    os.chdir(dir_name)
+	    for filename in file_list:
+	        file_ext = os.path.splitext(filename)[1]
+	        if file_ext == '.pdf':
+	            full_path = dir_name + '/' + filename
+	            print(full_path)
+	            cmd = ["ocrmypdf",  "--deskew", filename, filename]
+	            logging.info(cmd)
+	            proc = subprocess.Popen(
+	            	cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+	            result = proc.stdout.read()
+	            if proc.returncode == 6:
+	                print("Skipped document because it already contained text")
+	            elif proc.returncode == 0:
+	                print("OCR complete")
+	            logging.info(result)
+
+API
+"""
+
+OCRmyPDF is currently supported as a command line interface. This means that even if you are using OCRmyPDF in a Python script, you should run it in a subprocess rather importing the ocrmypdf package.
+
+The reason for this limitation is that the `ruffus <https://github.com/bunbun/ruffus/>`_ library that OCRmyPDF depends on is unfortunately not reentrant. OCRmyPDF works by defining each operation it does as a ruffus task that takes one or more files as input and generates one or more files as output. As such ruffus is fairly fundamental.
+
+(If you find individual functions implemented in OCRmyPDF useful (such as ``ocrmypdf.pdfinfo``), you can use these if you wish to.)
+
+
+Synology DiskStations
+"""""""""""""""""""""
+
+Synology DiskStations (Network Attached Storage devices) can run the Docker image of OCRmyPDF if the Synology `Docker package <https://www.synology.com/en-global/dsm/packages/Docker>`_ is installed. Attached is a script to address particular quirks of using OCRmyPDF on one of these devices.
+
+This is only possible for x86-based Synology products. Some Synology products use ARM or Power processors and do not support Docker. Further adjustments might be needed to deal with the Synology's relatively limited CPU and RAM.
+
+.. code-block:: python
+
+	#!/bin/env python3
+	# Contributed by github.com/Enantiomerie
+
+	# script needs 2 arguments
+	# 1. source dir with *.pdf - default is location of script
+	# 2. move dir where *.pdf and *_OCR.pdf are moved to
+
+	import logging
+	import os
+	import subprocess
+	import sys
+	import time
+	import shutil
+
+	script_dir = os.path.dirname(os.path.realpath(__file__))
+	timestamp = time.strftime("%Y-%m-%d-%H%M_")
+	log_file = script_dir + '/' + timestamp + 'ocrmypdf.log'
+	logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s', filename=log_file, filemode='w')
+
+	if len(sys.argv) > 1:
+	    start_dir = sys.argv[1]
+	else:
+	    start_dir = '.'
+
+	for dir_name, subdirs, file_list in os.walk(start_dir):
+	    logging.info('\n')
+	    logging.info(dir_name + '\n')
+	    os.chdir(dir_name)
+	    for filename in file_list:
+	        file_ext = os.path.splitext(filename)[1]
+	        if file_ext == '.pdf':
+	            full_path = dir_name + '/' + filename
+	            file_noext = os.path.splitext(filename)[0]
+	            timestamp_OCR = time.strftime("%Y-%m-%d-%H%M_OCR_")
+	            filename_OCR = timestamp_OCR + file_noext + '.pdf'
+	            docker_mount = dir_name + ':/home/docker'
+	# create string for pdf processing 
+	# diskstation needs a user:group docker:docker. find uid:gid of your diskstation docker:docker with id docker.
+	# use this uid:gid in -u flag
+	# rw rights for docker:docker at source dir are also necessary
+	# the script is processed as root user via chron 
+	            cmd = ['docker', 'run', '--rm', '-v', docker_mount, '-u=1030:65538', 'jbarlow83/ocrmypdf', , '--deskew' , filename, filename_OCR]
+	            logging.info(cmd)
+	            proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+	            result = proc.stdout.read()
+	            logging.info(result)
+	            full_path_OCR = dir_name + '/' + filename_OCR
+	            os.chmod(full_path_OCR, 0o666)
+	            os.chmod(full_path, 0o666)
+	            full_path_OCR_archive = sys.argv[2]
+	            full_path_archive = sys.argv[2] + '/no_ocr'
+	            shutil.move(full_path_OCR,full_path_OCR_archive)
+	            shutil.move(full_path, full_path_archive)
+	logging.info('Finished.\n')
+
+
+
+Huge batch jobs
+"""""""""""""""
+
+If you have thousands of files to work with, contact the author. Consulting work related to OCRmyPDF helps fund this open source project and all inquiries are appreciated.
+
+
+Hot (watched) folders
+---------------------
+
+To set up a "hot folder" that will trigger OCR for every file inserted, use a program like Python `watchdog <https://pypi.python.org/pypi/watchdog>`_ (supports all major OS).
+
+One could then configure a scanner to automatically place scanned files in a hot folder, so that they will be queued for OCR and copied to the destination.
+
+.. code-block:: bash
+
+	pip install watchdog
+
+watchdog installs the command line program ``watchmedo``, which can be told to run ``ocrmypdf`` on any .pdf added to the current directory (``.``) and place the result in the previously created ``out/`` folder.
+
+.. code-block:: bash
+
+	cd hot-folder
+	mkdir out
+	watchmedo shell-command \
+		--patterns="*.pdf" \
+		--ignore-directories \
+		--command='ocrmypdf "${watch_src_path}" "out/${watch_src_path}" ' \
+		.  # don't forget the final dot
+
+For more complex behavior you can write a Python script around to use the watchdog API.
+
+On file servers, you could configure watchmedo as a system service so it will run all the time.
+
+Caveats
+"""""""
+
+* ``watchmedo`` may not work properly on a networked file system, depending on the capabilities of the file system client and server.
+* This simple recipe does not filter for the type of file system event, so file copies, deletes and moves, and directory operations, will all be sent to ocrmypdf, producing errors in several cases. Disable your watched folder if you are doing anything other than copying files to it.
+* If the source and destination directory are the same, watchmedo may create an infinite loop.
+* On BSD, FreeBSD and older versions of macOS, you may need to increase the number of file descriptors to monitor more files, using ``ulimit -n 1024`` to watch a folder of up to 1024 files.
+
+Alternatives
+""""""""""""
+
+* `Watchman <https://facebook.github.io/watchman/>`_ is a more powerful alternative to ``watchmedo``.
+
+

docs/bitmap_vs_svg.svg

@@ -0,0 +1,35 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg xmlns="http://www.w3.org/2000/svg" width="100%" height="100%" viewBox="0 0 915 585" xmlns:xlink="http://www.w3.org/1999/xlink">
+<defs>
+<linearGradient id="b" y2="445" gradientUnits="userSpaceOnUse" y1="179" gradientTransform="translate(0 -2.06)" x2="-29.7" x1="322">
+<stop stop-color="#333" offset="0"/>
+<stop stop-color="#fff" stop-opacity="0" offset="1"/>
+</linearGradient>
+<linearGradient id="a" y2="414" gradientUnits="userSpaceOnUse" y1="159" x2="490" x1="815">
+<stop stop-color="#33f" offset="0"/>
+<stop stop-color="#3f3fff" stop-opacity="0" offset="1"/>
+</linearGradient>
+</defs>
+<path fill="url(#b)" d="m403 247c-12 115-135 122-368 123-4.3-1.07-7.32-7.33-6-41 76-37 151-124 167-236 93 123 201 40.9 207 154z"/>
+<g stroke-width="3.45" fill="none">
+<path stroke="#000" d="m11.8 11.8h411v411l-411 0.01v-411z"/>
+<path stroke="#448" d="m489 11.7h415v411h-415v-411z"/>
+</g>
+<path d="m876 244c-12 115-133 120-366 121-6-14-10-40-3-43 76-37.3 136-106 152-218 38 48 209 101 217 140z" fill="url(#a)"/>
+<g id="RasterLarge" transform="matrix(1.36 0 0 1.28 -161 -636)">
+<path fill="#999" d="m287 730h-30v-60h20v20h10v10h10v10h20v10h40v-10h20v-20h10v-20h-10v-10h-10v-10h-40v-10h-30v-10h-20v-10h-10v-10h-10v-60h10v-10h10v-10h10v-10h90v10h30v60h-20v-10h-10v-20h-10v-10h-20v-10h-30v10h-20v10h-10v20h10v20h30v10h30v10h30v10h10v10h10v10h10v60h-10v10h-10v10h-10v10h-100v-10z"/>
+<path fill="#555" d="m297 730h-30v-10h-10v-50h20v20h10v20h10 10v10h60v-10h10v-10h10v-40h-20v-10h-20v-10h-40v-10h-20v-10h-20v-20h-10v-50h10v-10h10v-10h20v-10h70v10h30v10h10v50h-20v-20h-10v-20h-20v-10h-50v10h-10v10h-10v30h10v10h20v10h30v10h30v10h20v10h10v10h10v50h-10v20h-10v10h-20v10h-80z"/>
+<path d="m307 730h-30v-10h-20v-50h20v30h10v10h10v10h70v-10h20v-50h-10v-10h-20v-10h-40v-10h-30v-10h-10v-10h-10v-20h-10v-30h10v-20h20v-10h20v-10h50v10h30v10h20v50h-20v-30h-10v-10h-10v-10h-70v10h-10v40h10v10h10v10h30v10h40v10h20v10h10v20h10v40h-10v20h-20v10h-30v10h-50v-10z"/>
+</g>
+<g font-size="40" font-family="sans-serif" text-anchor="middle">
+<g font-size="100">
+<text y="518" x="210">Raster</text>
+<text y="518" x="695" fill="#338">Vector</text>
+</g>
+<text y="563" x="210">.jpeg  .gif  .png</text>
+<text y="563" x="696" fill="#338">.svg</text>
+</g>
+<path id="VectorLarge" fill="#005" d="m661 294v-62.5l23.4 0.184c0.678 20.8 7.32 36.3 19.9 46.4 12.7 9.93 32.1 14.9 57.9 14.9 24.1 0 42.5-4.29 55.1-12.9 12.7-8.71 19.1-21.3 19.1-37.9 0-13.2-3.86-23.4-11.6-30.5-7.59-7.11-23.7-14-48.4-20.8l-40.1-10.9c-29-7.97-49.5-17.9-61.4-29.8-11.8-11.9-17.7-28.2-17.7-48.9 0-23.3 9.15-41.4 27.4-54.3 18.3-12.9 44-19.3 77.1-19.3 14.1 0 29.6 1.41 46.4 4.23 16.8 2.7 34.7 6.68 53.7 12v58.5h-23c-2.3-19.4-9.49-33.4-21.6-41.9-11.9-8.71-30.2-13.1-54.7-13.1-21.4 0-37.7 3.99-49 12-11.1 7.85-16.7 19.3-16.7 34.4 0 13.1 4.2 23.4 12.6 30.9 8.4 7.48 26.2 14.9 53.5 22.3l37.6 10.1c27.5 7.48 47.1 17 58.8 28.7 11.8 11.5 17.7 27 17.7 46.5 0 26.6-9.42 46.7-28.3 60.2s-46.9 20.2-84.2 20.2c-16.7 0-33.7-1.53-51-4.6-18-3-35-7-53-13z"/>
+<use xlink:href="#VectorLarge" transform="matrix(.17 0 0 .17 392 313)"/>
+<use xlink:href="#RasterLarge" transform="matrix(.173 0 0 .173 -8.25 314)" height="100%" width="100%" y="0" x="0"/>
+</svg>

docs/conf.py

@@ -0,0 +1,350 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+#
+# ocrmypdf documentation build configuration file, created by
+# sphinx-quickstart on Sun Sep  4 14:29:43 2016.
+#
+# This file is execfile()d with the current directory set to its
+# containing dir.
+#
+# Note that not all possible configuration values are present in this
+# autogenerated file.
+#
+# All configuration values have a default; values that are commented out
+# serve to show the default.
+
+# If extensions (or modules to document with autodoc) are in another directory,
+# add these directories to sys.path here. If the directory is relative to the
+# documentation root, use os.path.abspath to make it absolute, like shown here.
+#
+# import os
+# import sys
+# sys.path.insert(0, os.path.abspath('.'))
+
+# -- General configuration ------------------------------------------------
+
+# If your documentation needs a minimal Sphinx version, state it here.
+#
+# needs_sphinx = '1.0'
+
+# Add any Sphinx extension module names here, as strings. They can be
+# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
+# ones.
+extensions = [
+    # 'sphinx.ext.mathjax',
+]
+
+# Add any paths that contain templates here, relative to this directory.
+templates_path = ['_templates']
+
+# The suffix(es) of source filenames.
+# You can specify multiple suffix as a list of string:
+#
+# source_suffix = ['.rst', '.md']
+source_suffix = '.rst'
+
+# The encoding of source files.
+#
+# source_encoding = 'utf-8-sig'
+
+# The master toctree document.
+master_doc = 'index'
+
+# General information about the project.
+project = 'ocrmypdf'
+copyright = '2018, James R. Barlow. Licensed under Creative Commons Attribution-ShareAlike 4.0.'
+author = 'James R. Barlow'
+
+# The version info for the project you're documenting, acts as replacement for
+# |version| and |release|, also used in various other places throughout the
+# built documents.
+#
+# The short X.Y version.
+from ocrmypdf import __version__ as OCRMYPDF_VERSION
+
+_version_parts = OCRMYPDF_VERSION.split('.')
+
+version = '.'.join(_version_parts[0:2])
+# The full version, including alpha/beta/rc tags.
+release = OCRMYPDF_VERSION
+
+# The language for content autogenerated by Sphinx. Refer to documentation
+# for a list of supported languages.
+#
+# This is also used if you do content translation via gettext catalogs.
+# Usually you set "language" from the command line for these cases.
+language = None
+
+# There are two options for replacing |today|: either, you set today to some
+# non-false value, then it is used:
+#
+# today = ''
+#
+# Else, today_fmt is used as the format for a strftime call.
+#
+today_fmt = '%Y-%m-%d'
+
+# List of patterns, relative to source directory, that match files and
+# directories to ignore when looking for source files.
+# This patterns also effect to html_static_path and html_extra_path
+exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
+
+# The reST default role (used for this markup: `text`) to use for all
+# documents.
+#
+# default_role = None
+
+# If true, '()' will be appended to :func: etc. cross-reference text.
+#
+# add_function_parentheses = True
+
+# If true, the current module name will be prepended to all description
+# unit titles (such as .. function::).
+#
+# add_module_names = True
+
+# If true, sectionauthor and moduleauthor directives will be shown in the
+# output. They are ignored by default.
+#
+# show_authors = False
+
+# The name of the Pygments (syntax highlighting) style to use.
+pygments_style = 'sphinx'
+
+# A list of ignored prefixes for module index sorting.
+# modindex_common_prefix = []
+
+# If true, keep warnings as "system message" paragraphs in the built documents.
+# keep_warnings = False
+
+# If true, `todo` and `todoList` produce output, else they produce nothing.
+todo_include_todos = False
+
+
+# -- Options for HTML output ----------------------------------------------
+
+import sphinx_rtd_theme
+
+# The theme to use for HTML and HTML Help pages.  See the documentation for
+# a list of builtin themes.
+#
+html_theme = 'sphinx_rtd_theme'
+html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]
+
+# Theme options are theme-specific and customize the look and feel of a theme
+# further.  For a list of options available for each theme, see the
+# documentation.
+#
+html_theme_options = {
+    'display_version': False
+}
+
+# Add any paths that contain custom themes here, relative to this directory.
+# html_theme_path = []
+
+# The name for this set of Sphinx documents.
+# "<project> v<release> documentation" by default.
+#
+# html_title = 'ocrmypdf v4.2'
+
+# A shorter title for the navigation bar.  Default is the same as html_title.
+#
+# html_short_title = None
+
+# The name of an image file (relative to this directory) to place at the top
+# of the sidebar.
+#
+# html_logo = None
+
+# The name of an image file (relative to this directory) to use as a favicon of
+# the docs.  This file should be a Windows icon file (.ico) being 16x16 or 32x32
+# pixels large.
+#
+# html_favicon = None
+
+# Add any paths that contain custom static files (such as style sheets) here,
+# relative to this directory. They are copied after the builtin static files,
+# so a file named "default.css" will overwrite the builtin "default.css".
+html_static_path = ['_static']
+
+# Add any extra paths that contain custom files (such as robots.txt or
+# .htaccess) here, relative to this directory. These files are copied
+# directly to the root of the documentation.
+#
+# html_extra_path = []
+
+# If not None, a 'Last updated on:' timestamp is inserted at every page
+# bottom, using the given strftime format.
+# The empty string is equivalent to '%b %d, %Y'.
+#
+# html_last_updated_fmt = None
+
+# If true, SmartyPants will be used to convert quotes and dashes to
+# typographically correct entities.
+#
+# html_use_smartypants = True
+
+# Custom sidebar templates, maps document names to template names.
+#
+# html_sidebars = {}
+
+# Additional templates that should be rendered to pages, maps page names to
+# template names.
+#
+# html_additional_pages = {}
+
+# If false, no module index is generated.
+#
+# html_domain_indices = True
+
+# If false, no index is generated.
+#
+# html_use_index = True
+
+# If true, the index is split into individual pages for each letter.
+#
+# html_split_index = False
+
+# If true, links to the reST sources are added to the pages.
+#
+# html_show_sourcelink = True
+
+# If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
+#
+# html_show_sphinx = True
+
+# If true, "(C) Copyright ..." is shown in the HTML footer. Default is True.
+#
+# html_show_copyright = True
+
+# If true, an OpenSearch description file will be output, and all pages will
+# contain a <link> tag referring to it.  The value of this option must be the
+# base URL from which the finished HTML is served.
+#
+# html_use_opensearch = ''
+
+# This is the file name suffix for HTML files (e.g. ".xhtml").
+# html_file_suffix = None
+
+# Language to be used for generating the HTML full-text search index.
+# Sphinx supports the following languages:
+#   'da', 'de', 'en', 'es', 'fi', 'fr', 'h', 'it', 'ja'
+#   'nl', 'no', 'pt', 'ro', 'r', 'sv', 'tr', 'zh'
+#
+# html_search_language = 'en'
+
+# A dictionary with options for the search language support, empty by default.
+# 'ja' uses this config value.
+# 'zh' user can custom change `jieba` dictionary path.
+#
+# html_search_options = {'type': 'default'}
+
+# The name of a javascript file (relative to the configuration directory) that
+# implements a search results scorer. If empty, the default will be used.
+#
+# html_search_scorer = 'scorer.js'
+
+# Output file base name for HTML help builder.
+htmlhelp_basename = 'ocrmypdfdoc'
+
+# -- Options for LaTeX output ---------------------------------------------
+
+latex_elements = {
+     # The paper size ('letterpaper' or 'a4paper').
+     #
+     # 'papersize': 'letterpaper',
+
+     # The font size ('10pt', '11pt' or '12pt').
+     #
+     # 'pointsize': '10pt',
+
+     # Additional stuff for the LaTeX preamble.
+     #
+     # 'preamble': '',
+
+     # Latex figure (float) alignment
+     #
+     # 'figure_align': 'htbp',
+}
+
+# Grouping the document tree into LaTeX files. List of tuples
+# (source start file, target name, title,
+#  author, documentclass [howto, manual, or own class]).
+latex_documents = [
+    (master_doc, 'ocrmypdf.tex', 'ocrmypdf Documentation',
+     'James R. Barlow', 'manual'),
+]
+
+# The name of an image file (relative to this directory) to place at the top of
+# the title page.
+#
+# latex_logo = None
+
+# For "manual" documents, if this is true, then toplevel headings are parts,
+# not chapters.
+#
+# latex_use_parts = False
+
+# If true, show page references after internal links.
+#
+# latex_show_pagerefs = False
+
+# If true, show URL addresses after external links.
+#
+# latex_show_urls = False
+
+# Documents to append as an appendix to all manuals.
+#
+# latex_appendices = []
+
+# It false, will not define \strong, \code, 	itleref, \crossref ... but only
+# \sphinxstrong, ..., \sphinxtitleref, ... To help avoid clash with user added
+# packages.
+#
+# latex_keep_old_macro_names = True
+
+# If false, no module index is generated.
+#
+# latex_domain_indices = True
+
+
+# -- Options for manual page output ---------------------------------------
+
+# One entry per manual page. List of tuples
+# (source start file, name, description, authors, manual section).
+man_pages = [
+    (master_doc, 'ocrmypdf', 'ocrmypdf Documentation',
+     [author], 1)
+]
+
+# If true, show URL addresses after external links.
+#
+# man_show_urls = False
+
+
+# -- Options for Texinfo output -------------------------------------------
+
+# Grouping the document tree into Texinfo files. List of tuples
+# (source start file, target name, title, author,
+#  dir menu entry, description, category)
+texinfo_documents = [
+    (master_doc, 'ocrmypdf', 'ocrmypdf Documentation',
+     author, 'ocrmypdf', 'One line description of project.',
+     'Miscellaneous'),
+]
+
+# Documents to append as an appendix to all manuals.
+#
+# texinfo_appendices = []
+
+# If false, no module index is generated.
+#
+# texinfo_domain_indices = True
+
+# How to display URL addresses: 'footnote', 'no', or 'inline'.
+#
+# texinfo_show_urls = 'footnote'
+
+# If true, do not generate a @detailmenu in the "Top" node's menu.
+#
+# texinfo_no_detailmenu = False

docs/cookbook.rst

@@ -0,0 +1,201 @@
+Cookbook
+========
+
+Basic examples
+--------------
+
+Help!
+"""""
+
+ocrmypdf has built-in help.
+
+.. code-block:: bash
+
+    ocrmypdf --help
+
+
+Add an OCR layer and convert to PDF/A
+"""""""""""""""""""""""""""""""""""""
+
+.. code-block:: bash
+
+    ocrmypdf input.pdf output.pdf
+
+Add an OCR layer and output a standard PDF
+""""""""""""""""""""""""""""""""""""""""""
+
+.. code-block:: bash
+
+    ocrmypdf --output-type pdf input.pdf output.pdf
+
+Create a PDF/A with all color and grayscale images converted to JPEG
+""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
+
+.. code-block:: bash
+
+    ocrmypdf --output-type pdfa --pdfa-image-compression jpeg input.pdf output.pdf
+
+Modify a file in place
+""""""""""""""""""""""
+
+The file will only be overwritten if OCRmyPDF is successful.
+
+.. code-block:: bash
+
+    ocrmypdf myfile.pdf myfile.pdf
+
+Correct page rotation
+"""""""""""""""""""""
+
+OCR will attempt to automatic correct the rotation of each page. This can help fix a scanning job that contains a mix of landscape and portrait pages.
+
+.. code-block:: bash
+
+    ocrmypdf --rotate-pages myfile.pdf myfile.pdf
+
+You can increase (decrease) the parameter ``--rotate-pages-threshold`` to make page rotation more (less) aggressive.
+
+If the page is "just a little off horizontal", like a crooked picture, then you want ``--deskew``. ``--rotate-pages`` is for when the cardinal angle is wrong.
+
+
+OCR languages other than English
+""""""""""""""""""""""""""""""""
+
+By default OCRmyPDF assumes the document is English. 
+
+.. code-block:: bash
+
+    ocrmypdf -l fra LeParisien.pdf LeParisien.pdf
+    ocrmypdf -l eng+fra Bilingual-English-French.pdf Bilingual-English-French.pdf
+
+Language packs must be installed for all languages specified. See :ref:`Installing additional language packs <lang-packs>`.
+
+
+Produce PDF and text file containing OCR text
+"""""""""""""""""""""""""""""""""""""""""""""
+
+This produces a file named "output.pdf" and a companion text file named "output.txt".
+
+.. code-block:: bash
+
+    ocrmypdf --sidecar output.txt input.pdf output.pdf
+
+OCR images, not PDFs
+--------------------
+
+If you are starting with images, you can just use Tesseract 3.04 or later directly to convert images to PDFs:
+
+.. code-block:: bash
+
+    tesseract my-image.jpg output-prefix pdf
+
+.. code-block:: bash
+
+    # When there are multiple images
+    tesseract text-file-containing-list-of-image-filenames.txt output-prefix pdf
+
+Tesseract's PDF output is quite good – OCRmyPDF uses it by internally by default. However, OCRmyPDF has many features not available in Tesseract like like image processing, metadata control, and PDF/A generation.
+
+Use a program like `img2pdf <https://gitlab.mister-muffin.de/josch/img2pdf>`_ to convert your images to PDFs, and then pipe the results to run ocrmypdf.  The `-` tells ocrmypdf to read standard input.
+
+.. code-block:: bash
+
+    img2pdf my-images*.jpg | ocrmypdf - myfile.pdf
+
+``img2pdf`` is recommended because it does an excellent job at generating PDFs without transcoding images.
+
+For convenience, OCRmyPDF can also convert single images to PDFs on its own. If the resolution (dots per inch, DPI) of an image is not set or is incorrect, it can be overridden with ``--image-dpi``. (As 1 inch is 2.54 cm, 1 dpi = 0.39 dpcm).
+
+.. code-block:: bash
+
+    ocrmypdf --image-dpi 300 image.png myfile.pdf
+
+If you have multiple images, you must use ``img2pdf`` to convert the images to PDF.
+
+.. note::
+
+    ImageMagick ``convert`` can also convert a group of images to PDF, but in the author's experience it takes a long time, transcodes unnecessarily and gives poor results.
+
+
+Image processing
+----------------
+
+OCRmyPDF perform some image processing on each page of a PDF, if desired.  The same processing is applied to each page.  It is suggested that the user review files after image processing as these commands might remove desirable content, especially from poor quality scans.
+
+* ``--rotate-pages`` attempts to determine the correct orientation for each page and rotates the page if necessary.
+
+* ``--remove-background`` attempts to detect and remove a noisy background from grayscale or color images.  Monochrome images are ignored. This should not be used on documents that contain color photos as it may remove them.
+
+* ``--deskew`` will correct pages were scanned at a skewed angle by rotating them back into place.  Skew determination and correction is performed using `Postl's variance of line sums <http://www.leptonica.com/skew-measurement.html>`_ algorithm as implemented in `Leptonica <http://www.leptonica.com/index.html>`_.
+  
+* ``--clean`` uses `unpaper <https://www.flameeyes.eu/projects/unpaper>`_ to clean up pages before OCR, but does not alter the final output.  This makes it less likely that OCR will try to find text in background noise.
+
+* ``--clean-final`` uses unpaper to clean up pages before OCR and inserts the page into the final output.  You will want to review each page to ensure that unpaper did not remove something important.
+
+.. note::
+
+    In many cases image processing will rasterize PDF pages as images, potentially losing quality.
+
+.. warning::
+
+    ``--clean-final`` and ``-remove-background`` may leave undesirable visual artifacts in some images where their algorithms have shortcomings. Files should be visually reviewed after using these options.
+
+
+OCR and correct document skew (crooked scan)
+""""""""""""""""""""""""""""""""""""""""""""
+
+Deskew:
+
+.. code-block:: bash
+
+    ocrmypdf --deskew input.pdf output.pdf
+
+Image processing commands can be combined. The order in which options are given does not matter. OCRmyPDF always applies the steps of the image processing pipeline in the same order (rotate, remove background, deskew, clean).
+
+.. code-block:: bash
+
+    ocrmypdf --deskew --clean --rotate-pages input.pdf output.pdf
+
+
+Don't actually OCR my PDF
+"""""""""""""""""""""""""
+
+If you set ``--tesseract-timeout 0`` OCRmyPDF will apply its image processing without performing OCR, if all you want to is to apply image processing or PDF/A conversion.
+
+.. code-block:: bash
+
+    ocrmypdf --tesseract-timeout=0 --remove-background input.pdf output.pdf
+
+
+Redo OCR
+""""""""
+
+To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--force-ocr`` argument. Normally, OCRmyPDF does not modify files that already appear to contain OCR text.
+
+.. code-block:: bash
+
+    ocrmypdf --force-ocr input.pdf output.pdf
+
+Note that the method above will force rasterization of all pages, potentially reducing quality or losing vector content. 
+
+To ensure quality is preserved, one could extract all of the images and rebuild the PDF for a lossless transformation. This recipe does not work when PDFs contain multiple images per page, as many do in practice. It will also lose any page rotation information.
+
+.. code-block:: bash
+
+    pdfimages -all old-ocr.pdf prefix  # extract all images
+    img2pdf -o temp.pdf prefix*        # construct new PDF from the images
+    # review the new PDF to ensure it visually matches the old one
+    ocrmypdf --output-type pdf temp.pdf new-ocr.pdf
+
+``--output-type pdf`` is used here to avoid using Ghostscript which will also rasterize images.
+
+
+Improving OCR quality
+---------------------
+
+The `Image processing`_ features can improve OCR quality.
+
+Rotating pages and deskewing helps to ensure that the page orientation is correct before OCR begins. Removing the background and/or cleaning the page can also improve results. The ``--oversample DPI`` argument can be specified to resample images to higher resolution before attempting OCR; this can improve results as well.
+
+OCR quality will suffer if the resolution of input images is not correct (since the range of pixel sizes that will be checked for possible fonts will also be incorrect).
+

docs/errors.rst

@@ -0,0 +1,33 @@
+Common error messages
+=====================
+
+Page already has text
+---------------------
+
+.. code::
+
+	ERROR -    1: page already has text! – aborting (use --force-ocr to force OCR)
+
+You ran ocrmypdf on a file that already contains printable text or a hidden OCR text layer (it can't quite tell the difference). You probably don't want to do this, because the file is already searchable.
+
+As the error message suggests, your options are:
+
+- ``ocrmypdf --force-ocr`` to :ref:`rasterize <raster-vector>` all vector content and run OCR on the images. This is useful if a previous OCR program failed, or if the document contains a text watermark.
+
+- ``ocrmypdf --skip-text`` to skip OCR and other processing on any pages that contain text. Text pages will be copied into the output PDF without modification.
+
+
+Input file 'filename' is not a valid PDF
+----------------------------------------
+
+OCRmyPDF passes files through qpdf, a program that fixes errors in PDFs, before it tries to work on them. In most cases this happens because the PDF is corrupt and
+truncated (incomplete file copying) and not much can be done.
+
+You can try rewriting the file with Ghostscript or pdftk:
+
+- ``gs -o output.pdf -dSAFER -sDEVICE=pdfwrite input.pdf``
+
+- ``pdftk input.pdf cat output output.pdf``
+
+Sometimes Acrobat can repair PDFs with its `Preflight tool <https://helpx.adobe.com/acrobat/using/correcting-problem-areas-preflight-tool.html>`_.
+

docs/index.rst

@@ -0,0 +1,39 @@
+.. ocrmypdf documentation master file, created by
+   sphinx-quickstart on Sun Sep  4 14:29:43 2016.
+   You can adapt this file completely to your liking, but it should at least
+   contain the root `toctree` directive.
+
+OCRmyPDF documentation
+======================
+
+OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to
+be searched.
+
+PDF is the best format for storing and exchanging scanned documents.  Unfortunately, PDFs can be difficult to modify. OCRmyPDF makes it easy to apply image processing and OCR to existing PDFs.
+
+.. toctree::
+   :maxdepth: 1
+
+   introduction
+   release_notes
+   installation
+   languages
+
+.. toctree::
+   :caption: Usage
+   :maxdepth: 2
+
+   cookbook
+   advanced
+   batch
+   security
+   errors
+
+
+Indices and tables
+==================
+
+* :ref:`genindex`
+* :ref:`modindex`
+* :ref:`search`
+

docs/installation.rst

@@ -0,0 +1,453 @@
+Installation
+============
+
+The easiest way to install OCRmyPDF to follow the steps for your operating system/platform.
+
+If you want to use the latest version of OCRmyPDF, your best bet is to install the most recent version your platform provides, and then upgrade that version by installing the Python binary wheels.
+
+.. contents:: Platform-specific steps
+    :depth: 1
+    :local:
+
+Installing on Debian and Ubuntu 16.10 or newer
+----------------------------------------------
+
+Users of Debian 9 ("stretch") or later or Ubuntu 16.10 or later may simply
+
+.. code-block:: bash
+
+    apt-get install ocrmypdf
+
+To see what versions are available, check the `Debian Package Tracker <https://tracker.debian.org/pkg/ocrmypdf>`_ or `Ubuntu launchpad.net <https://launchpad.net/ocrmypdf>`_.
+
+Installing on macOS with Homebrew
+---------------------------------
+
+.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg
+    :alt: homebrew
+    :target: http://brewformulas.org/Ocrmypdf
+
+OCRmyPDF is now a standard `Homebrew <https://brew.sh>`_ formula. To install on macOS:
+
+.. code-block:: bash
+
+    brew install ocrmypdf
+
+.. note::
+
+    Users who previously installed OCRmyPDF on macOS using ``pip install ocrmypdf`` should remove the pip version (``pip3 uninstall ocrmypdf``) before switching to the Homebrew version.
+
+.. note::
+
+    Users who previously installed OCRmyPDF from the private tap should switch to the mainline version (``brew untap jbarlow83/ocrmypdf``) and install from there.
+
+.. _Docker-install:
+
+Installing the Docker image
+---------------------------
+
+For many users, installing the Docker image will be easier than installing all of OCRmyPDF's dependencies. For Windows, it is the only option.
+
+If you have `Docker <https://docs.docker.com/>`_ installed on your system, you can install a Docker image of the latest release.
+
+Follow the Docker installation instructions for your platform.  If you can run this command successfully, your system is ready to download and execute the image:
+
+.. code-block:: bash
+
+    docker run hello-world
+
+OCRmyPDF will use all available CPU cores.  By default, the VirtualBox machine instance on Windows and macOS has only a single CPU core enabled. Use the VirtualBox Manager to determine the name of your Docker engine host, and then follow these optional steps to enable multiple CPUs:
+
+.. code-block:: bash
+
+    # Optional step for Mac OS X users
+    docker-machine stop "yourVM"
+    VBoxManage modifyvm "yourVM" --cpus 2  # or whatever number of core is desired
+    docker-machine start "yourVM"
+    eval $(docker-machine env "yourVM")
+
+Assuming you have a Docker engine running, you can download one of the three available images:
+
+.. list-table::
+    :widths: auto
+    :header-rows: 1
+
+    *   - Image name
+        - Download command
+        - Notes
+    *   - ocrmypdf
+        - ``docker pull jbarlow83/ocrmypdf``
+        - Latest ocrmypdf with Tesseract 4.0.0-beta1 on Ubuntu 18.04. Includes English, French, German, Spanish, Portugeuse and Simplified Chinese.
+    *   - ocrmypdf-polyglot
+        - ``docker pull jbarlow83/ocrmypdf-polyglot``
+        - As above, with all available language packs.
+
+For example:
+
+.. code-block:: bash
+
+    docker pull jbarlow83/ocrmypdf
+
+Then tag it to give a more convenient name, just ocrmypdf:
+
+.. code-block:: bash
+
+    docker tag jbarlow83/ocrmypdf ocrmypdf
+
+.. _docker-polyglot:
+
+The alternative "polyglot" image provides `all available language packs <https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#languages>`_.
+
+You can then run ocrmypdf using the command:
+
+.. code-block:: bash
+
+    docker run --rm ocrmypdf --help
+
+To execute the OCRmyPDF on a local file, you must `provide a writable volume to the Docker image <https://docs.docker.com/userguide/dockervolumes/>`_, and both the input and output file must be inside the writable volume.  This example command uses the current working directory as the writable volume:
+
+.. code-block:: bash
+
+    docker run --rm -v "$(pwd):/home/docker" <other docker arguments>   ocrmypdf <your arguments to ocrmypdf>
+
+In this worked example, the current working directory contains an input file called ``test.pdf`` and the output will go to ``output.pdf``:
+
+.. code-block:: bash
+
+    docker run --rm -v "$(pwd):/home/docker"   ocrmypdf --skip-text test.pdf output.pdf
+
+.. note:: The working directory should be a writable local volume or Docker may not have permission to access it.
+
+Note that ``ocrmypdf`` has its own separate ``-v VERBOSITYLEVEL`` argument to control debug verbosity. All Docker arguments should before the ``ocrmypdf`` image name and all arguments to ``ocrmypdf`` should be listed after.
+
+In some environments the permissions associated with Docker can be complex to configure. The process that executes Docker may end up not having the permissions to write the specified file system. In that case one can stream the file into and out of the Docker process and avoid all permission hassles, using ``-`` as the input and output filename:
+
+.. code-block:: bash
+
+    docker run --rm -i   ocrmypdf <other arguments to ocrmypdf> - - <input.pdf >output.pdf
+
+For convenience, a shell alias can hide the docker command:
+
+.. code-block:: bash
+
+    alias ocrmypdf='docker run --rm -v "$(pwd):/home/docker" ocrmypdf'
+    ocrmypdf --version  # runs docker version
+
+Or in the wonderful `fish shell <https://fishshell.com/>`_:
+
+.. code-block:: fish
+
+    alias ocrmypdf 'docker run --rm -v (pwd):/home/docker ocrmypdf'
+    funcsave ocrmypdf
+
+.. note::
+
+    The ocrmypdf Docker containers are designed to be used for a single OCR job. The ``docker run --rm`` argument tells Docker to delete temporary storage associated with container when it is done executing.
+
+Manual installation on macOS
+----------------------------
+
+These instructions probably work on all macOS supported by Homebrew.
+
+If it's not already present, `install Homebrew <http://brew.sh/>`_.
+
+Update Homebrew:
+
+.. code-block:: bash
+
+    brew update
+
+Install or upgrade the required Homebrew packages, if any are missing:
+
+.. code-block:: bash
+
+    brew install libpng openjpeg jbig2dec libtiff     # image libraries
+    brew install qpdf
+    brew install ghostscript
+    brew install python3
+    brew install libxml2 libffi leptonica
+    brew install unpaper   # optional
+
+Python 3.5, 3.6 and 3.7 are supported.
+
+Install the required Tesseract OCR engine with the language packs you plan to use:
+
+.. code-block:: bash
+
+    brew install tesseract                       # Option 1: for English, French, German, Spanish
+
+.. _macos-all-languages:
+
+.. code-block:: bash
+
+    brew install tesseract --with-all-languages  # Option 2: for all language packs
+
+Update the homebrew pip:
+
+.. code-block:: bash
+
+    pip3 install --upgrade pip
+
+You can then install OCRmyPDF from PyPI, for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf[fitz]
+
+or system-wide:
+
+.. code-block:: bash
+
+    pip3 install ocrmypdf[fitz]
+
+``[fitz]`` includes the optional dependency on PyMuPDF, which improves OCRmyPDF's output in many cases.
+
+The command line program should now be available:
+
+.. code-block:: bash
+
+    ocrmypdf --help
+
+Installing the latest version on Ubuntu 18.04 LTS
+-------------------------------------------------
+
+Ubuntu 18.04 includes ocrmypdf 6.1.2. To install a more recent version, first
+install the system version to get all the dependencies:
+
+.. code-block:: bash
+
+    sudo apt-get update
+    sudo apt-get install \
+        ocrmypdf \
+        python3-pip
+
+Then install ocrmypdf 6.1.5 for the local user and set the user's ``PATH`` to check for the user's Python packages.
+
+.. code-block:: bash
+
+    export PATH=$HOME/.local/bin:$PATH
+    pip3 install --user ocrmypdf[fitz]
+
+
+Installing on Ubuntu 16.04 LTS
+------------------------------
+
+No package is currently available for Ubuntu 16.04, but you can install the dependencies manually:
+
+.. code-block:: bash
+
+    sudo apt-get update
+    sudo apt-get install \
+        unpaper \
+        ghostscript \
+        tesseract-ocr \
+        qpdf \
+        python3-pip \
+        python3-cffi
+
+If you wish install OCRmyPDF for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf[fitz]
+
+Alternately, system-wide. Note that this may modify the system Python environment:
+
+.. code-block:: bash
+
+    sudo pip3 install ocrmypdf[fitz]
+
+If you wish to install OCRmyPDF to a virtual environment to isolate the system Python, you can follow these steps.
+
+.. code-block:: bash
+
+    python3 -m venv venv-ocrmypdf
+    source venv-ocrmypdf/bin/activate
+    pip3 install ocrmypdf[fitz]
+
+``[fitz]`` includes the optional dependency on PyMuPDF, which improves OCRmyPDF's output in many cases.
+
+
+Installing on Ubuntu 14.04 LTS
+------------------------------
+
+Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other options, because it is older and does not provide ``pip``.
+
+Update apt-get:
+
+.. code-block:: bash
+
+    sudo apt-get update
+
+Install system dependencies:
+
+.. code-block:: bash
+
+    sudo apt-get install \
+        software-properties-common python-software-properties \
+        zlib1g-dev \
+        libjpeg-dev \
+        libffi-dev \
+        qpdf
+
+We will need backports of Ghostscript 9.16, libav-11 (for unpaper 6.1), Tesseract 4.00 (alpha), and Python 3.6. This will replace Ghostscript and Tesseract 3.x on your system. Python 3.6 will be installed alongside the system Python 3.
+
+If you prefer to not modify your system in this matter, consider using a Docker container.
+
+.. code-block:: bash
+
+    sudo add-apt-repository ppa:vshn/ghostscript -y
+    sudo add-apt-repository ppa:heyarje/libav-11 -y
+    sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
+    sudo add-apt-repository ppa:jonathonf/python-3.6 -y
+
+    sudo apt-get update
+
+    sudo apt-get install \
+        python3.6-dev \
+        ghostscript \
+        tesseract-ocr \
+        tesseract-ocr-eng \
+        libavformat56 libavcodec56 libavutil54 \
+        wget
+
+Now we need to install ``pip`` and let it install ocrmypdf:
+
+.. code-block:: bash
+
+    curl https://bootstrap.pypa.io/ez_setup.py -o - | python3.6 && python3.6 -m easy_install pip
+    pip3.6 install ocrmypdf[fitz]
+
+The ``wget`` command will download a program and run it.
+
+These installation instructions omit the optional dependency ``unpaper``, which is only available at version 0.4.2 in Ubuntu 14.04. The author could not find a backport of ``unpaper``, and created a .deb package to do the job of installing unpaper 6.1 (for x86 64-bit only):
+
+.. code-block:: bash
+
+    wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O unpaper_6.1-1.deb
+    sudo dpkg -i unpaper_6.1-1.deb
+
+
+Installing on ArchLinux
+-----------------------
+
+The author is aware of an `ArchLinux package for ocrmypdf <https://aur.archlinux.org/packages/ocrmypdf/>`_. It seems like the following command might work.
+
+.. code-block:: bash
+
+    pacman -S ocrmypdf
+
+
+Installing on Windows
+---------------------
+
+Direct installation on Windows is not possible.  Install the _`Docker` container as described above.  Ensure that your command prompt can run the docker "hello world" container.
+
+It would probably not be too difficult to run on Windows.  The main reason this has been avoided is the difficulty of packaging and installing the various non-Python dependencies: Tesseract, QPDF, Ghostscript, Leptonica.  Pull requests to add or improve Windows support would be quite welcome.
+
+
+Running on Windows
+~~~~~~~~~~~~~~~~~~
+
+The command line syntax to run ocrmypdf from a command prompt will resemble:
+
+.. code-block:: bat
+
+    docker run -v /c/Users/sampleuser:/home/docker ocrmypdf --skip-text test.pdf output.pdf
+
+where /c/Users/sampleuser is a Unix representation of the Windows path C:\\Users\\sampleuser, assuming a user named "sampleuser" is running ocrmypdf on a file in their home directory, and the files "test.pdf" and "output.pdf" are in the sampleuser folder. The Windows user must have read and write permissions.
+
+`Bash on Ubuntu on Windows <https://github.com/Microsoft/BashOnWindows>`_ should also be a viable route for running the OCRmyPDF Docker container.
+
+
+Installing with Python pip
+--------------------------
+
+First, install `your platform's version <https://repology.org/metapackage/ocrmypdf/versions>`_ of ``ocrmypdf``, if available, as a way of ensuring that external dependencies are (mostly) satisified, even though the platform version may be out of date. Use ``ocrmypdf --version`` to confirm what version was installed.
+
+Then you can install the latest OCRmyPDF from the Python wheels. First try:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf
+
+You should then be able to run ``ocrmypdf --version`` and see that the latest version was located.
+
+Since ``pip3 install --user`` does not work correctly on some platforms, notably Ubuntu 16.04 and older, and the Homebrew version of Python, instead use this for a system wide installation:
+
+.. code-block:: bash
+
+    pip3 install ocrmypdf
+
+Requirements for pip and HEAD install
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+OCRmyPDF currently requires these external programs to be installed:
+
+- Python 3.5 or newer
+- Tesseract 3.04 or newer
+- Ghostscript 9.15 or newer
+- qpdf 7.0.0 or newer
+
+The following dependencies are recommended:
+
+- Python 3.6
+- Tesseract 4.00 or newer
+- Ghostscript 9.22 or newer
+- qpdf 8.0.2 or newer
+- unpaper 6.1
+- PyMuPDF 1.12.5 or newer
+
+These are in addition to the Python packaging dependencies, meaning that unfortunately, the ``pip install`` command cannot satisfy all of them.
+
+Python 3.6 and Tesseract 4.0.0-beta.1 are recommended for best OCR results and best performance.
+
+The library PyMuPDF is not widely available in platform distributions, and it improves OCRmyPDF in certain conditions. Consider installing OCRmyPDF from the Python binary wheels, which include a precompiled version of this library.
+
+
+Installing HEAD revision from sources
+-------------------------------------
+
+If you have ``git`` and Python 3.5 or newer installed, you can install from source. When the ``pip`` installer runs, it will alert you if dependencies are missing.
+
+To install the HEAD revision from sources in the current Python 3 environment:
+
+.. code-block:: bash
+
+    pip3 install git+https://github.com/jbarlow83/OCRmyPDF.git
+
+Or, to install in `development mode <https://pythonhosted.org/setuptools/setuptools.html#development-mode>`_,  allowing customization of OCRmyPDF, use the ``-e`` flag:
+
+.. code-block:: bash
+
+    pip3 install -e git+https://github.com/jbarlow83/OCRmyPDF.git
+
+You may find it easiest to install in a virtual environment, rather than system-wide:
+
+.. code-block:: bash
+
+    git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
+    python3 -m venv
+    source venv/bin/activate
+    cd OCRmyPDF
+    pip3 install .
+
+However, ``ocrmypdf`` will only be accessible on the system PATH after
+you activate the virtual environment.
+
+To run the program:
+
+.. code-block:: bash
+
+    ocrmypdf --help
+
+If not yet installed, the script will notify you about dependencies that
+need to be installed. The script requires specific versions of the
+dependencies. Older version than the ones mentioned in the release notes
+are likely not to be compatible to OCRmyPDF.
+
+
+Other Linux packages
+--------------------
+
+See the `Repology <https://repology.org/metapackage/ocrmypdf/versions>`_ page.

docs/introduction.rst

@@ -0,0 +1,115 @@
+Introduction
+============
+
+OCRmyPDF is a Python 3 package that adds OCR layers to PDFs.
+
+
+About OCR
+---------
+
+`Optical character recognition <https://en.wikipedia.org/wiki/Optical_character_recognition>`_ is technology that converts images of typed or handwritten text, such as in a scanned document, to computer text that can be searched and copied.
+
+OCRmyPDF uses `Tesseract <https://github.com/tesseract-ocr/tesseract>`_, the best available open source OCR engine, to perform OCR.
+
+.. _raster-vector:
+
+About PDFs
+----------
+
+PDFs are page description files that attempts to preserve a layout exactly. They  contain `vector graphics <http://vector-conversions.com/vectorizing/raster_vs_vector.html>`_ that can contain raster objects such as scanned images. Because PDFs can contain multiple pages (unlike many image formats) and can contain fonts and text, it is a good formats for exchanging scanned documents.
+
+.. image:: bitmap_vs_svg.svg
+
+A PDF page might contain multiple images, even if it only appears to have one image.  Some scanners or scanning software will segment pages into monochromatic text and color regions for example, to improve the compression ratio and appearance of the page.
+
+Rasterizing a PDF is the process of generating an image suitable for display or analyzing with an OCR engine.  OCR engines like Tesseract work with images, not vector objects.
+
+
+About PDF/A
+-----------
+
+`PDF/A <https://en.wikipedia.org/wiki/PDF/A>`_ is an ISO-standardized subset of the full PDF specification that is designed for archiving (the 'A' stands for Archive).  PDF/A differs from PDF primarily by omitting features that would make it difficult to read the file in the future, such as embedded Javascript, video, audio and references to external fonts.  All fonts and resources needed to interpret the PDF must be contained within it. Because PDF/A disables Javascript and other types of embedded content, it is probably more secure.
+
+There are various conformance levels and versions, such as "PDF/A-2b".
+
+Generally speaking, the best format for scanned documents is PDF/A. Some governments and jurisdictions, US Courts in particular, `mandate the use of PDF/A <https://pdfblog.com/2012/02/13/what-is-pdfa/>`_ for scanned documents.
+
+Since most people who scan documents are interested in reading them indefinitely into the future, OCRmyPDF generates PDF/A-2b by default.
+
+PDF/A has a few drawbacks.  Some PDF viewers include an alert that the file is a PDF/A, which may confuse some users.  It also tends to produce larger files than PDF, because it embeds certain resources even if they are commonly available. PDF/A files can be digitally signed, but may not be encrypted, to ensure they can be read in the future.  Fortunately, converting from PDF/A to a regular PDF is trivial, and any PDF viewer can view PDF/A.
+
+
+What OCRmyPDF does
+------------------
+
+OCRmyPDF analyzes each page of a PDF to determine the colorspace and resolution (DPI) needed to capture all of the information on that page without losing content.  It uses `Ghostscript <http://ghostscript.com/>`_ to rasterize the page, and then performs on OCR on the rasterized image to create an OCR "layer". The layer is then grafted back onto the original PDF.
+
+While one can use a program like Ghostscript or ImageMagick to get an image and put the image through Tesseract, that actually creates a new PDF and many details may be lost. OCRmyPDF can produce a minimally changed PDF as output.
+
+OCRmyPDF also some image processing options like deskew which improve the appearance of files and quality of OCR. When these are used, the OCR layer is grafted onto the processed image instead.
+
+By default, OCRmyPDF produces archival PDFs – PDF/A, which are a stricter subset of PDF features designed for long term archives. If regular PDFs are desired, this can be disabled with ``--output-type pdf``.
+
+
+Why you shouldn't do this manually
+----------------------------------
+
+A PDF is similar to an HTML file, in that it contains document structure along with images.  Sometimes a PDF does nothing more than present a full page image, but often there is additional content that would be lost.
+
+A manual process could work like either of these:
+
+1. Rasterize each page as an image, OCR the images, and combine the output into a PDF. This preserves the layout of each page, but resamples all images (possibly losing quality, increasing file size, introducing compression artifacts, etc.).
+
+2. Extract each image, OCR, and combine the output into a PDF. This loses the context in which images are used in the PDF, meaning that cropping, rotation and scaling of pages may be lost. Some scanned PDFs use multiple images segmented into black and white, grayscale and color regions, with stencil masks to prevent overlap, as this can enhance the appearance of a file while reducing file size. Clearly, reassembling these images will be easy. This also loses and text or vector art on any pages in a PDF with both scanned and pure digital content.
+
+In the case of a PDF that is nothing other than a container of images (no rotation, scaling, cropping, one image per page), the second approach can be lossless.
+
+OCRmyPDF uses several strategies depending on input options and the input PDF itself, but generally speaking it rasterizes a page for OCR and then grafts the OCR back onto the original. As such it can handle complex PDFs and still preserve their contents as much as possible.
+
+
+Limitations
+-----------
+
+OCRmyPDF is limited by the Tesseract OCR engine.  As such it experiences these limitations, as do any other programs that rely on Tesseract:
+
+* The OCR is not as accurate as commercial solutions such as Abbyy.
+* It is not capable of recognizing handwriting.
+* It may find gibberish and report this as OCR output.
+* If a document contains languages outside of those given in the ``-l LANG`` arguments, results may be poor.
+* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns and join text across the columns.
+* Poor quality scans may produce poor quality OCR. Garbage in, garbage out.
+* PDFs that use transparent layers are not currently checked in the test suite, so they may not work correctly.
+* It does not expose information about what font family text belongs to.
+
+OCRmyPDF is also limited by the PDF specification:
+
+* PDF encodes the position of text glyphs but does not encode document structure.  There is no markup that divides a document in sections, paragraphs, sentences, or even words (since blank spaces are not represented). As such all elements of document structure including the spaces between words must be derived heuristically.  Some PDF viewers do a better job of this than others.
+* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround. While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones.
+
+Ghostscript also imposes some limitations:
+
+* PDFs containing JBIG2-encoded content will be converted to CCITT Group4 encoding, which has lower compression ratios, if Ghostscript PDF/A is enabled.
+* PDFs containing JPEG 2000-encoded content will be converted to JPEG encoding, which may introduce compression artifacts, if Ghostscript PDF/A is enabled.
+* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format.
+  
+OCRmyPDF is currently not designed to be used as a Python API; it is designed to be run as a command line tool. ``import ocrmypf`` currently attempts to process the command line on ``sys.argv`` at import time so it has side effects that will interfere with its use as a package. The API it presents should not be considered stable.
+
+
+Similar programs
+----------------
+
+To the author's knowledge, OCRmyPDF is the most feature-rich and thoroughly tested command line OCR PDF conversion tool. If it does not meet your needs, contributions and suggestions are welcome. If not, consider one of these similar open source programs:
+
+* pdf2pdfocr
+* pdfsandwich
+* pypdfocr
+* pdfbeads
+
+Web front-ends
+--------------
+
+* `Nextcloud OCR <https://github.com/janis91/ocr>`_ is a free software plugin for the Nextcloud private cloud software
+* `OCRmyPDF-web <https://github.com/sseemayer/OCRmyPDF-web>`_, a micro web-frontend for OCRmyPDF (third-party, not actively maintained)
+
+Bear in mind that OCRmyPDF is not designed to be secure against malware-bearing PDFs (see `Using OCRmyPDF online`_).
+

docs/languages.rst

@@ -0,0 +1,76 @@
+.. _lang-packs:
+
+Installing additional language packs
+====================================
+
+OCRmyPDF uses Tesseract for OCR, and relies on its language packs for languages other than English. 
+
+Tesseract supports `most languages <https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#languages>`_.
+
+For Linux users, you can often find packages that provide language packs:
+
+Debian and Ubuntu users
+-----------------------
+
+.. code-block:: bash
+
+   # Display a list of all Tesseract language packs
+   apt-cache search tesseract-ocr
+
+   # Debian/Ubuntu users
+   apt-get install tesseract-ocr-chi-sim  # Example: Install Chinese Simplified language back
+   
+You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple
+languages can be requested using either ``-l eng+fre`` (English and French) or ``-l eng -l fre``.
+
+macOS users
+-----------
+
+You can install additional language packs by :ref:`installing Tesseract using Homebrew with all language packs <macos-all-languages>`.
+
+Docker users
+------------
+
+Users of the Docker image may use the alternative :ref:`"polyglot" container <docker-polyglot>` which includes all languages.
+
+Adding individual language packs to a Docker image
+""""""""""""""""""""""""""""""""""""""""""""""""""
+
+If you wish to add a single language pack, you could do the following:
+
+* Download the desired ``.trainedata`` file from the `tessdata <https://github.com/tesseract-ocr/tessdata>`_ repository. Let's use Hebrew in this example (``heb.traineddata``)
+
+* Copy the file to ``/home/user/downloads/heb.traineddata``.
+
+* Create a new container based on the ocrmypdf-tess4 image and jump into it with a terminal:
+
+.. code-block:: bash
+
+	host$ docker run  -v /home/user/downloads:/home/docker -it --entrypoint /bin/bash ocrmypdf-tess4
+
+* Put the file where Tesseract expects it:
+
+.. code-block:: bash
+
+	docker$ cp /home/docker/heb.traineddata /usr/share/tesseract-ocr/tessdata
+
+* Note the container id, and save it as a new image (in this example, ``ocrmypdf-tess4-heb``)
+
+.. code-block:: bash
+
+    host$ docker commit <container_id> ocrmypdf-tess4-heb
+
+
+Known limitations
+-----------------
+
+As of v4.2, users of ocrmypdf working languages outside the Latin alphabet should use the following syntax:
+
+.. code-block:: bash
+
+	ocrmypdf -l eng+gre --output-type pdf --pdf-renderer tesseract
+
+The reasons for this are:
+
+* The latest version of Ghostscript (9.19 as of this writing) has unfixed bugs in Unicode handling that generate invalid character maps, so Ghostscript cannot be used for PDF/A conversion
+* The default "hocr" PDF renderer does not handle Asian fonts properly

docs/pipeline.svg

@@ -0,0 +1,553 @@
+<?xml version="1.0" encoding="UTF-8" standalone="no"?>
+<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN"
+ "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
+<!-- Generated by graphviz version 2.40.1 (20161225.0304)
+ -->
+<!-- Title: Pipeline: Pages: 1 -->
+<svg width="1952pt" height="1277pt"
+ viewBox="0.00 0.00 1952.00 1277.18" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
+<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 1273.1825)">
+<title>Pipeline:</title>
+<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-1273.1825 1948,-1273.1825 1948,4 -4,4"/>
+<g id="clust1" class="cluster">
+<title>clustertasks</title>
+<polygon fill="none" stroke="#000000" points="8,-8 8,-1261.1825 1936,-1261.1825 1936,-8 8,-8"/>
+<text text-anchor="middle" x="972" y="-1233.1825" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
+</g>
+<!-- t0 -->
+<g id="node1" class="node">
+<title>t0</title>
+<polygon fill="#efa03b" stroke="#000000" points="876.7053,-1215.1825 667.2947,-1215.1825 663.2947,-1211.1825 663.2947,-1179.1825 872.7053,-1179.1825 876.7053,-1183.1825 876.7053,-1215.1825"/>
+<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 663.2947,-1211.1825 "/>
+<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 872.7053,-1179.1825 "/>
+<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 876.7053,-1215.1825 "/>
+<text text-anchor="middle" x="770" y="-1191.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.triage</text>
+</g>
+<!-- t1 -->
+<g id="node2" class="node">
+<title>t1</title>
+<polygon fill="#efa03b" stroke="#000000" points="941.0433,-1157.1825 602.9567,-1157.1825 598.9567,-1153.1825 598.9567,-1121.1825 937.0433,-1121.1825 941.0433,-1125.1825 941.0433,-1157.1825"/>
+<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 598.9567,-1153.1825 "/>
+<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 937.0433,-1121.1825 "/>
+<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 941.0433,-1157.1825 "/>
+<text text-anchor="middle" x="770" y="-1133.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.repair_and_parse_pdf</text>
+</g>
+<!-- t0&#45;&gt;t1 -->
+<g id="edge1" class="edge">
+<title>t0&#45;&gt;t1</title>
+<path fill="none" stroke="#0044a0" d="M770,-1179.1488C770,-1175.4743 770,-1171.5242 770,-1167.5888"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="773.5001,-1167.4047 770,-1157.4047 766.5001,-1167.4047 773.5001,-1167.4047"/>
+</g>
+<!-- t2 -->
+<g id="node3" class="node">
+<title>t2</title>
+<polygon fill="#efa03b" stroke="#000000" points="1043.9447,-1078.2585 787,-1095.158 530.0553,-1078.2585 530.2951,-1050.9147 1043.7049,-1050.9147 1043.9447,-1078.2585"/>
+<polygon fill="none" stroke="#000000" points="1047.9814,-1081.9991 787,-1099.164 526.0186,-1081.9991 526.3263,-1046.9173 1047.6737,-1046.9173 1047.9814,-1081.9991"/>
+<text text-anchor="middle" x="787" y="-1064.7009" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.pre_split_pages</text>
+</g>
+<!-- t1&#45;&gt;t2 -->
+<g id="edge2" class="edge">
+<title>t1&#45;&gt;t2</title>
+<path fill="none" stroke="#0044a0" d="M774.5557,-1120.8306C775.4962,-1117.0421 776.5194,-1112.92 777.5618,-1108.7212"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="780.986,-1109.4542 779.9984,-1098.9055 774.1922,-1107.7677 780.986,-1109.4542"/>
+</g>
+<!-- t20 -->
+<g id="node22" class="node">
+<title>t20</title>
+<polygon fill="#efa03b" stroke="#000000" points="649.5156,-1020.2193 286.4844,-1020.2193 282.4844,-1016.2193 282.4844,-984.2193 645.5156,-984.2193 649.5156,-988.2193 649.5156,-1020.2193"/>
+<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 282.4844,-1016.2193 "/>
+<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 645.5156,-984.2193 "/>
+<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 649.5156,-1020.2193 "/>
+<text text-anchor="middle" x="466" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.generate_postscript_stub</text>
+</g>
+<!-- t1&#45;&gt;t20 -->
+<g id="edge31" class="edge">
+<title>t1&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M598.785,-1123.5822C563.105,-1117.7214 531.8209,-1109.8012 517,-1099.1825 493.9382,-1082.6593 480.4369,-1052.4741 473.1842,-1030.1561"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="476.5199,-1029.095 470.2797,-1020.5329 469.8185,-1031.1177 476.5199,-1029.095"/>
+</g>
+<!-- t24 -->
+<g id="node24" class="node">
+<title>t24</title>
+<polygon fill="#efa03b" stroke="#000000" points="20.4213,-156.4247 242,-129.112 463.5787,-156.4247 463.3719,-200.6177 20.6281,-200.6177 20.4213,-156.4247"/>
+<polygon fill="none" stroke="#000000" points="16.4083,-152.8864 242,-125.079 467.5917,-152.8864 467.3496,-204.6205 16.6504,-204.6205 16.4083,-152.8864"/>
+<text text-anchor="middle" x="242" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_mupdf</text>
+</g>
+<!-- t1&#45;&gt;t24 -->
+<g id="edge41" class="edge">
+<title>t1&#45;&gt;t24</title>
+<path fill="none" stroke="#0044a0" d="M598.6617,-1124.8567C376.4578,-1103.9215 17,-1060.9746 17,-1002.2193 17,-1002.2193 17,-1002.2193 17,-335.7123 17,-276.266 68.2524,-235.5065 121.8346,-209.1377"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="123.6227,-212.1629 131.1479,-204.7049 120.6143,-205.8423 123.6227,-212.1629"/>
+</g>
+<!-- t23 -->
+<g id="node25" class="node">
+<title>t23</title>
+<polygon fill="#efa03b" stroke="#000000" points="1006.4902,-156.4239 1218,-129.1092 1429.5098,-156.4239 1429.3124,-200.6199 1006.6876,-200.6199 1006.4902,-156.4239"/>
+<polygon fill="none" stroke="#000000" points="1002.4742,-152.9063 1218,-125.073 1433.5258,-152.9063 1433.2948,-204.623 1002.7052,-204.623 1002.4742,-152.9063"/>
+<text text-anchor="middle" x="1218" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_qpdf</text>
+</g>
+<!-- t1&#45;&gt;t23 -->
+<g id="edge46" class="edge">
+<title>t1&#45;&gt;t23</title>
+<path fill="none" stroke="#0044a0" d="M941.3419,-1136.2202C1263.4143,-1128.3561 1927,-1100.2573 1927,-1002.2193 1927,-1002.2193 1927,-1002.2193 1927,-335.7123 1927,-234.7734 1651.3069,-194.7548 1443.7522,-178.9295"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1443.8443,-175.4267 1433.6111,-178.1712 1443.3223,-182.4072 1443.8443,-175.4267"/>
+</g>
+<!-- t3 -->
+<g id="node4" class="node">
+<title>t3</title>
+<polygon fill="#efa03b" stroke="#000000" points="1055.3851,-1020.2193 808.6149,-1020.2193 804.6149,-1016.2193 804.6149,-984.2193 1051.3851,-984.2193 1055.3851,-988.2193 1055.3851,-1020.2193"/>
+<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 804.6149,-1016.2193 "/>
+<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1051.3851,-984.2193 "/>
+<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1055.3851,-1020.2193 "/>
+<text text-anchor="middle" x="930" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.split_page</text>
+</g>
+<!-- t2&#45;&gt;t3 -->
+<g id="edge3" class="edge">
+<title>t2&#45;&gt;t3</title>
+<path fill="none" stroke="#0044a0" d="M837.0145,-1046.7494C851.946,-1039.5987 868.2558,-1031.7881 882.9879,-1024.733"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="884.9035,-1027.6964 892.4109,-1020.2204 881.88,-1021.383 884.9035,-1027.6964"/>
+</g>
+<!-- t4 -->
+<g id="node5" class="node">
+<title>t4</title>
+<polygon fill="#efa03b" stroke="#000000" points="1194.5194,-941.2939 964,-958.1945 733.4806,-941.2939 733.6958,-913.948 1194.3042,-913.948 1194.5194,-941.2939"/>
+<polygon fill="none" stroke="#000000" points="1198.5491,-945.0069 964,-962.203 729.4509,-945.0069 729.7267,-909.9503 1198.2733,-909.9503 1198.5491,-945.0069"/>
+<text text-anchor="middle" x="964" y="-927.7357" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_or_skip</text>
+</g>
+<!-- t3&#45;&gt;t4 -->
+<g id="edge4" class="edge">
+<title>t3&#45;&gt;t4</title>
+<path fill="none" stroke="#0044a0" d="M939.1114,-983.867C941.197,-979.6659 943.4863,-975.0549 945.8053,-970.3839"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="948.9439,-971.9327 950.2559,-961.4194 942.6741,-968.8199 948.9439,-971.9327"/>
+</g>
+<!-- t5 -->
+<g id="node6" class="node">
+<title>t5</title>
+<polygon fill="#efa03b" stroke="#000000" points="1119.3166,-883.252 812.6834,-883.252 808.6834,-879.252 808.6834,-847.252 1115.3166,-847.252 1119.3166,-851.252 1119.3166,-883.252"/>
+<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 808.6834,-879.252 "/>
+<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1115.3166,-847.252 "/>
+<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1119.3166,-883.252 "/>
+<text text-anchor="middle" x="964" y="-859.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_preview</text>
+</g>
+<!-- t4&#45;&gt;t5 -->
+<g id="edge5" class="edge">
+<title>t4&#45;&gt;t5</title>
+<path fill="none" stroke="#0044a0" d="M964,-909.7834C964,-904.5184 964,-898.8955 964,-893.4799"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="967.5001,-893.4695 964,-883.4695 960.5001,-893.4695 967.5001,-893.4695"/>
+</g>
+<!-- t6 -->
+<g id="node7" class="node">
+<title>t6</title>
+<polygon fill="#efa03b" stroke="#000000" points="1269.0824,-825.252 1010.9176,-825.252 1006.9176,-821.252 1006.9176,-789.252 1265.0824,-789.252 1269.0824,-793.252 1269.0824,-825.252"/>
+<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1006.9176,-821.252 "/>
+<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1265.0824,-789.252 "/>
+<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1269.0824,-825.252 "/>
+<text text-anchor="middle" x="1138" y="-801.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.orient_page</text>
+</g>
+<!-- t4&#45;&gt;t6 -->
+<g id="edge7" class="edge">
+<title>t4&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M1072.6871,-909.8532C1097.768,-902.2558 1119.5887,-893.2079 1128,-883.252 1138.9559,-870.2843 1141.4763,-851.3696 1141.2475,-835.6545"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1144.7291,-835.2052 1140.7199,-825.3983 1137.7384,-835.5649 1144.7291,-835.2052"/>
+</g>
+<!-- t5&#45;&gt;t6 -->
+<g id="edge6" class="edge">
+<title>t5&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M1018.1009,-847.2184C1035.9411,-841.2716 1055.9459,-834.6034 1074.4267,-828.4431"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1075.5526,-831.7572 1083.9326,-825.2745 1073.3389,-825.1165 1075.5526,-831.7572"/>
+</g>
+<!-- t7 -->
+<g id="node8" class="node">
+<title>t7</title>
+<polygon fill="#efa03b" stroke="#000000" points="917.7325,-767.252 542.2675,-767.252 538.2675,-763.252 538.2675,-731.252 913.7325,-731.252 917.7325,-735.252 917.7325,-767.252"/>
+<polyline fill="none" stroke="#000000" points="913.7325,-763.252 538.2675,-763.252 "/>
+<polyline fill="none" stroke="#000000" points="913.7325,-763.252 913.7325,-731.252 "/>
+<polyline fill="none" stroke="#000000" points="913.7325,-763.252 917.7325,-767.252 "/>
+<text text-anchor="middle" x="728" y="-743.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
+</g>
+<!-- t6&#45;&gt;t7 -->
+<g id="edge8" class="edge">
+<title>t6&#45;&gt;t7</title>
+<path fill="none" stroke="#0044a0" d="M1010.5209,-789.2184C964.4341,-782.6988 912.2171,-775.312 865.3222,-768.6781"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="865.792,-765.2098 855.4003,-767.2745 864.8114,-772.1408 865.792,-765.2098"/>
+</g>
+<!-- t17 -->
+<g id="node15" class="node">
+<title>t17</title>
+<polygon fill="#ff69b4" stroke="#000000" points="1224.2696,-447.252 843.7304,-447.252 839.7304,-443.252 839.7304,-411.252 1220.2696,-411.252 1224.2696,-415.252 1224.2696,-447.252"/>
+<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 839.7304,-443.252 "/>
+<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1220.2696,-411.252 "/>
+<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1224.2696,-447.252 "/>
+<text text-anchor="middle" x="1032" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
+</g>
+<!-- t6&#45;&gt;t17 -->
+<g id="edge16" class="edge">
+<title>t6&#45;&gt;t17</title>
+<path fill="none" stroke="#0044a0" d="M1158.4397,-788.9107C1179.7261,-767.6923 1210,-730.5987 1210,-691.252 1210,-691.252 1210,-691.252 1210,-575.252 1210,-532.3545 1227.3937,-512.2638 1200,-479.252 1190.5284,-467.8379 1163.9108,-457.8666 1134.5783,-449.8944"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1135.1079,-446.4151 1124.5475,-447.2681 1133.3349,-453.1869 1135.1079,-446.4151"/>
+</g>
+<!-- t14 -->
+<g id="node17" class="node">
+<title>t14</title>
+<polygon fill="#00cc66" stroke="#000000" points="1565,-457.252 1242.4141,-429.252 1565,-401.252 1887.5859,-429.252 1565,-457.252"/>
+<text text-anchor="middle" x="1565" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_image_layer</text>
+</g>
+<!-- t6&#45;&gt;t14 -->
+<g id="edge22" class="edge">
+<title>t6&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M1211.6459,-789.1984C1287.5889,-768.3738 1395,-731.7602 1395,-691.252 1395,-691.252 1395,-691.252 1395,-575.252 1395,-519.4485 1448.3697,-480.3436 1495.3026,-456.7437"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1497.1229,-459.7505 1504.5857,-452.2301 1494.0619,-453.4552 1497.1229,-459.7505"/>
+</g>
+<!-- t21 -->
+<g id="node20" class="node">
+<title>t21</title>
+<polygon fill="#efa03b" stroke="#000000" points="1414.7716,-270.1725 1169.2284,-270.1725 1165.2284,-266.1725 1165.2284,-234.1725 1410.7716,-234.1725 1414.7716,-238.1725 1414.7716,-270.1725"/>
+<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1165.2284,-266.1725 "/>
+<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1410.7716,-234.1725 "/>
+<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1414.7716,-270.1725 "/>
+<text text-anchor="middle" x="1290" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.skip_page</text>
+</g>
+<!-- t6&#45;&gt;t21 -->
+<g id="edge28" class="edge">
+<title>t6&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M1269.322,-800.6326C1488.9576,-787.7851 1907,-754.9633 1907,-691.252 1907,-691.252 1907,-691.252 1907,-429.252 1907,-329.7211 1606.6136,-282.8737 1425.0278,-263.647"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1425.1684,-260.1428 1414.8592,-262.586 1424.4419,-267.105 1425.1684,-260.1428"/>
+</g>
+<!-- t19 -->
+<g id="node21" class="node">
+<title>t19</title>
+<polygon fill="#66ccff" stroke="#000000" points="465.6739,-447.252 60.3261,-447.252 56.3261,-443.252 56.3261,-411.252 461.6739,-411.252 465.6739,-415.252 465.6739,-447.252"/>
+<polyline fill="none" stroke="#000000" points="461.6739,-443.252 56.3261,-443.252 "/>
+<polyline fill="none" stroke="#000000" points="461.6739,-443.252 461.6739,-411.252 "/>
+<polyline fill="none" stroke="#000000" points="461.6739,-443.252 465.6739,-447.252 "/>
+<text text-anchor="middle" x="261" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_and_render_pdf</text>
+</g>
+<!-- t6&#45;&gt;t19 -->
+<g id="edge30" class="edge">
+<title>t6&#45;&gt;t19</title>
+<path fill="none" stroke="#0044a0" d="M1006.6775,-802.4744C883.1032,-797.0687 693.0454,-786.4114 529,-767.252 327.1018,-743.6717 78,-894.5226 78,-691.252 78,-691.252 78,-691.252 78,-575.252 78,-532.3545 60.4757,-512.1551 88,-479.252 97.684,-467.6756 124.6714,-457.7191 154.5327,-449.8074"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="155.4872,-453.176 164.3106,-447.311 153.7556,-446.3936 155.4872,-453.176"/>
+</g>
+<!-- t8 -->
+<g id="node9" class="node">
+<title>t8</title>
+<polygon fill="#efa03b" stroke="#000000" points="942.0801,-709.252 517.9199,-709.252 513.9199,-705.252 513.9199,-673.252 938.0801,-673.252 942.0801,-677.252 942.0801,-709.252"/>
+<polyline fill="none" stroke="#000000" points="938.0801,-705.252 513.9199,-705.252 "/>
+<polyline fill="none" stroke="#000000" points="938.0801,-705.252 938.0801,-673.252 "/>
+<polyline fill="none" stroke="#000000" points="938.0801,-705.252 942.0801,-709.252 "/>
+<text text-anchor="middle" x="728" y="-685.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_remove_background</text>
+</g>
+<!-- t7&#45;&gt;t8 -->
+<g id="edge9" class="edge">
+<title>t7&#45;&gt;t8</title>
+<path fill="none" stroke="#0044a0" d="M728,-731.2184C728,-727.5438 728,-723.5938 728,-719.6583"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="731.5001,-719.4742 728,-709.4743 724.5001,-719.4743 731.5001,-719.4742"/>
+</g>
+<!-- t13 -->
+<g id="node16" class="node">
+<title>t13</title>
+<polygon fill="#efa03b" stroke="#000000" points="481,-535.252 97.1874,-507.252 481,-479.252 864.8126,-507.252 481,-535.252"/>
+<text text-anchor="middle" x="481" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_visible_page_image</text>
+</g>
+<!-- t7&#45;&gt;t13 -->
+<g id="edge20" class="edge">
+<title>t7&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M592.7552,-731.2177C555.8805,-724.9263 521.5464,-717.3855 505,-709.252 470.5499,-692.3178 456.557,-686.3456 441,-651.252 425.2764,-615.7827 441.5685,-572.6191 457.8398,-542.8513"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="460.9365,-544.4864 462.8533,-534.0664 454.8568,-541.0167 460.9365,-544.4864"/>
+</g>
+<!-- t9 -->
+<g id="node10" class="node">
+<title>t9</title>
+<polygon fill="#efa03b" stroke="#000000" points="773.6759,-651.252 454.3241,-651.252 450.3241,-647.252 450.3241,-615.252 769.6759,-615.252 773.6759,-619.252 773.6759,-651.252"/>
+<polyline fill="none" stroke="#000000" points="769.6759,-647.252 450.3241,-647.252 "/>
+<polyline fill="none" stroke="#000000" points="769.6759,-647.252 769.6759,-615.252 "/>
+<polyline fill="none" stroke="#000000" points="769.6759,-647.252 773.6759,-651.252 "/>
+<text text-anchor="middle" x="612" y="-627.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_deskew</text>
+</g>
+<!-- t8&#45;&gt;t9 -->
+<g id="edge10" class="edge">
+<title>t8&#45;&gt;t9</title>
+<path fill="none" stroke="#0044a0" d="M691.9327,-673.2184C680.9122,-667.7081 668.6526,-661.5783 657.1104,-655.8072"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="658.5545,-652.6162 648.045,-651.2745 655.424,-658.8772 658.5545,-652.6162"/>
+</g>
+<!-- t8&#45;&gt;t13 -->
+<g id="edge19" class="edge">
+<title>t8&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M763.3121,-673.1037C771.2083,-667.2594 778.5531,-659.9847 783,-651.252 790.2604,-636.9942 792.214,-628.3326 783,-615.252 759.3602,-581.6921 662.7714,-550.9641 585.1057,-530.9385"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="585.806,-527.5052 575.2515,-528.4283 584.078,-534.2886 585.806,-527.5052"/>
+</g>
+<!-- t10 -->
+<g id="node11" class="node">
+<title>t10</title>
+<polygon fill="#efa03b" stroke="#000000" points="1117.8828,-593.252 816.1172,-593.252 812.1172,-589.252 812.1172,-557.252 1113.8828,-557.252 1117.8828,-561.252 1117.8828,-593.252"/>
+<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 812.1172,-589.252 "/>
+<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1113.8828,-557.252 "/>
+<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1117.8828,-593.252 "/>
+<text text-anchor="middle" x="965" y="-569.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_clean</text>
+</g>
+<!-- t9&#45;&gt;t10 -->
+<g id="edge11" class="edge">
+<title>t9&#45;&gt;t10</title>
+<path fill="none" stroke="#0044a0" d="M721.7564,-615.2184C760.9379,-608.7806 805.2658,-601.4973 845.2464,-594.9283"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="846.0113,-598.3496 855.3115,-593.2745 844.8763,-591.4422 846.0113,-598.3496"/>
+</g>
+<!-- t9&#45;&gt;t13 -->
+<g id="edge18" class="edge">
+<title>t9&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M593.2309,-615.1993C573.0988,-595.8356 540.6771,-564.6514 515.8453,-540.7673"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="518.0381,-538.0203 508.4046,-533.6106 513.1856,-543.0654 518.0381,-538.0203"/>
+</g>
+<!-- t11 -->
+<g id="node12" class="node">
+<title>t11</title>
+<polygon fill="#efa03b" stroke="#000000" points="1191.1094,-525.252 886.8906,-525.252 882.8906,-521.252 882.8906,-489.252 1187.1094,-489.252 1191.1094,-493.252 1191.1094,-525.252"/>
+<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 882.8906,-521.252 "/>
+<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1187.1094,-489.252 "/>
+<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1191.1094,-525.252 "/>
+<text text-anchor="middle" x="1037" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_ocr_image</text>
+</g>
+<!-- t10&#45;&gt;t11 -->
+<g id="edge12" class="edge">
+<title>t10&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M984.2947,-557.0293C992.2846,-549.4832 1001.6851,-540.6049 1010.318,-532.4517"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1012.7851,-534.9359 1017.6521,-525.5251 1007.9787,-529.8468 1012.7851,-534.9359"/>
+</g>
+<!-- t10&#45;&gt;t13 -->
+<g id="edge17" class="edge">
+<title>t10&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M836.5676,-557.2078C770.772,-547.9638 690.4915,-536.6847 622.8731,-527.1846"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="622.9531,-523.6616 612.5634,-525.7361 621.9792,-530.5935 622.9531,-523.6616"/>
+</g>
+<!-- t12 -->
+<g id="node13" class="node">
+<title>t12</title>
+<polygon fill="#00cc66" stroke="#000000" points="822.0996,-447.252 507.9004,-447.252 503.9004,-443.252 503.9004,-411.252 818.0996,-411.252 822.0996,-415.252 822.0996,-447.252"/>
+<polyline fill="none" stroke="#000000" points="818.0996,-443.252 503.9004,-443.252 "/>
+<polyline fill="none" stroke="#000000" points="818.0996,-443.252 818.0996,-411.252 "/>
+<polyline fill="none" stroke="#000000" points="818.0996,-443.252 822.0996,-447.252 "/>
+<text text-anchor="middle" x="663" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
+</g>
+<!-- t11&#45;&gt;t12 -->
+<g id="edge13" class="edge">
+<title>t11&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M950.2568,-489.1612C893.2425,-477.2705 818.5749,-461.6981 759.6569,-449.4104"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="760.1,-445.9276 749.5961,-447.3122 758.6708,-452.7802 760.1,-445.9276"/>
+</g>
+<!-- t11&#45;&gt;t17 -->
+<g id="edge15" class="edge">
+<title>t11&#45;&gt;t17</title>
+<path fill="none" stroke="#0044a0" d="M1035.8403,-489.1612C1035.2388,-479.7779 1034.4904,-468.1019 1033.8144,-457.5566"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1037.3033,-457.2705 1033.1707,-447.5149 1030.3177,-457.7184 1037.3033,-457.2705"/>
+</g>
+<!-- t15 -->
+<g id="node14" class="node">
+<title>t15</title>
+<polygon fill="#00cc66" stroke="#000000" points="1229.0433,-353.7123 920.9567,-353.7123 916.9567,-349.7123 916.9567,-317.7123 1225.0433,-317.7123 1229.0433,-321.7123 1229.0433,-353.7123"/>
+<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 916.9567,-349.7123 "/>
+<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1225.0433,-317.7123 "/>
+<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1229.0433,-353.7123 "/>
+<text text-anchor="middle" x="1073" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_page</text>
+</g>
+<!-- t12&#45;&gt;t15 -->
+<g id="edge14" class="edge">
+<title>t12&#45;&gt;t15</title>
+<path fill="none" stroke="#0044a0" d="M754.3855,-411.2269C799.2649,-402.117 854.0781,-390.6067 903,-379.252 933.2642,-372.2277 966.3927,-363.8908 995.3012,-356.4015"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="996.5416,-359.6955 1005.3389,-353.7914 994.7799,-352.9208 996.5416,-359.6955"/>
+</g>
+<!-- t16 -->
+<g id="node19" class="node">
+<title>t16</title>
+<polygon fill="#00cc66" stroke="#000000" points="651.9202,-270.1725 284.0798,-270.1725 280.0798,-266.1725 280.0798,-234.1725 647.9202,-234.1725 651.9202,-238.1725 651.9202,-270.1725"/>
+<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 280.0798,-266.1725 "/>
+<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 647.9202,-234.1725 "/>
+<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 651.9202,-270.1725 "/>
+<text text-anchor="middle" x="466" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_debug_page</text>
+</g>
+<!-- t12&#45;&gt;t16 -->
+<g id="edge27" class="edge">
+<title>t12&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M772.353,-411.2135C825.7002,-401.3017 880.0px85,-389.0648 888,-379.252 912.2169,-349.0628 914.0433,-320.8011 888,-292.1725 871.9898,-274.5731 764.4698,-264.5473 661.9779,-258.9344"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="661.9319,-255.4271 651.7593,-258.3883 661.5583,-262.4171 661.9319,-255.4271"/>
+</g>
+<!-- t25 -->
+<g id="node27" class="node">
+<title>t25</title>
+<polygon fill="#efa03b" stroke="#000000" points="488.7205,-323.495 682,-296.1762 875.2795,-323.495 875.0991,-367.6976 488.9009,-367.6976 488.7205,-323.495"/>
+<polygon fill="none" stroke="#000000" points="484.7033,-320.0193 682,-292.1327 879.2967,-320.0193 879.0858,-371.7013 484.9142,-371.7013 484.7033,-320.0193"/>
+<text text-anchor="middle" x="682" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_sidecars</text>
+</g>
+<!-- t12&#45;&gt;t25 -->
+<g id="edge52" class="edge">
+<title>t12&#45;&gt;t25</title>
+<path fill="none" stroke="#0044a0" d="M666.6634,-411.2166C668.3795,-402.768 670.5105,-392.2769 672.6326,-381.8293"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="676.09,-382.3905 674.6507,-371.8939 669.2301,-380.9971 676.09,-382.3905"/>
+</g>
+<!-- t18 -->
+<g id="node18" class="node">
+<title>t18</title>
+<polygon fill="#00cc66" stroke="#000000" points="1147.2989,-270.1725 856.7011,-270.1725 852.7011,-266.1725 852.7011,-234.1725 1143.2989,-234.1725 1147.2989,-238.1725 1147.2989,-270.1725"/>
+<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 852.7011,-266.1725 "/>
+<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1143.2989,-234.1725 "/>
+<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1147.2989,-270.1725 "/>
+<text text-anchor="middle" x="1000" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.combine_layers</text>
+</g>
+<!-- t15&#45;&gt;t18 -->
+<g id="edge24" class="edge">
+<title>t15&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1057.1601,-317.5854C1047.0201,-305.9814 1033.7081,-290.7474 1022.4822,-277.9007"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1025.0247,-275.4911 1015.809,-270.264 1019.7536,-280.0972 1025.0247,-275.4911"/>
+</g>
+<!-- t17&#45;&gt;t18 -->
+<g id="edge23" class="edge">
+<title>t17&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1160.5767,-411.2248C1195.7121,-403.5317 1227.1378,-393.066 1238,-379.252 1261.9222,-348.8288 1262.9959,-321.72 1238,-292.1725 1231.6408,-284.6553 1197.8632,-277.2656 1157.6598,-270.9231"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1158.0117,-267.4362 1147.596,-269.3761 1156.9481,-274.3549 1158.0117,-267.4362"/>
+</g>
+<!-- t17&#45;&gt;t25 -->
+<g id="edge51" class="edge">
+<title>t17&#45;&gt;t25</title>
+<path fill="none" stroke="#0044a0" d="M964.5165,-411.2166C925.304,-400.7368 874.3329,-387.1145 826.4989,-374.3305"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="827.2267,-370.9023 816.6621,-371.7016 825.4193,-377.6649 827.2267,-370.9023"/>
+</g>
+<!-- t13&#45;&gt;t14 -->
+<g id="edge21" class="edge">
+<title>t13&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M672.8683,-493.1993C736.8314,-488.6166 808.4777,-483.5937 874,-479.252 1033.5051,-468.6828 1073.5671,-468.8587 1233,-457.252 1285.8429,-453.4051 1343.3743,-448.7264 1395.4741,-444.3037"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1395.81,-447.7879 1405.477,-443.4521 1395.2161,-440.8131 1395.81,-447.7879"/>
+</g>
+<!-- t13&#45;&gt;t16 -->
+<g id="edge26" class="edge">
+<title>t13&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M479.4495,-479.3529C478.2545,-457.9761 476.5497,-427.7492 475,-401.252 472.546,-359.2945 469.5914,-310.7019 467.754,-280.6981"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="471.2362,-280.2988 467.1306,-270.5318 464.2493,-280.7273 471.2362,-280.2988"/>
+</g>
+<!-- t13&#45;&gt;t19 -->
+<g id="edge29" class="edge">
+<title>t13&#45;&gt;t19</title>
+<path fill="none" stroke="#0044a0" d="M415.2907,-483.9551C385.6007,-473.4286 350.7427,-461.0699 321.6782,-450.7652"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="322.6542,-447.3978 312.0595,-447.3549 320.315,-453.9954 322.6542,-447.3978"/>
+</g>
+<!-- t14&#45;&gt;t18 -->
+<g id="edge25" class="edge">
+<title>t14&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1519.9776,-405.138C1460.3688,-374.2505 1351.1978,-321.2463 1252,-292.1725 1221.8888,-283.3473 1189.1477,-276.3434 1157.5373,-270.8287"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1157.7641,-267.3171 1147.3187,-269.0901 1156.59,-274.2179 1157.7641,-267.3171"/>
+</g>
+<!-- t22 -->
+<g id="node23" class="node">
+<title>t22</title>
+<polygon fill="#efa03b" stroke="#000000" points="489.918,-156.4263 735,-129.1169 980.082,-156.4263 979.8532,-200.6137 490.1468,-200.6137 489.918,-156.4263"/>
+<polygon fill="none" stroke="#000000" points="485.9027,-152.8467 735,-125.09 984.0973,-152.8467 983.8293,-204.6159 486.1707,-204.6159 485.9027,-152.8467"/>
+<text text-anchor="middle" x="735" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_ghostscript</text>
+</g>
+<!-- t18&#45;&gt;t22 -->
+<g id="edge35" class="edge">
+<title>t18&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M942.8252,-234.1499C918.2191,-226.3936 888.4232,-217.0013 859.1708,-207.7804"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="859.8692,-204.3309 849.2796,-204.6625 857.7647,-211.007 859.8692,-204.3309"/>
+</g>
+<!-- t18&#45;&gt;t24 -->
+<g id="edge39" class="edge">
+<title>t18&#45;&gt;t24</title>
+<path fill="none" stroke="#0044a0" d="M852.4716,-245.1051C747.4645,-239.0474 603.1919,-228.6387 477,-212.1725 463.5912,-210.4229 449.788,-208.4132 435.908,-206.238"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="436.2183,-202.7434 425.7921,-204.6263 435.1168,-209.6562 436.2183,-202.7434"/>
+</g>
+<!-- t18&#45;&gt;t23 -->
+<g id="edge44" class="edge">
+<title>t18&#45;&gt;t23</title>
+<path fill="none" stroke="#0044a0" d="M1047.0344,-234.1499C1066.9243,-226.5285 1090.9364,-217.3275 1114.596,-208.2617"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1115.9033,-211.509 1123.9889,-204.6625 1113.3986,-204.9724 1115.9033,-211.509"/>
+</g>
+<!-- t16&#45;&gt;t22 -->
+<g id="edge36" class="edge">
+<title>t16&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M524.0378,-234.1499C549.1239,-226.3599 579.5241,-216.9196 609.3422,-207.6602"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="610.4833,-210.9707 618.9955,-204.6625 608.4073,-204.2857 610.4833,-210.9707"/>
+</g>
+<!-- t16&#45;&gt;t24 -->
+<g id="edge40" class="edge">
+<title>t16&#45;&gt;t24</title>
+<path fill="none" stroke="#0044a0" d="M417.6711,-234.1499C397.2337,-226.5285 372.5607,-217.3275 348.2499,-208.2617"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="349.1912,-204.8773 338.5986,-204.6625 346.7453,-211.4361 349.1912,-204.8773"/>
+</g>
+<!-- t16&#45;&gt;t23 -->
+<g id="edge45" class="edge">
+<title>t16&#45;&gt;t23</title>
+<path fill="none" stroke="#0044a0" d="M651.7596,-235.3105C656.5547,-234.9215 661.308,-234.5415 666,-234.1725 811.2137,-222.7545 848.5512,-230.9322 993,-212.1725 1006.176,-210.4614 1019.7432,-208.4558 1033.3753,-206.2619"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1034.0068,-209.7051 1043.3083,-204.6325 1032.8736,-202.7974 1034.0068,-209.7051"/>
+</g>
+<!-- t21&#45;&gt;t22 -->
+<g id="edge32" class="edge">
+<title>t21&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M1165.0941,-235.3973C1085.9035,-224.7541 997.501,-212.8552 993,-212.1725 980.474,-210.2728 967.615,-208.2703 954.6442,-206.2099"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="955.1339,-202.7439 944.7072,-204.6238 954.0305,-209.6564 955.1339,-202.7439"/>
+</g>
+<!-- t21&#45;&gt;t24 -->
+<g id="edge37" class="edge">
+<title>t21&#45;&gt;t24</title>
+<path fill="none" stroke="#0044a0" d="M1165.2215,-234.9906C1162.1187,-234.699 1159.041,-234.4254 1156,-234.1725 855.1023,-209.1529 777.1841,-244.6523 477,-212.1725 462.026,-210.5524 446.5737,-208.5164 431.0798,-206.2151"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="431.3094,-202.7099 420.8968,-204.6663 430.2567,-209.6303 431.3094,-202.7099"/>
+</g>
+<!-- t21&#45;&gt;t23 -->
+<g id="edge42" class="edge">
+<title>t21&#45;&gt;t23</title>
+<path fill="none" stroke="#0044a0" d="M1274.3771,-234.0471C1268.8557,-227.6412 1262.3807,-220.0px91 1255.8248,-212.5231"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1258.4543,-210.2128 1249.2743,-204.9233 1253.152,-214.783 1258.4543,-210.2128"/>
+</g>
+<!-- t19&#45;&gt;t22 -->
+<g id="edge34" class="edge">
+<title>t19&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M260.6159,-410.9505C259.8316,-364.871 259.0249,-246.8273 271,-234.1725 302.6435,-200.7331 431.3919,-218.4498 477,-212.1725 490.4888,-210.316 504.3558,-208.2974 518.3278,-206.1801"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="519.1602,-209.5936 528.5161,-204.6219 518.1019,-202.6741 519.1602,-209.5936"/>
+</g>
+<!-- t19&#45;&gt;t24 -->
+<g id="edge38" class="edge">
+<title>t19&#45;&gt;t24</title>
+<path fill="none" stroke="#0044a0" d="M259.0498,-411.2136C258.0595,-401.7112 256.8826,-389.8585 256,-379.252 251.2976,-322.7418 247.1648,-257.5963 244.6312,-214.9019"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="248.1192,-214.5935 244.0377,-204.8165 241.1313,-215.0048 248.1192,-214.5935"/>
+</g>
+<!-- t19&#45;&gt;t23 -->
+<g id="edge43" class="edge">
+<title>t19&#45;&gt;t23</title>
+<path fill="none" stroke="#0044a0" d="M282.1528,-411.0561C318.8331,-380.6455 397.8323,-320.0855 476,-292.1725 553.979,-264.3269 579.2078,-283.062 661,-270.1725 742.8817,-257.269 762.3086,-248.2304 844,-234.1725 909.9705,-222.82 927.0119,-223.4221 993,-212.1725 1003.5851,-210.368 1014.439,-208.4601 1025.3874,-206.4907"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1026.2474,-209.892 1035.4637,-204.666 1025,-203.004 1026.2474,-209.892"/>
+</g>
+<!-- t19&#45;&gt;t25 -->
+<g id="edge50" class="edge">
+<title>t19&#45;&gt;t25</title>
+<path fill="none" stroke="#0044a0" d="M342.173,-411.2166C389.8324,-400.6274 451.9326,-386.8297 509.989,-373.9305"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="511.0179,-377.2873 520.0207,-371.7016 509.4996,-370.4539 511.0179,-377.2873"/>
+</g>
+<!-- t20&#45;&gt;t22 -->
+<g id="edge33" class="edge">
+<title>t20&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M323.4472,-984.1943C198.6099,-964.2959 37,-926.6746 37,-865.252 37,-865.252 37,-865.252 37,-429.252 37,-211.1917 260.8906,-241.2767 477,-212.1725 490.808,-210.313 505.0108,-208.2741 519.3157,-206.1255"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="519.8697,-209.5816 529.2317,-204.6214 518.8199,-202.6608 519.8697,-209.5816"/>
+</g>
+<!-- t26 -->
+<g id="node26" class="node">
+<title>t26</title>
+<polygon fill="#efa03b" stroke="#000000" points="568.0652,-47.332 735,-20.0049 901.9348,-47.332 901.779,-91.5483 568.221,-91.5483 568.0652,-47.332"/>
+<polygon fill="none" stroke="#000000" points="564.0538,-43.9306 735,-15.9468 905.9462,-43.9306 905.7643,-95.5532 564.2357,-95.5532 564.0538,-43.9306"/>
+<text text-anchor="middle" x="735" y="-53.5531" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.copy_final</text>
+</g>
+<!-- t22&#45;&gt;t26 -->
+<g id="edge49" class="edge">
+<title>t22&#45;&gt;t26</title>
+<path fill="none" stroke="#0044a0" d="M735,-124.8984C735,-118.6708 735,-112.2475 735,-105.9711"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="738.5001,-105.6074 735,-95.6074 731.5001,-105.6075 738.5001,-105.6074"/>
+</g>
+<!-- t24&#45;&gt;t26 -->
+<g id="edge47" class="edge">
+<title>t24&#45;&gt;t26</title>
+<path fill="none" stroke="#0044a0" d="M368.6395,-140.6178C427.4047,-127.6148 498.1821,-111.9539 561.7955,-97.8781"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="562.9329,-101.2112 571.9406,-95.6333 561.4205,-94.3765 562.9329,-101.2112"/>
+</g>
+<!-- t23&#45;&gt;t26 -->
+<g id="edge48" class="edge">
+<title>t23&#45;&gt;t26</title>
+<path fill="none" stroke="#0044a0" d="M1095.1818,-140.9007C1037.3475,-127.8387 967.4279,-112.0473 904.6388,-97.8663"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="905.1511,-94.3939 894.6256,-95.6048 903.6089,-101.2219 905.1511,-94.3939"/>
+</g>
+</g>
+</svg>

docs/release_notes.rst

@@ -0,0 +1,861 @@
+Release notes
+=============
+
+OCRmyPDF uses `semantic versioning <http://semver.org/>`_ for its command line interface.
+
+The OCRmyPDF package itself does not contain a public API, although it is fairly stable and breaking changes are usually timed with a major release. A future release will clearly define the stable public API.
+
+.. Issue regex
+   find:    [^`]\#([0-9]{1,3})[^0-9]
+   replace: `#$1 <https://github.com/jbarlow83/OCRmyPDF/issues/$1>`_
+
+
+v6.2.4
+------
+
+-   Backport Ghostscript 9.25 compatibility fixes, which removes support for setting Unicode metadata
+-   Backport blacklisting Ghostscript 9.24
+-   Older versions of Ghostscript are still supported
+
+
+v6.2.3
+------
+
+-   Fix compatibility with img2pdf >= 0.3.0 by rejecting input images that have an alpha channel
+
+
+v6.2.2
+------
+
+-   Backport compatibility fixes for Python 3.7 and ruffus 2.7.0 from v7.0.0
+-   Backport fix to ignore masks when deciding what colors are on a page
+-   Backport some minor improvements from v7.0.0: better argument validation and warnings about the Tesseract 4.0.0 ``--user-words`` regression
+
+v6.2.1
+------
+
+-   Fix recent versions of Tesseract (after 4.0.0-beta1) not being detected as supporting the ``sandwich`` renderer (`#271 <https://github.com/ppjbarlow83/OCRmyPDF/issues/271>`_).
+
+
+v6.2.0
+------
+
+-   **Docker**: The Docker image ``ocrmypdf-tess4`` has been removed. The main Docker images, ``ocrmypdf`` and ``ocrmypdf-polyglot`` now use Ubuntu 18.04 as a base image, and as such Tesseract 4.0.0-beta1 is now the Tesseract version they use. There is no Docker image based on Tesseract 3.05 anymore.
+
+-   Creation of PDF/A-3 is now supported. However, there is no ability to attach files to PDF/A-3.
+
+-   List more reasons why the file size might grow.
+
+-   Fix issue `#262 <https://github.com/ppjbarlow83/OCRmyPDF/issues/262>`_, ``--remove-background`` error on PDFs contained colormapped (paletted) images.
+
+-   Fix another XMP metadata validation issue, in cases where the input file's creation date has no timezone and the creation date is not overridden.
+
+
+v6.1.5
+------
+
+-   Fix issue `#253 <https://github.com/jbarlow83/OCRmyPDF/issues/253>`_, a possible division by zero when using the ``hocr`` renderer.
+
+-   Fix incorrectly formatted ``<xmp:ModifyDate>`` field inside XMP metadata for PDF/As.  veraPDF flags this as a PDF/A validation failure. The error is caused the timezone and final digit of the seconds of modified time to be omitted, so at worst the modification time stamp is rounded to the nearest 10 seconds.
+
+
+v6.1.4
+------
+
+-   Fix issue `#248 <https://github.com/jbarlow83/OCRmyPDF/issues/248>`_ ``--clean`` argument may remove OCR from left column of text on certain documents. We now set ``--layout none`` to suppress this.
+
+-   The test cache was updated to reflect the change above.
+
+-   Change test suite to accommodate Ghostscript 9.23's new ability to insert JPEGs into PDFs without transcoding.
+
+-   XMP metadata in PDFs is now examined using ``defusedxml`` for safety.
+
+-   If an external process exits with a signal when asked to report its version, we now print the system error message instead of suppressing it.  This occurred when the required executable was found but was missing a shared library.
+
+-   qpdf 7.0.0 or newer is now required as the test suite can no longer pass without it.
+
+Notes
+~~~~~
+
+-   An apparent `regression in Ghostscript 9.23 <https://bugs.ghostscript.com/show_bug.cgi?id=699216>`_ will cause some ocrmypdf output files to become invalid in rare cases; the workaround for the moment is to set ``--force-ocr``.
+
+
+v6.1.3
+------
+
+-   Fix issue `#247 <https://github.com/jbarlow83/OCRmyPDF/issues/247>`_, ``/CreationDate`` metadata not copied from input to output.
+
+-   A warning is now issued when Python 3.5 is used on files with a large page count, as this case is known to regress to single core performance. The cause of this problem is unknown.
+
+
+v6.1.2
+------
+
+-   Upgrade to PyMuPDF v1.12.5 which includes a more complete fix to `#239 <https://github.com/jbarlow83/OCRmyPDF/issues/239>`_.
+
+-   Add ``defusedxml`` dependency.
+
+
+v6.1.1
+------
+
+-   Fix text being reported as found on all pages if PyMuPDF is not installed.
+
+
+v6.1.0
+------
+
+-   PyMuPDF is now an optional but recommended dependency, to alleviate installation difficulties on platforms that have less access to PyMuPDF than the author anticipated.  Install OCRmyPDF with ``pip install ocrmypdf[fitz]`` to use it to its full potential.
+
+-   Fix ``FileExistsError`` that could occur if OCR timed out while it was generating the output file. (`#218 <https://github.com/jbarlow83/OCRmyPDF/issues/218>`_)
+
+-   Fix table of contents/bookmarks all being redirected to page 1 when generating a PDF/A (with PyMuPDF).  (Without PyMuPDF the table of contents is removed in PDF/A mode.)
+
+-   Fix "RuntimeError: invalid key in dict" when table of contents/bookmarks titles contained the character ``)``. (`#239 <https://github.com/jbarlow83/OCRmyPDF/issues/239>`_)
+
+-   Added a new argument ``--skip-repair`` to skip the initial PDF repair step if the PDF is already well-formed (because another program repaired it).
+
+
+v6.0.0
+------
+
+-   The software license has been changed to GPLv3. Test resource files and some individual sources may have other licenses.
+
+-   OCRmyPDF now depends on `PyMuPDF <https://pymupdf.readthedocs.io/en/latest/installation/>`_. Including PyMuPDF is the primary reason for the change to GPLv3.
+
+-   Other backward incompatible changes
+
+    + The ``OCRMYPDF_TESSERACT``, ``OCRMYPDF_QPDF``, ``OCRMYPDF_GS`` and ``OCRMYPDF_UNPAPER`` environment variables are no longer used. Change ``PATH`` if you need to override the external programs OCRmyPDF uses.
+
+    + The ``ocrmypdf`` package has been moved to ``src/ocrmypdf`` to avoid issues with accidental import.
+
+    + The function ``ocrmypdf.exec.get_program`` was removed.
+
+    + The deprecated module ``ocrmypdf.pageinfo`` was removed.
+
+    + The ``--pdf-renderer tess4`` alias for ``sandwich`` was removed.
+
+-   Fixed an issue where OCRmyPDF failed to detect existing text on pages, depending on how the text and fonts were encoded within the PDF. (`#233 <https://github.com/jbarlow83/OCRmyPDF/issues/233>`_, `#232 <https://github.com/jbarlow83/OCRmyPDF/issues/232>`_)
+
+-   Fixed an issue that caused dramatic inflation of file sizes when ``--skip-text --output-type pdf`` was used. OCRmyPDF now removes duplicate resources such as fonts, images and other objects that it generates. (`#237 <https://github.com/jbarlow83/OCRmyPDF/issues/237>`_)
+
+-   Improved performance of the initial page splitting step. Originally this step was not believed to be expensive and ran in a process. Large file testing revealed it to be a bottleneck, so it is now parallelized. On a 700 page file with quad core machine, this change saves about 2 minutes. (`#234 <https://github.com/jbarlow83/OCRmyPDF/issues/234>`_)
+
+-   The test suite now includes a cache that can be used to speed up test runs across platforms. This also does not require computing checksums, so it's faster. (`#217 <https://github.com/jbarlow83/OCRmyPDF/issues/217>`_)
+
+
+v5.7.0
+------
+
+-   Fixed an issue that caused poor CPU utilization on machines more than 4 cores when running Tesseract 4. (Related to issue `#217 <https://github.com/jbarlow83/OCRmyPDF/issues/217>`_.)
+
+-   The 'hocr' renderer has been improved. The 'sandwich' and 'tesseract' renderers are still better for most use cases, but 'hocr' may be useful for people who work with the PDF.js renderer in English/ASCII languages. (`#225 <https://github.com/jbarlow83/OCRmyPDF/issues/225>`_)
+
+    + It now formats text in a matter that is easier for certain PDF viewers to select and extract copy and paste text. This should help macOS Preview and PDF.js in particular.
+    + The appearance of selected text and behavior of selecting text is improved.
+    + The PDF content stream now uses relative moves, making it more compact and easier for viewers to determine when two words on the same line.
+    + It can now deal with text on a skewed baseline.
+    + Thanks to @cforcey for the pull request, @jbreiden for many helpful suggestions, @ctbarbour for another round of improvements, and @acaloiaro for an independent review.
+
+v5.6.3
+------
+
+-   Suppress two debug messages that were too verbose
+
+
+v5.6.2
+------
+
+-   Development branch accidentally tagged as release. Do not use.
+
+
+v5.6.1
+------
+
+-   Fix issue `#219 <https://github.com/jbarlow83/OCRmyPDF/issues/219>`_: change how the final output file is created to avoid triggering permission errors when the output is a special file such as ``/dev/null``
+-   Fix test suite failures due to a qpdf 8.0.0 regression and Python 3.5's handling of symlink
+-   The "encrypted PDF" error message was different depending on the type of PDF encryption. Now a single clear message appears for all types of PDF encryption.
+-   ocrmypdf is now in Homebrew. Homebrew users are advised to the version of ocrmypdf in the official homebrew-core formulas rather than the private tap.
+-   Some linting
+
+
+v5.6.0
+------
+
+-   Fix issue `#216 <https://github.com/jbarlow83/OCRmyPDF/issues/216>`_: preserve "text as curves" PDFs without rasterizing file
+-   Related to the above, messages about rasterizing are more consistent
+-   For consistency versions minor releases will now get the trailing .0 they always should have had.
+
+
+v5.5
+----
+
+-   Add new argument ``--max-image-mpixels``. Pillow 5.0 now raises an exception when images may be decompression bombs. This argument can be used to override the limit Pillow sets.
+-   Fix output page cropped when using the sandwich renderer and OCR is skipped on a rotated and image-processed page
+-   A warning is now issued when old versions of Ghostscript are used in cases known to cause issues with non-Latin characters
+-   Fix a few parameter validation checks for ``-output-type pdfa-1`` and ``pdfa-2``
+
+
+v5.4.4
+------
+
+-   Fix issue `#181 <https://github.com/jbarlow83/OCRmyPDF/issues/181>`_: fix final merge failure for PDFs with more pages than the system file handle limit (``ulimit -n``)
+-   Fix issue `#200 <https://github.com/jbarlow83/OCRmyPDF/issues/200>`_: an uncommon syntax for formatting decimal numbers in a PDF would cause qpdf to issue a warning, which ocrmypdf treated as an error. Now this the warning is relayed.
+-   Fix an issue where intermediate PDFs would be created at version 1.3 instead of the version of the original file. It's possible but unlikely this had side effects.
+-   A warning is now issued when older versions of qpdf are used since issues like `#200 <https://github.com/jbarlow83/OCRmyPDF/issues/200>`_ cause qpdf to infinite-loop
+-   Address issue `#140 <https://github.com/jbarlow83/OCRmyPDF/issues/140>`_: if Tesseract outputs invalid UTF-8, escape it and print its message instead of aborting with a Unicode error
+-   Adding previously unlisted setup requirement, pytest-runner
+-   Update documentation: fix an error in the example script for Synology with Docker images, improved security guidance, advised ``pip install --user``
+
+
+v5.4.3
+------
+
+-   If a subprocess fails to report its version when queried, exit cleanly with an error instead of throwing an exception
+-   Added test to confirm that the system locale is Unicode-aware and fail early if it's not
+-   Clarified some copyright information
+-   Updated pinned requirements.txt so the homebrew formula captures more recent versions
+
+
+v5.4.2
+------
+
+-   Fixed a regression from v5.4.1 that caused sidecar files to be created as empty files
+
+
+v5.4.1
+------
+
+-   Add workaround for Tesseract v4.00alpha crash when trying to obtain orientation and the latest language packs are installed
+
+
+v5.4
+----
+
+-   Change wording of a deprecation warning to improve clarity
+-   Added option to generate PDF/A-1b output if desired (``--output-type pdfa-1``); default remains PDF/A-2b generation
+-   Update documentation
+
+
+v5.3.3
+------
+
+-   Fixed missing error message that should occur when trying to force ``--pdf-renderer sandwich`` on old versions of Tesseract
+-   Update copyright information in test files
+-   Set system ``LANG`` to UTF-8 in Dockerfiles to avoid UTF-8 encoding errors
+
+
+v5.3.2
+------
+
+-   Fixed a broken test case related to language packs
+
+
+v5.3.1
+------
+
+-   Fixed wrong return code given for missing Tesseract language packs
+-   Fixed "brew audit" crashing on Travis when trying to auto-brew
+
+
+v5.3
+----
+
+-   Added ``--user-words`` and ``--user-patterns`` arguments which are forwarded to Tesseract OCR as words and regular expressions respective to use to guide OCR. Supplying a list of subject-domain words should assist Tesseract with resolving words. (`#165 <https://github.com/jbarlow83/OCRmyPDF/issues/165>`_)
+-   Using a non Latin-1 language with the "hocr" renderer now warns about possible OCR quality and recommends workarounds (`#176 <https://github.com/jbarlow83/OCRmyPDF/issues/176>`_)
+-   Output file path added to error message when that location is not writable (`#175 <https://github.com/jbarlow83/OCRmyPDF/issues/175>`_)
+-   Otherwise valid PDFs with leading whitespace at the beginning of the file are now accepted
+
+
+v5.2
+----
+
+-   When using Tesseract 3.05.01 or newer, OCRmyPDF will select the "sandwich" PDF renderer by default, unless another PDF renderer is specified with the ``--pdf-renderer`` argument. The previous behavior was to select ``--pdf-renderer=hocr``.
+-   The "tesseract" PDF renderer is now deprecated, since it can cause problems with Ghostscript on Tesseract 3.05.00
+-   The "tess4" PDF renderer has been renamed to "sandwich". "tess4" is now a deprecated alias for "sandwich".
+
+
+v5.1
+----
+
+-   Files with pages larger than 200" (5080 mm) in either dimension are now supported with ``--output-type=pdf`` with the page size preserved (in the PDF specification this feature is called UserUnit scaling). Due to Ghostscript limitations this is not available in conjunction with PDF/A output.
+
+
+v5.0.1
+------
+
+-   Fixed issue `#169 <https://github.com/jbarlow83/OCRmyPDF/issues/169>`_, exception due to failure to create sidecar text files on some versions of Tesseract 3.04, including the jbarlow83/ocrmypdf Docker image
+
+
+v5.0
+----
+
+-   Backward incompatible changes
+
+     + Support for Python 3.4 dropped. Python 3.5 is now required.
+     + Support for Tesseract 3.02 and 3.03 dropped. Tesseract 3.04 or newer is required. Tesseract 4.00 (alpha) is supported.
+     + The OCRmyPDF.sh script was removed.
+
+-   Add a new feature, ``--sidecar``, which allows creating "sidecar" text files which contain the OCR results in plain text. These OCR text is more reliable than extracting text from PDFs. Closes `#126 <https://github.com/jbarlow83/OCRmyPDF/issues/126>`_.
+-   New feature: ``--pdfa-image-compression``, which allows overriding Ghostscript's lossy-or-lossless image encoding heuristic and making all images JPEG encoded or lossless encoded as desired. Fixes `#163 <https://github.com/jbarlow83/OCRmyPDF/issues/163>`_.
+-   Fixed issue `#143 <https://github.com/jbarlow83/OCRmyPDF/issues/143>`_, added ``--quiet`` to suppress "INFO" messages
+-   Fixed issue `#164 <https://github.com/jbarlow83/OCRmyPDF/issues/164>`_, a typo
+-   Removed the command line parameters ``-n`` and ``--just-print`` since they have not worked for some time (reported as Ubuntu bug `#1687308 <https://bugs.launchpad.net/ubuntu/+source/ocrmypdf/+bug/1687308>`_)
+
+
+v4.5.6
+------
+
+-   Fixed issue `#156 <https://github.com/jbarlow83/OCRmyPDF/issues/156>`_, 'NoneType' object has no attribute 'getObject' on pages with no optional /Contents record.  This should resolve all issues related to pages with no /Contents record.
+-   Fixed issue `#158 <https://github.com/jbarlow83/OCRmyPDF/issues/158>`_, ocrmypdf now stops and terminates if Ghostscript fails on an intermediate step, as it is not possible to proceed.
+-   Fixed issue `#160 <https://github.com/jbarlow83/OCRmyPDF/issues/160>`_, exception thrown on certain invalid arguments instead of error message
+
+
+v4.5.5
+------
+
+-   Automated update of macOS homebrew tap
+-   Fixed issue `#154 <https://github.com/jbarlow83/OCRmyPDF/issues/154>`_, KeyError '/Contents' when searching for text on blank pages that have no /Contents record.  Note: incomplete fix for this issue.
+
+
+v4.5.4
+------
+
+-   Fix ``--skip-big`` raising an exception if a page contains no images (`#152 <https://github.com/jbarlow83/OCRmyPDF/issues/152>`_) (thanks to @TomRaz)
+-   Fix an issue where pages with no images might trigger "cannot write mode P as JPEG" (`#151 <https://github.com/jbarlow83/OCRmyPDF/issues/151>`_)
+
+
+v4.5.3
+------
+
+-   Added a workaround for Ghostscript 9.21 and probably earlier versions would fail with the error message "VMerror -25", due to a Ghostscript bug in XMP metadata handling
+-   High Unicode characters (U+10000 and up) are no longer accepted for setting metadata on the command line, as Ghostscript may not handle them correctly.
+-   Fixed an issue where the ``tess4`` renderer would duplicate content onto output pages if tesseract failed or timed out
+-   Fixed ``tess4`` renderer not recognized when lossless reconstruction is possible
+
+
+v4.5.2
+------
+
+-   Fix issue `#147 <https://github.com/jbarlow83/OCRmyPDF/issues/147>`_. ``--pdf-renderer tess4 --clean`` will produce an oversized page containing the original image in the bottom left corner, due to loss DPI information.
+-   Make "using Tesseract 4.0" warning less ominous
+-   Set up machinery for homebrew OCRmyPDF tap
+
+
+v4.5.1
+------
+
+-   Fix issue `#137 <https://github.com/jbarlow83/OCRmyPDF/issues/137>`_, proportions of images with a non-square pixel aspect ratio would be distorted in output for ``--force-ocr`` and some other combinations of flags
+
+
+v4.5
+----
+
+-   Exotic PDFs containing "Form XObjects" are now supported (issue `#134 <https://github.com/jbarlow83/OCRmyPDF/issues/134>`_; PDF reference manual 8.10), and images they contain are taken into account when determining the resolution for rasterizing
+-   The Tesseract 4 Docker image no longer includes all languages, because it took so long to build something would tend to fail
+-   OCRmyPDF now warns about using ``--pdf-renderer tesseract`` with Tesseract 3.04 or lower due to issues with Ghostscript corrupting the OCR text in these cases
+
+
+v4.4.2
+------
+
+-   The Docker images (ocrmypdf, ocrmypdf-polyglot, ocrmypdf-tess4) are now based on Ubuntu 16.10 instead of Debian stretch
+
+    + This makes supporting the Tesseract 4 image easier
+    + This could be a disruptive change for any Docker users who built customized these images with their own changes, and made those changes in a way that depends on Debian and not Ubuntu
+
+-   OCRmyPDF now prevents running the Tesseract 4 renderer with Tesseract 3.04, which was permitted in v4.4 and v4.4.1 but will not work
+
+
+v4.4.1
+------
+
+-   To prevent a `TIFF output error <https://github.com/python-pillow/Pillow/issues/2206>`_ caused by img2pdf >= 0.2.1 and Pillow <= 3.4.2, dependencies have been tightened
+-   The Tesseract 4.00 simultaneous process limit was increased from 1 to 2, since it was observed that 1 lowers performance
+-   Documentation improvements to describe the ``--tesseract-config`` feature
+-   Added test cases and fixed error handling for ``--tesseract-config``
+-   Tweaks to setup.py to deal with issues in the v4.4 release
+
+v4.4
+----
+
+-   Tesseract 4.00 is now supported on an experimental basis.
+
+    +  A new rendering option ``--pdf-renderer tess4`` exploits Tesseract 4's new text-only output PDF mode. See the documentation on PDF Renderers for details.
+    +  The ``--tesseract-oem`` argument allows control over the Tesseract 4 OCR engine mode (tesseract's ``--oem``). Use ``--tesseract-oem 2`` to enforce the new LSTM mode.
+    +  Fixed poor performance with Tesseract 4.00 on Linux
+
+-   Fixed an issue that caused corruption of output to stdout in some cases
+-   Removed test for Pillow JPEG and PNG support, as the minimum supported version of Pillow now enforces this
+-   OCRmyPDF now tests that the intended destination file is writable before proceeding
+-   The test suite now requires ``pytest-helpers-namespace`` to run (but not install)
+-   Significant code reorganization to make OCRmyPDF re-entrant and improve performance. All changes should be backward compatible for the v4.x series.
+
+    + However, OCRmyPDF's dependency "ruffus" is not re-entrant, so no Python API is available. Scripts should continue to use the command line interface.
+
+
+v4.3.5
+------
+
+-   Update documentation to confirm Python 3.6.0 compatibility. No code changes were needed, so many earlier versions are likely supported.
+
+
+v4.3.4
+------
+
+-   Fixed "decimal.InvalidOperation: quantize result has too many digits" for high DPI images
+
+
+v4.3.3
+------
+
+-   Fixed PDF/A creation with Ghostscript 9.20 properly
+-   Fixed an exception on inline stencil masks with a missing optional parameter
+
+
+v4.3.2
+------
+
+-   Fixed a PDF/A creation issue with Ghostscript 9.20 (note: this fix did not actually work)
+
+
+v4.3.1
+------
+
+-   Fixed an issue where pages produced by the "hocr" renderer after a Tesseract timeout would be rotated incorrectly if the input page was rotated with a /Rotate marker
+-   Fixed a file handle leak in LeptonicaErrorTrap that would cause a "too many open files" error for files around hundred pages of pages long when ``--deskew`` or ``--remove-background`` or other Leptonica based image processing features were in use, depending on the system value of ``ulimit -n``
+-   Ability to specify multiple languages for multilingual documents is now advertised in documentation
+-   Reduced the file sizes of some test resources
+-   Cleaned up debug output
+-   Tesseract caching in test cases is now more cautious about false cache hits and reproducing exact output, not that any problems were observed
+
+
+v4.3
+----
+
+-   New feature ``--remove-background`` to detect and erase the background of color and grayscale images
+-   Better documentation
+-   Fixed an issue with PDFs that draw images when the raster stack depth is zero
+-   ocrmypdf can now redirect its output to stdout for use in a shell pipeline
+
+    +  This does not improve performance since temporary files are still used for buffering
+    +  Some output validation is disabled in this mode
+
+v4.2.5
+------
+
+-   Fixed an issue (`#100 <https://github.com/jbarlow83/OCRmyPDF/issues/100>`_) with PDFs that omit the optional /BitsPerComponent parameter on images
+-   Removed non-free file milk.pdf
+
+
+v4.2.4
+------
+
+-   Fixed an error (`#90 <https://github.com/jbarlow83/OCRmyPDF/issues/90>`_) caused by PDFs that use stencil masks properly
+-   Fixed handling of PDFs that try to draw images or stencil masks without properly setting up the graphics state (such images are now ignored for the purposes of calculating DPI)
+
+v4.2.3
+------
+
+-   Fixed an issue with PDFs that store page rotation (/Rotate) in an indirect object
+-   Integrated a few fixes to simplify downstream packaging (Debian)
+
+    +  The test suite no longer assumes it is installed
+    +  If running Linux, skip a test that passes Unicode on the command line
+
+-   Added a test case to check explicit masks and stencil masks
+-   Added a test case for indirect objects and linearized PDFs
+-   Deprecated the OCRmyPDF.sh shell script
+
+
+v4.2.2
+------
+
+-   Improvements to documentation
+
+
+v4.2.1
+------
+
+-   Fixed an issue where PDF pages that contained stencil masks would report an incorrect DPI and cause Ghostscript to abort
+-   Implemented stdin streaming
+
+
+v4.2
+----
+
+-   ocrmypdf will now try to convert single image files to PDFs if they are provided as input (`#15 <https://github.com/jbarlow83/OCRmyPDF/issues/15>`_)
+
+    +  This is a basic convenience feature. It only supports a single image and always makes the image fill the whole page.
+    +  For better control over image to PDF conversion, use ``img2pdf`` (one of ocrmypdf's dependencies)
+
+-   New argument ``--output-type {pdf|pdfa}`` allows disabling Ghostscript PDF/A generation
+
+    +  ``pdfa`` is the default, consistent with past behavior
+    +  ``pdf`` provides a workaround for users concerned about the increase in file size from Ghostscript forcing JBIG2 images to CCITT and transcoding JPEGs
+    +  ``pdf`` preserves as much as it can about the original file, including problems that PDF/A conversion fixes
+
+-   PDFs containing images with "non-square" pixel aspect ratios, such as 200x100 DPI, are now handled and converted properly (fixing a bug that caused to be cropped)
+-   ``--force-ocr`` rasterizes pages even if they contain no images
+
+    +  supports users who want to use OCRmyPDF to reconstruct text information in PDFs with damaged Unicode maps (copy and paste text does not match displayed text)
+    +  supports reinterpreting PDFs where text was rendered as curves for printing, and text needs to be recovered
+    +  fixes issue `#82 <https://github.com/jbarlow83/OCRmyPDF/issues/82>`_
+
+-   Fixes an issue where, with certain settings, monochrome images in PDFs would be converted to 8-bit grayscale, increasing file size (`#79 <https://github.com/jbarlow83/OCRmyPDF/issues/79>`_)
+-   Support for Ubuntu 12.04 LTS "precise" has been dropped in favor of (roughly) Ubuntu 14.04 LTS "trusty"
+
+    +  Some Ubuntu "PPAs" (backports) are needed to make it work
+
+-   Support for some older dependencies dropped
+
+    +  Ghostscript 9.15 or later is now required (available in Ubuntu trusty with backports)
+    +  Tesseract 3.03 or later is now required (available in Ubuntu trusty)
+
+-   Ghostscript now runs in "safer" mode where possible
+
+v4.1.4
+------
+
+-   Bug fix: monochrome images with an ICC profile attached were incorrectly converted to full color images if lossless reconstruction was not possible due to other settings; consequence was increased file size for these images
+
+
+v4.1.3
+------
+
+-   More helpful error message for PDFs with version 4 security handler
+-   Update usage instructions for Windows/Docker users
+-   Fix order of operations for matrix multiplication (no effect on most users)
+-   Add a few leptonica wrapper functions (no effect on most users)
+
+
+v4.1.2
+------
+
+-   Replace IEC sRGB ICC profile with Debian's sRGB (from icc-profiles-free) which is more compatible with the MIT license
+-   More helpful error message for an error related to certain types of malformed PDFs
+
+
+v4.1
+----
+
+-   ``--rotate-pages`` now only rotates pages when reasonably confidence in the orientation. This behavior can be adjusted with the new argument ``--rotate-pages-threshold``
+-   Fixed problems in error checking if ``unpaper`` is uninstalled or missing at run-time
+-   Fixed problems with "RethrownJobError" errors during error handling that suppressed the useful error messages
+
+
+v4.0.7
+------
+
+-   Minor correction to Ghostscript output settings
+
+
+v4.0.6
+------
+
+-   Update install instructions
+-   Provide a sRGB profile instead of using Ghostscript's
+
+
+v4.0.5
+------
+
+-   Remove some verbose debug messages from v4.0.4
+-   Fixed temporary that wasn't being deleted
+-   DPI is now calculated correctly for cropped images, along with other image transformations
+-   Inline images are now checked during DPI calculation instead of rejecting the image
+
+v4.0.4
+------
+
+Released with verbose debug message turned on. Do not use. Skip to v4.0.5.
+
+
+v4.0.3
+------
+
+New features
+^^^^^^^^^^^^
+
+-   Page orientations detected are now reported in a summary comment
+
+
+Fixes
+^^^^^
+
+-   Show stack trace if unexpected errors occur
+-   Treat "too few characters" error message from Tesseract as a reason to skip that page rather than
+    abort the file
+-   Docker: fix blank JPEG2000 issue by insisting on Ghostscript versions that have this fixed
+
+
+v4.0.2
+------
+
+Fixes
+^^^^^
+
+-   Fixed compatibility with Tesseract 3.04.01 release, particularly its different way of outputting
+    orientation information
+-   Improved handling of Tesseract errors and crashes
+-   Fixed use of chmod on Docker that broke most test cases
+
+
+v4.0.1
+------
+
+Fixes
+^^^^^
+
+-   Fixed a KeyError if tesseract fails to find page orientation information
+
+
+v4.0
+----
+
+New features
+^^^^^^^^^^^^
+
+-   Automatic page rotation (``-r``) is now available. It uses ignores any prior rotation information
+    on PDFs and sets rotation based on the dominant orientation of detectable text. This feature is
+    fairly reliable but some false positives occur especially if there is not much text to work with. (`#4 <https://github.com/jbarlow83/OCRmyPDF/issues/4>`_)
+-   Deskewing is now performed using Leptonica instead of unpaper. Leptonica is faster and more reliable
+    at image deskewing than unpaper.
+
+
+Fixes
+^^^^^
+
+-   Fixed an issue where lossless reconstruction could cause some pages to be appear incorrectly
+    if the page was rotated by the user in Acrobat after being scanned (specifically if it a /Rotate tag)
+-   Fixed an issue where lossless reconstruction could misalign the graphics layer with respect to
+    text layer if the page had been cropped such that its origin is not (0, 0) (`#49 <https://github.com/jbarlow83/OCRmyPDF/issues/49>`_)
+
+
+Changes
+^^^^^^^
+
+-   Logging output is now much easier to read
+-   ``--deskew`` is now performed by Leptonica instead of unpaper (`#25 <https://github.com/jbarlow83/OCRmyPDF/issues/25>`_)
+-   libffi is now required
+-   Some changes were made to the Docker and Travis build environments to support libffi
+-   ``--pdf-renderer=tesseract`` now displays a warning if the Tesseract version is less than 3.04.01,
+    the planned release that will include fixes to an important OCR text rendering bug in Tesseract 3.04.00.
+    You can also manually install ./share/sharp2.ttf on top of pdf.ttf in your Tesseract tessdata folder
+    to correct the problem.
+
+
+v3.2.1
+------
+
+Changes
+^^^^^^^
+
+-   Fixed issue `#47 <https://github.com/jbarlow83/OCRmyPDF/issues/47>`_ "convert() got and unexpected keyword argument 'dpi'" by upgrading to img2pdf 0.2
+-   Tweaked the Dockerfiles
+
+
+v3.2
+----
+
+New features
+^^^^^^^^^^^^
+
+-   Lossless reconstruction: when possible, OCRmyPDF will inject text layers without
+    otherwise manipulating the content and layout of a PDF page. For example, a PDF containing a mix
+    of vector and raster content would see the vector content preserved. Images may still be transcoded
+    during PDF/A conversion.  (``--deskew`` and ``--clean-final`` disable this mode, necessarily.)
+-   New argument ``--tesseract-pagesegmode`` allows you to pass page segmentation arguments to Tesseract OCR.
+    This helps for two column text and other situations that confuse Tesseract.
+-   Added a new "polyglot" version of the Docker image, that generates Tesseract with all languages packs installed,
+    for the polyglots among us. It is much larger.
+
+Changes
+^^^^^^^
+
+-   JPEG transcoding quality is now 95 instead of the default 75. Bigger file sizes for less degradation.
+
+
+
+v3.1.1
+------
+
+Changes
+^^^^^^^
+
+-   Fixed bug that caused incorrect page size and DPI calculations on documents with mixed page sizes
+
+v3.1
+----
+
+Changes
+^^^^^^^
+
+-   Default output format is now PDF/A-2b instead of PDF/A-1b
+-   Python 3.5 and macOS El Capitan are now supported platforms - no changes were
+    needed to implement support
+-   Improved some error messages related to missing input files
+-   Fixed issue `#20 <https://github.com/jbarlow83/OCRmyPDF/issues/20>`_ - uppercase .PDF extension not accepted
+-   Fixed an issue where OCRmyPDF failed to text that certain pages contained previously OCR'ed text,
+    such as OCR text produced by Tesseract 3.04
+-   Inserts /Creator tag into PDFs so that errors can be traced back to this project
+-   Added new option ``--pdf-renderer=auto``, to let OCRmyPDF pick the best PDF renderer.
+    Currently it always chooses the 'hocrtransform' renderer but that behavior may change.
+-   Set up Travis CI automatic integration testing
+
+v3.0
+----
+
+New features
+^^^^^^^^^^^^
+
+-   Easier installation with a Docker container or Python's ``pip`` package manager
+-   Eliminated many external dependencies, so it's easier to setup
+-   Now installs ``ocrmypdf`` to ``/usr/local/bin`` or equivalent for system-wide
+    access and easier typing
+-   Improved command line syntax and usage help (``--help``)
+-   Tesseract 3.03+ PDF page rendering can be used instead for better positioning
+    of recognized text (``--pdf-renderer tesseract``)
+-   PDF metadata (title, author, keywords) are now transferred to the
+    output PDF
+-   PDF metadata can also be set from the command line (``--title``, etc.)
+-   Automatic repairs malformed input PDFs if possible
+-   Added test cases to confirm everything is working
+-   Added option to skip extremely large pages that take too long to OCR and are
+    often not OCRable (e.g. large scanned maps or diagrams); other pages are still
+    processed (``--skip-big``)
+-   Added option to kill Tesseract OCR process if it seems to be taking too long on
+    a page, while still processing other pages (``--tesseract-timeout``)
+-   Less common colorspaces (CMYK, palette) are now supported by conversion to RGB
+-   Multiple images on the same PDF page are now supported
+
+Changes
+^^^^^^^
+
+-   New, robust rewrite in Python 3.4+ with ruffus_ pipelines
+-   Now uses Ghostscript 9.14's improved color conversion model to preserve PDF colors
+-   OCR text is now rendered in the PDF as invisible text. Previous versions of OCRmyPDF
+    incorrectly rendered visible text with an image on top.
+-   All "tasks" in the pipeline can be executed in parallel on any
+    available CPUs, increasing performance
+-   The ``-o DPI`` argument has been phased out, in favor of ``--oversample DPI``, in
+    case we need ``-o OUTPUTFILE`` in the future
+-   Removed several dependencies, so it's easier to install.  We no
+    longer use:
+
+    - GNU parallel_
+    - ImageMagick_
+    - Python 2.7
+    - Poppler
+    - MuPDF_ tools
+    - shell scripts
+    - Java and JHOVE_
+    - libxml2
+
+-   Some new external dependencies are required or optional, compared to v2.x:
+
+    - Ghostscript 9.14+
+    - qpdf_ 5.0.0+
+    - Unpaper_ 6.1 (optional)
+    - some automatically managed Python packages
+
+.. _ruffus: http://www.ruffus.org.uk/index.html
+.. _parallel: https://www.gnu.org/software/parallel/
+.. _ImageMagick: http://www.imagemagick.org/script/index.php
+.. _MuPDF: http://mupdf.com/docs/
+.. _qpdf: http://qpdf.sourceforge.net/
+.. _Unpaper: https://github.com/Flameeyes/unpaper
+.. _JHOVE: http://jhove.sourceforge.net/
+
+Release candidates
+^^^^^^^^^^^^^^^^^^
+
+-   rc9:
+
+    - fix issue `#118 <https://github.com/jbarlow83/OCRmyPDF/issues/118>`_: report error if ghostscript iccprofiles are missing
+    - fixed another issue related to `#111 <https://github.com/jbarlow83/OCRmyPDF/issues/111>`_: PDF rasterized to palette file
+    - add support image files with a palette
+    - don't try to validate PDF file after an exception occurs
+
+-   rc8:
+
+    - fix issue `#111 <https://github.com/jbarlow83/OCRmyPDF/issues/111>`_: exception thrown if PDF is missing DocumentInfo dictionary
+
+-   rc7:
+
+    - fix error when installing direct from pip, "no such file 'requirements.txt'"
+
+-   rc6:
+
+    - dropped libxml2 (Python lxml) since Python 3's internal XML parser is sufficient
+    - set up Docker container
+    - fix Unicode errors if recognized text contains Unicode characters and system locale is not UTF-8
+
+-   rc5:
+
+    - dropped Java and JHOVE in favour of qpdf
+    - improved command line error output
+    - additional tests and bug fixes
+    - tested on Ubuntu 14.04 LTS
+
+-   rc4:
+
+    - dropped MuPDF in favour of qpdf
+    - fixed some installer issues and errors in installation instructions
+    - improve performance: run Ghostscript with multithreaded rendering
+    - improve performance: use multiple cores by default
+    - bug fix: checking for wrong exception on process timeout
+
+-   rc3: skipping version number intentionally to avoid confusion with Tesseract
+-   rc2: first release for public testing to test-PyPI, Github
+-   rc1: testing release process
+
+Compatibility notes
+-------------------
+
+-   ``./OCRmyPDF.sh`` script is still available for now
+-   Stacking the verbosity option like ``-vvv`` is no longer supported
+
+-   The configuration file ``config.sh`` has been removed.  Instead, you can
+    feed a file to the arguments for common settings:
+
+::
+
+    ocrmypdf input.pdf output.pdf @settings.txt
+
+where ``settings.txt`` contains *one argument per line*, for example:
+
+::
+
+    -l
+    deu
+    --author
+    A. Merkel
+    --pdf-renderer
+    tesseract
+
+
+Fixes
+^^^^^
+
+-   Handling of filenames containing spaces: fixed
+
+Notes and known issues
+^^^^^^^^^^^^^^^^^^^^^^
+
+-   Some dependencies may work with lower versions than tested, so try
+    overriding dependencies if they are "in the way" to see if they work.
+
+-   ``--pdf-renderer tesseract`` will output files with an incorrect page size in Tesseract 3.03,
+    due to a bug in Tesseract.
+
+-   PDF files containing "inline images" are not supported and won't be for the 3.0 release. Scanned
+    images almost never contain inline images.
+
+
+v2.2-stable (2014-09-29)
+------------------------
+
+OCRmyPDF versions 1 and 2 were implemented as shell scripts. OCRmyPDF 3.0+ is a fork that gradually replaced all shell scripts with Python while maintaining the existing command line arguments. No one is maintaining old versions.
+
+For details on older versions, see the `final version of its release notes <https://github.com/fritz-hh/OCRmyPDF/blob/7fd3dbdf42ca53a619412ce8add7532c5e81a9d1/RELEASE_NOTES.md>`_.

docs/security.rst

@@ -0,0 +1,79 @@
+PDF security issues
+===================
+
+	OCRmyPDF should only be used on PDFs you trust. It is not designed to protect you against malware. 
+
+Recognizing that many users have an interest in handling PDFs and applying OCR to PDFs they did not generate themselves, this article discusses the security implications of PDFs and how users can protect themselves.
+
+The disclaimer applies: this software has no warranties of any kind.
+
+PDFs may contain malware
+------------------------
+
+PDF is a rich, complex file format. The official PDF 1.7 specification, ISO 32000:2008, is hundreds of pages long and references several annexes each of which are similar in length. PDFs can contain video, audio, XML, JavaScript and other programming, and forms. In some cases, they can open internet connections to pre-selected URLs. All of these possible attack vectors.
+
+In short, PDFs `may contain viruses <https://security.stackexchange.com/questions/64052/can-a-pdf-file-contain-a-virus>`_.
+
+This `article <https://theinvisiblethings.blogspot.ca/2013/02/converting-untrusted-pdfs-into-trusted.html>`_ describes a high-paranoia method which allows potentially hostile PDFs to be viewed and rasterized safely in a disposable virtual machine. A trusted PDF created in this manner is converted to images and loses all information making it searchable and losing all compression. OCRmyPDF could be used restore searchability.
+
+How OCRmyPDF processes PDFs
+---------------------------
+
+OCRmyPDF must open and interpret your PDF in order to insert an OCR layer. First, it runs all PDFs through `qpdf <https://github.com/qpdf/qpdf>`_, a program that repairs PDFs with syntax errors. This is done because, in the author's experience, a significant number of PDFs in the wild especially those created by scanners are not well-formed files. qpdf makes it more likely that OCRmyPDF will succeed, but offers no security guarantees. qpdf is also used to split the PDF into single page PDFs.
+
+After qpdf, OCRmyPDF examines each page using `PyPDF2 <https://github.com/mstamy2/PyPDF2>`_. This library also has no warranties or guarantees. OCRmyPDF works with qpdf 5.0 and up, but version 7.0 is recommended because of known security vulnerabilities in early versions.
+
+Finally, OCRmyPDF rasterizes each page of the PDF using `Ghostscript <http://ghostscript.com/>`_ in ``-dSAFER`` mode.
+
+Depending on the options specified, OCRmyPDF may graft the OCR layer into the existing PDF or it may essentially reconstruct ("re-fry") a visually identical PDF that may be quite different at the binary level. That said, OCRmyPDF is not a tool designed for sanitizing PDFs.
+
+Using OCRmyPDF online or as a service
+-------------------------------------
+
+OCRmyPDF should not be deployed as a public-facing service, such as a website where a potential attacker could upload a PDF of their choice for OCR. OCRmyPDF is not designed to be secure against PDF malware. Another concern is PDFs specifically designed to be a denial of service attack: PDFs can contain recursive data structures that sometimes send parsers into infinite loops, and issue complex graphics drawing commands.
+
+Setting aside these concerns, a side effect of OCRmyPDF is it may incidentally sanitize PDFs that contain malware. It runs ``qpdf`` to repair the PDF, which could correct malformed PDF structures that are part of an attack. When PDF/A output is selected (the default), the input PDF is partially reconstructed by Ghostscript. When ``--force-ocr`` is used, all pages are rasterized and reconverted to PDF, which could remove malware in embedded images. No guarantees.
+
+OCRmyPDF should be relatively safe to use in a trusted intranet, with some considerations:
+
+Limiting CPU usage
+^^^^^^^^^^^^^^^^^^
+
+OCRmyPDF will attempt to use all available CPUs and storage, so executing ``nice ocrmypdf`` or limiting the number of jobs with the ``-j`` argument may ensure the server remains available. Another option would be run OCRmyPDF jobs inside a Docker container, a virtual machine, or a cloud instance, which can impose its own limits on CPU usage and be terminated "from orbit" if it fails to complete.
+
+Temporary storage requirements
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+OCRmyPDF will use a large amount of temporary storage for its work, proportional to the total number of pixels needed to rasterize the PDF. The raster image of a 8.5×11" color page at 300 DPI takes 25 MB uncompressed; OCRmyPDF saves its intermediates as PNG, but that still means it requires about 9 MB per intermediate based on average compression ratios. Multiple intermediates per page are also required, depending on the command line given. A rule of thumb would be to allow 100 MB of temporary storage per page in a file – meaning that a small cloud servers or small VM partitions should be provisioned with plenty of extra space, if say, a 500 page file might be sent.
+
+To check temporary storage usage on actual files, run ``ocrmypdf -k ...`` which will preserve and print the path to temporary storage when the job is done.
+
+To change where temporary files are stored, change the ``TMPDIR`` environment variable for ocrmypdf's environment. (Python's ``tempfile.gettempdir()`` returns the root directory in which temporary files will be stored.) For example, one could redirect ``TMPDIR`` to a large RAM disk to avoid wear on HDD/SSD and potentially improve performance. On Amazon Web Services, ``TMPDIR`` can be set to `empheral storage <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/InstanceStorage.html>`_.
+
+Timeouts
+^^^^^^^^
+
+To prevent excessively long OCR jobs consider setting ``--tesseract-timeout`` and/or ``--skip-big`` arguments. ``--skip-big`` is particularly helpful if your PDFs include documents such as reports on standard page sizes with large images attached - often large images are not worth OCR'ing anyway.
+
+Commercial alternatives
+^^^^^^^^^^^^^^^^^^^^^^^
+
+The author also provides professional services that include OCR and building databases around PDFs, and is happy to provide consultation.
+
+Abbyy Cloud OCR is a viable commercial alternative with a web services API. 
+
+
+Password protection, digital signatures and certification
+---------------------------------------------------------
+
+Password protected PDFs usually have two passwords, and owner and user password. When the user password is set to empty, PDF readers will open the file automatically and marked it as "(SECURED)". While not as reliable as a digital signature, this indicates that whoever set the password approved of the file at that time. When the user password is set, the document cannot be viewed without the password. 
+
+Either way, OCRmyPDF does not remove passwords from PDFs and exits with an error on encountering them.
+
+``qpdf``, one of OCRmyPDF's dependencies, can remove passwords. If the owner and user password are set, a password is required for ``qpdf``. If only the owner password is set, then the password can be stripped, even if one does not have the owner password.
+
+After OCR is applied, password protection is not permitted on PDF/A documents but the file can be converted to regular PDF.
+
+Many programs exist which are capable of inserting an image of someone's signature. On its own, this offers no security guarantees. It is trivial to remove the signature image and apply it to other files. This practice offers no real security.
+
+Important documents can be digitally signed and certified to attest to their authorship. OCRmyPDF cannot do this. Open source tools such as pdfbox (Java) have this capability as does Adobe Acrobat. 

ocrmypdf/__init__.py

@@ -1,12 +0,0 @@
-from enum import IntEnum
-
-
-class ExitCode(IntEnum):
-    ok = 0
-    bad_args = 1
-    input_file = 2
-    missing_dependency = 3
-    invalid_output_pdfa = 4
-    file_access_error = 5
-    already_done_ocr = 6
-    other_error = 15

ocrmypdf/ghostscript.py

@@ -1,54 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from tempfile import NamedTemporaryFile
-from subprocess import Popen, PIPE, check_call
-from shutil import copy
-
-
-def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log):
-    with NamedTemporaryFile(delete=True) as tmp:
-        args_gs = [
-            'gs',
-            '-dQUIET',
-            '-dBATCH',
-            '-dNOPAUSE',
-            '-sDEVICE=%s' % raster_device,
-            '-o', tmp.name,
-            '-r{0}x{1}'.format(str(xres), str(yres)),
-            input_file
-        ]
-
-        p = Popen(args_gs, close_fds=True, stdout=PIPE, stderr=PIPE,
-                  universal_newlines=True)
-        stdout, stderr = p.communicate()
-        if stdout:
-            log.debug(stdout)
-        if stderr:
-            log.error(stderr)
-
-        if p.returncode == 0:
-            copy(tmp.name, output_file)
-        else:
-            log.error('Ghostscript rendering failed')
-
-
-def generate_pdfa(pdf_pages, output_file, threads=1):
-    with NamedTemporaryFile(delete=True) as gs_pdf:
-        args_gs = [
-            "gs",
-            "-dQUIET",
-            "-dBATCH",
-            "-dNOPAUSE",
-            '-dNumRenderingThreads=' + str(threads),
-            "-sDEVICE=pdfwrite",
-            "-sColorConversionStrategy=/RGB",
-            "-sProcessColorModel=DeviceRGB",
-            "-dPDFA",
-            "-sPDFACompatibilityPolicy=2",
-            "-sOutputICCProfile=srgb.icc",
-            "-sOutputFile=" + gs_pdf.name,
-        ]
-        args_gs.extend(pdf_pages)
-        check_call(args_gs)
-        copy(gs_pdf.name, output_file)

ocrmypdf/hocrtransform.py

@@ -1,230 +0,0 @@
-#!/usr/local/bin/python3
-##############################################################################
-# Copyright (c) 2013-14: fritz-hh from Github
-#   (https://github.com/fritz-hh)
-#
-# Copyright (c) 2010: Jonathan Brinley from Github
-#   (https://github.com/jbrinley/HocrConverter)
-#   Initial version by Jonathan Brinley, jonathanbrinley@gmail.com
-##############################################################################
-from reportlab.pdfgen.canvas import Canvas
-from reportlab.lib.units import inch
-from xml.etree import ElementTree
-from PIL import Image
-from collections import namedtuple
-import re
-import argparse
-
-
-Rect = namedtuple('Rect', ['x1', 'y1', 'x2', 'y2'])
-
-
-class HocrTransformError(Exception):
-    pass
-
-
-class HocrTransform():
-
-    """
-    A class for converting documents from the hOCR format.
-    For details of the hOCR format, see:
-    http://docs.google.com/View?docid=dfxcv4vc_67g844kf
-    """
-
-    def __init__(self, hocrFileName, dpi):
-        self.dpi = dpi
-        self.boxPattern = re.compile(r'bbox((\s+\d+){4})')
-
-        self.hocr = ElementTree.parse(hocrFileName)
-
-        # if the hOCR file has a namespace, ElementTree requires its use to
-        # find elements
-        matches = re.match(r'({.*})html', self.hocr.getroot().tag)
-        self.xmlns = ''
-        if matches:
-            self.xmlns = matches.group(1)
-
-        # get dimension in pt (not pixel!!!!) of the OCRed image
-        self.width, self.height = None, None
-        for div in self.hocr.findall(
-                ".//%sdiv[@class='ocr_page']" % (self.xmlns)):
-            coords = self.element_coordinates(div)
-            pt_coords = self.pt_from_pixel(coords)
-            self.width = pt_coords.x2 - pt_coords.x1
-            self.height = pt_coords.y2 - pt_coords.y1
-            # there shouldn't be more than one, and if there is, we don't want
-            # it
-            break
-        if self.width is None or self.height is None:
-            raise HocrTransformError("hocr file is missing page dimensions")
-
-    def __str__(self):
-        """
-        Return the textual content of the HTML body
-        """
-        if self.hocr is None:
-            return ''
-        body = self.hocr.find(".//%sbody" % (self.xmlns))
-        if body:
-            return self._get_element_text(body)
-        else:
-            return ''
-
-    def _get_element_text(self, element):
-        """
-        Return the textual content of the element and its children
-        """
-        text = ''
-        if element.text is not None:
-            text += element.text
-        for child in element.getchildren():
-            text += self._get_element_text(child)
-        if element.tail is not None:
-            text += element.tail
-        return text
-
-    def element_coordinates(self, element):
-        """
-        Returns a tuple containing the coordinates of the bounding box around
-        an element
-        """
-        out = (0, 0, 0, 0)
-        if 'title' in element.attrib:
-            matches = self.boxPattern.search(element.attrib['title'])
-            if matches:
-                coords = matches.group(1).split()
-                out = Rect._make(int(coords[n]) for n in range(4))
-        return out
-
-    def pt_from_pixel(self, pxl):
-        """
-        Returns the quantity in PDF units (pt) given quantity in pixels
-        """
-        return Rect._make(
-            (c / self.dpi * inch) for c in pxl)
-
-    def replace_unsupported_chars(self, s):
-        """
-        Given an input string, returns the corresponding string that:
-        - is available in the helvetica facetype
-        - does not contain any ligature (to allow easy search in the PDF file)
-        """
-        # The 'u' before the character to replace indicates that it is a
-        # unicode character
-        s = s.replace(u"fl", "fl")
-        s = s.replace(u"fi", "fi")
-        return s
-
-    def to_pdf(self, outFileName, imageFileName=None, showBoundingboxes=False,
-               fontname="Helvetica", invisibleText=False):
-        """
-        Creates a PDF file with an image superimposed on top of the text.
-        Text is positioned according to the bounding box of the lines in
-        the hOCR file.
-        The image need not be identical to the image used to create the hOCR
-        file.
-        It can have a lower resolution, different color mode, etc.
-        """
-        # create the PDF file
-        # page size in points (1/72 in.)
-        pdf = Canvas(
-            outFileName, pagesize=(self.width, self.height), pageCompression=1)
-
-        # draw bounding box for each paragraph
-        # light blue for bounding box of paragraph
-        pdf.setStrokeColorRGB(0, 1, 1)
-        # light blue for bounding box of paragraph
-        pdf.setFillColorRGB(0, 1, 1)
-        pdf.setLineWidth(0)		# no line for bounding box
-        for elem in self.hocr.findall(
-                ".//%sp[@class='%s']" % (self.xmlns, "ocr_par")):
-
-            elemtxt = self._get_element_text(elem).rstrip()
-            if len(elemtxt) == 0:
-                continue
-
-            pxl_coords = self.element_coordinates(elem)
-            pt = self.pt_from_pixel(pxl_coords)
-
-            # draw the bbox border
-            if showBoundingboxes:
-                pdf.rect(
-                    pt.x1, self.height - pt.y2, pt.x2 - pt.x1, pt.y2 - pt.y1,
-                    fill=1)
-
-        # check if element with class 'ocrx_word' are available
-        # otherwise use 'ocr_line' as fallback
-        elemclass = "ocr_line"
-        if self.hocr.find(
-                ".//%sspan[@class='ocrx_word']" % (self.xmlns)) is not None:
-            elemclass = "ocrx_word"
-
-        # itterate all text elements
-        # light green for bounding box of word/line
-        pdf.setStrokeColorRGB(1, 0, 0)
-        pdf.setLineWidth(0.5)		# bounding box line width
-        pdf.setDash(6, 3)		# bounding box is dashed
-        pdf.setFillColorRGB(0, 0, 0)  # text in black
-        for elem in self.hocr.findall(
-                ".//%sspan[@class='%s']" % (self.xmlns, elemclass)):
-
-            elemtxt = self._get_element_text(elem).rstrip()
-
-            elemtxt = self.replace_unsupported_chars(elemtxt)
-
-            if len(elemtxt) == 0:
-                continue
-
-            pxl_coords = self.element_coordinates(elem)
-            pt = self.pt_from_pixel(pxl_coords)
-
-            # draw the bbox border
-            if showBoundingboxes:
-                pdf.rect(
-                    pt.x1, self.height - pt.y2, pt.x2 - pt.x1, pt.y2 - pt.y1,
-                    fill=0)
-
-            text = pdf.beginText()
-            fontsize = pt.y2 - pt.y1
-            text.setFont(fontname, fontsize)
-            if invisibleText:
-                text.setTextRenderMode(3)  # Invisible (indicates OCR text)
-
-            # set cursor to bottom left corner of bbox (adjust for dpi)
-            text.setTextOrigin(pt.x1, self.height - pt.y2)
-
-            # scale the width of the text to fill the width of the bbox
-            text.setHorizScale(
-                100 * (pt.x2 - pt.x1) / pdf.stringWidth(
-                    elemtxt, fontname, fontsize))
-
-            # write the text to the page
-            text.textLine(elemtxt)
-            pdf.drawText(text)
-
-        # put the image on the page, scaled to fill the page
-        if imageFileName is not None:
-            pdf.drawImage(imageFileName, 0, 0,
-                          width=self.width, height=self.height)
-
-        # finish up the page and save it
-        pdf.showPage()
-        pdf.save()
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description='Convert hocr file to PDF')
-    parser.add_argument('-b', '--boundingboxes', action="store_true",
-                        default=False, help='Show bounding boxes borders')
-    parser.add_argument('-r', '--resolution', type=int,
-                        default=300,
-                        help='Resolution of the image that was OCRed')
-    parser.add_argument('-i', '--image', default=None,
-                        help='Path to the image to be placed above the text')
-    parser.add_argument('hocrfile', help='Path to the hocr file to be parsed')
-    parser.add_argument(
-        'outputfile', help='Path to the PDF file to be generated')
-    args = parser.parse_args()
-
-    hocr = HocrTransform(args.hocrfile, args.resolution)
-    hocr.to_pdf(args.outputfile, args.image, args.boundingboxes)

ocrmypdf/leptonica.py

@@ -1,331 +0,0 @@
-#!/usr/bin/env python2
-# -*- coding: utf-8 -*-
-#
-# © 2013-15: jbarlow83 from Github (https://github.com/jbarlow83)
-#
-#
-# Use Leptonica to detect find and remove page skew.  Leptonica uses the method
-# of differential square sums, which its author claim is faster and more robust
-# than the Hough transform used by ImageMagick.
-
-from __future__ import print_function, absolute_import, division
-import argparse
-import ctypes as C
-import sys
-import os
-import logging
-from tempfile import TemporaryFile
-
-logger = logging.getLogger(__name__)
-
-
-def stderr(*objs):
-    """Python 2/3 compatible print to stderr.
-    """
-    print("leptonica.py:", *objs, file=sys.stderr)
-
-
-from ctypes.util import find_library
-lept_lib = find_library('lept')
-if not lept_lib:
-    stderr("Could not find the Leptonica library")
-    sys.exit(3)
-try:
-    lept = C.cdll.LoadLibrary(lept_lib)
-except Exception:
-    stderr("Could not load the Leptonica library from %s", lept_lib)
-    sys.exit(3)
-
-
-class _PIXCOLORMAP(C.Structure):
-    """struct PixColormap from Leptonica src/pix.h
-    """
-
-    _fields_ = [
-        ("array", C.c_void_p),
-        ("depth", C.c_int32),
-        ("nalloc", C.c_int32),
-        ("n", C.c_int32)
-    ]
-
-
-class _PIX(C.Structure):
-    """struct Pix from Leptonica src/pix.h
-    """
-
-    _fields_ = [
-        ("w", C.c_uint32),
-        ("h", C.c_uint32),
-        ("d", C.c_uint32),
-        ("wpl", C.c_uint32),
-        ("refcount", C.c_uint32),
-        ("xres", C.c_int32),
-        ("yres", C.c_int32),
-        ("informat", C.c_int32),
-        ("text", C.POINTER(C.c_char)),
-        ("colormap", C.POINTER(_PIXCOLORMAP)),
-        ("data", C.POINTER(C.c_uint32))
-    ]
-
-
-PIX = C.POINTER(_PIX)
-
-lept.pixRead.argtypes = [C.c_char_p]
-lept.pixRead.restype = PIX
-lept.pixScale.argtypes = [PIX, C.c_float, C.c_float]
-lept.pixScale.restype = PIX
-lept.pixDeskew.argtypes = [PIX, C.c_int32]
-lept.pixDeskew.restype = PIX
-lept.pixFindSkew.argtypes = [PIX, C.POINTER(C.c_float), C.POINTER(C.c_float)]
-lept.pixFindSkew.restype = C.c_int32
-lept.pixWriteImpliedFormat.argtypes = [C.c_char_p, PIX, C.c_int32, C.c_int32]
-lept.pixWriteImpliedFormat.restype = C.c_int32
-lept.pixDestroy.argtypes = [C.POINTER(PIX)]
-lept.pixDestroy.restype = None
-lept.getLeptonicaVersion.argtypes = []
-lept.getLeptonicaVersion.restype = C.c_char_p
-
-
-class LeptonicaErrorTrap(object):
-    """Context manager to trap errors reported by Leptonica.
-
-    Leptonica's error return codes are unreliable to the point of being
-    almost useless.  It does, however, write errors to stderr provided that is
-    not disabled at its compile time.  Fortunately this is done using error
-    macros so it is very self-consistent.
-
-    This context manager redirects stderr to a temporary file which is then
-    read and parsed for error messages.  As a side benefit, debug messages
-    from Leptonica are also suppressed.
-
-    """
-    def __enter__(self):
-        self.tmpfile = TemporaryFile()
-
-        # Save the old stderr, and redirect stderr to temporary file
-        self.old_stderr_fileno = os.dup(sys.stderr.fileno())
-        os.dup2(self.tmpfile.fileno(), sys.stderr.fileno())
-        return
-
-    def __exit__(self, exc_type, exc_value, traceback):
-        # Restore old stderr
-        os.dup2(self.old_stderr_fileno, sys.stderr.fileno())
-
-        # Get data from tmpfile (in with block to ensure it is closed)
-        with self.tmpfile as tmpfile:
-            tmpfile.seek(0)  # Cursor will be at end, so move back to beginning
-            leptonica_output = tmpfile.read().decode(errors='replace')
-
-        # If there are Python errors, let them bubble up
-        if exc_type:
-            logger.warning(leptonica_output)
-            return False
-
-        # If there are Leptonica errors, wrap them in Python excpetions
-        if 'Error' in leptonica_output:
-            if 'image file not found' in leptonica_output:
-                raise FileNotFoundError()
-            if 'pixWrite: stream not opened' in leptonica_output:
-                raise LeptonicaIOError()
-            raise LeptonicaError(leptonica_output)
-
-        return False
-
-
-class LeptonicaError(Exception):
-    pass
-
-
-class LeptonicaIOError(LeptonicaError):
-    pass
-
-
-def pixRead(filename):
-    """Load an image file into a PIX object.
-
-    Leptonica can load TIFF, PNM (PBM, PGM, PPM), PNG, and JPEG.  If loading
-    fails then the object will wrap a C null pointer.
-
-    """
-    with LeptonicaErrorTrap():
-        return lept.pixRead(filename.encode(sys.getfilesystemencoding()))
-
-
-def pixScale(pix, scalex, scaley):
-    """Returns the pix object rescaled according to the proportions given."""
-    with LeptonicaErrorTrap():
-        return lept.pixScale(pix, scalex, scaley)
-
-
-def pixDeskew(pix, reduction_factor=0):
-    """Returns the deskewed pix object.
-
-    A clone of the original is returned when the algorithm cannot find a skew
-    angle with sufficient confidence.
-
-    reduction_factor -- amount to downsample (0 for default) when searching
-        for skew angle
-
-    """
-    with LeptonicaErrorTrap():
-        return lept.pixDeskew(pix, reduction_factor)
-
-
-def pixFindSkew(pix):
-    """Returns a tuple (deskew angle in degrees, confidence value).
-
-    Returns (None, None) if no angle is available.
-
-    """
-    with LeptonicaErrorTrap():
-        angle = C.c_float(0.0)
-        confidence = C.c_float(0.0)
-        result = lept.pixFindSkew(pix, C.byref(angle), C.byref(confidence))
-        if result == 0:
-            return (angle.value, confidence.value)
-        else:
-            return (None, None)
-
-
-def pixWriteImpliedFormat(filename, pix, jpeg_quality=0, jpeg_progressive=0):
-    """Write pix to the filename, with the extension indicating format.
-
-    jpeg_quality -- quality (iff JPEG; 1 - 100, 0 for default)
-    jpeg_progressive -- (iff JPEG; 0 for baseline seq., 1 for progressive)
-
-    """
-    fileroot, extension = os.path.splitext(filename)
-    fix_pnm = False
-    if extension.lower() in ('.pbm', '.pgm', '.ppm'):
-        # Leptonica does not process handle these extensions correctly, but
-        # does handle .pnm correctly.  Add another .pnm suffix.
-        filename += '.pnm'
-        fix_pnm = True
-
-    with LeptonicaErrorTrap():
-        lept.pixWriteImpliedFormat(
-            filename.encode(sys.getfilesystemencoding()),
-            pix, jpeg_quality, jpeg_progressive)
-
-    if fix_pnm:
-        from shutil import move
-        move(filename, filename[:-4])   # Remove .pnm suffix
-
-
-def pixDestroy(pix):
-    """Destroy the pix object.
-
-    Function signature is pixDestroy(struct Pix **), hence C.byref() to pass
-    the address of the pointer.
-
-    """
-    with LeptonicaErrorTrap():
-        lept.pixDestroy(C.byref(pix))
-
-
-def getLeptonicaVersion():
-    """Get Leptonica version string.
-
-    Caveat: Leptonica expects the caller to free this memory.  We don't,
-    since that would involve binding to libc to access libc.free(),
-    a pointless effort to reclaim 100 bytes of memory.
-
-    """
-    return lept.getLeptonicaVersion().decode()
-
-
-def deskew(infile, outfile, dpi):
-    try:
-        pix_source = pixRead(infile)
-    except LeptonicaIOError:
-        raise LeptonicaIOError("Failed to open file: %s" % infile)
-
-    if dpi < 150:
-        reduction_factor = 1  # Don't downsample too much if DPI is already low
-    else:
-        reduction_factor = 0  # Use default
-    pix_deskewed = pixDeskew(pix_source, reduction_factor)
-
-    try:
-        pixWriteImpliedFormat(outfile, pix_deskewed)
-    except LeptonicaIOError:
-        raise LeptonicaIOError("Failed to open destination file: %s" % outfile)
-    pixDestroy(pix_source)
-    pixDestroy(pix_deskewed)
-
-
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(
-        description="Python wrapper to access Leptonica")
-
-    subparsers = parser.add_subparsers(title='commands',
-                                       description='supported operations')
-
-    parser_deskew = subparsers.add_parser('deskew')
-    parser_deskew.add_argument('-r', '--dpi', dest='dpi', action='store',
-                               type=int, default=300, help='input resolution')
-    parser_deskew.add_argument('infile', help='image to deskew')
-    parser_deskew.add_argument('outfile', help='deskewed output image')
-    parser_deskew.set_defaults(func=deskew)
-
-    args = parser.parse_args()
-
-    if getLeptonicaVersion() != u'leptonica-1.69':
-        print("Unexpected leptonica version: %s" % getLeptonicaVersion())
-
-    args.func(args)
-
-
-def _test_output(mode, extension, im_format):
-    from PIL import Image
-    from tempfile import NamedTemporaryFile
-
-    with NamedTemporaryFile(prefix='test-lept-pnm', suffix=extension, delete=True) as tmpfile:
-        im = Image.new(mode=mode, size=(100, 100))
-        im.save(tmpfile)
-
-        pix = pixRead(tmpfile.name)
-        pixWriteImpliedFormat(tmpfile.name, pix)
-        pixDestroy(pix)
-
-        im_roundtrip = Image.open(tmpfile.name)
-        assert im_roundtrip.mode == im.mode, "leptonica mode differs"
-        assert im_roundtrip.format == im_format, \
-            "{0}: leptonica produced a {1}".format(
-                extension,
-                im_roundtrip.format)
-
-
-def test_pnm_output():
-    params = [['1', '.pbm', 'PPM'], ['L', '.pgm', 'PPM'],
-              ['RGB', '.ppm', 'PPM']]
-    for param in params:
-        _test_output(*param)
-
-
-def test_skew_angle():
-    from PIL import Image, ImageDraw
-    from tempfile import NamedTemporaryFile
-
-    im = Image.new(mode='1', size=(1000, 1000), color=1)
-
-    draw = ImageDraw.Draw(im)
-    for n in range(20):
-        draw.line([(50, 25 + 50*n), (950, 25 + 50*n)], width=1)
-    del draw
-
-    test_angles = [0.1 * ang for ang in range(1, 10)] + \
-                  [float(ang) for ang in range(1, 7)]
-    test_angles += [-ang for ang in test_angles]
-    test_angles = sorted(test_angles)
-
-    for rotate_angle in test_angles:
-        rotated_im = im.rotate(rotate_angle)
-        with NamedTemporaryFile(prefix='lept-skew', suffix='.png', delete=True) as tmpfile:
-            rotated_im.save(tmpfile)
-            pix = pixRead(tmpfile.name)
-            angle, confidence = pixFindSkew(pix)
-            pixDestroy(pix)
-            print('{0} {1}  {2}'.format(rotate_angle, angle, confidence), file=sys.stderr)
-
-

ocrmypdf/main.py

@@ -1,912 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from contextlib import suppress
-from tempfile import NamedTemporaryFile, mkdtemp
-import sys
-import os
-import re
-import shutil
-import warnings
-import multiprocessing
-import atexit
-import textwrap
-
-import PyPDF2 as pypdf
-from PIL import Image
-
-from subprocess import Popen, check_call, PIPE, CalledProcessError, \
-    TimeoutExpired, check_output, STDOUT
-try:
-    from subprocess import DEVNULL
-except ImportError:
-    DEVNULL = open(os.devnull, 'wb')
-
-
-from ruffus import transform, suffix, merge, active_if, regex, jobs_limit, \
-    formatter, follows, split, collate, check_if_uptodate
-import ruffus.ruffus_exceptions as ruffus_exceptions
-import ruffus.cmdline as cmdline
-
-from .hocrtransform import HocrTransform
-from .pageinfo import pdf_get_all_pageinfo
-from .pdfa import generate_pdfa_def
-from . import ghostscript
-from . import tesseract
-from . import ExitCode
-
-warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
-
-
-BASEDIR = os.path.dirname(os.path.realpath(__file__))
-JHOVE_PATH = os.path.realpath(os.path.join(BASEDIR, 'jhove'))
-JHOVE_JAR = os.path.join(JHOVE_PATH, 'bin', 'JhoveApp.jar')
-JHOVE_CFG = os.path.join(JHOVE_PATH, 'conf', 'jhove.conf')
-
-
-# -------------
-# External dependencies
-
-MINIMUM_TESS_VERSION = '3.02.02'
-
-
-def complain(message):
-    print(*textwrap.wrap(message), file=sys.stderr)
-
-
-if tesseract.version() < MINIMUM_TESS_VERSION:
-    complain(
-        "Please install tesseract {0} or newer "
-        "(currently installed version is {1})".format(
-            MINIMUM_TESS_VERSION, tesseract.version()))
-    sys.exit(ExitCode.missing_dependency)
-
-
-try:
-    import PIL.features
-    check_codec = PIL.features.check_codec
-except (ImportError, AttributeError):
-    def check_codec(codec_name):
-        if codec_name == 'jpg':
-            return 'jpeg_encoder' in dir(Image.core)
-        elif codec_name == 'zlib':
-            return 'zip_encoder' in dir(Image.core)
-        raise NotImplementedError(codec_name)
-
-
-def check_pil_encoder(codec_name, friendly_name):
-    try:
-        if check_codec(codec_name):
-            return
-    except Exception:
-        pass
-    complain(
-        "ERROR: Your version of the Python imaging library (Pillow) was "
-        "compiled without support for " + friendly_name + " encoding/decoding."
-        "\n"
-        "You will need to uninstall Pillow and reinstall it with PNG and JPEG "
-        "support (libjpeg and zlib)."
-        "\n"
-        "See installation instructions for your platform here:\n"
-        "    https://pillow.readthedocs.org/installation.html"
-    )
-    sys.exit(ExitCode.missing_dependency)
-
-
-check_pil_encoder('jpg', 'JPEG')
-check_pil_encoder('zlib', 'PNG')
-
-
-# -------------
-# Parser
-
-parser = cmdline.get_argparse(
-    prog="ocrmypdf",
-    description="Generate searchable PDF file from an image-only PDF file.",
-    version='3.0',
-    fromfile_prefix_chars='@',
-    ignored_args=[
-        'touch_files_only', 'recreate_database', 'checksum_file_name',
-        'key_legend_in_graph', 'draw_graph_horizontally', 'flowchart_format',
-        'forced_tasks', 'target_tasks'])
-
-parser.add_argument(
-    'input_file',
-    help="PDF file containing the images to be OCRed")
-parser.add_argument(
-    'output_file',
-    help="output searchable PDF file")
-parser.add_argument(
-    '-l', '--language', action='append',
-    help="languages of the file to be OCRed")
-
-metadata = parser.add_argument_group(
-    "Metadata options",
-    "Set output PDF/A metadata (default: use input document's title)")
-metadata.add_argument(
-    '--title', type=str,
-    help="set document title (place multiple words in quotes)")
-metadata.add_argument(
-    '--author', type=str,
-    help="set document author")
-metadata.add_argument(
-    '--subject', type=str,
-    help="set document")
-metadata.add_argument(
-    '--keywords', type=str,
-    help="set document keywords")
-
-
-preprocessing = parser.add_argument_group(
-    "Preprocessing options",
-    "Improve OCR quality and final image")
-preprocessing.add_argument(
-    '-d', '--deskew', action='store_true',
-    help="deskew each page before performing OCR")
-preprocessing.add_argument(
-    '-c', '--clean', action='store_true',
-    help="clean pages from scanning artifacts before performing OCR")
-preprocessing.add_argument(
-    '-i', '--clean-final', action='store_true',
-    help="incorporate the cleaned image in the final PDF file")
-preprocessing.add_argument(
-    '--oversample', metavar='DPI', type=int, default=0,
-    help="oversample images to at least the specified DPI, to improve OCR "
-         "results slightly")
-
-parser.add_argument(
-    '-f', '--force-ocr', action='store_true',
-    help="rasterize any fonts or vector images on each page and apply OCR")
-parser.add_argument(
-    '-s', '--skip-text', action='store_true',
-    help="skip OCR on any pages that already contain text, but include the"
-         " page in final output")
-parser.add_argument(
-    '--skip-big', type=float, metavar='MPixels',
-    help="skip OCR on pages larger than the specified amount of megapixels, "
-         "but include skipped pages in final output")
-# parser.add_argument(
-#     '--exact-image', action='store_true',
-#     help="Use original page from PDF without re-rendering")
-
-advanced = parser.add_argument_group(
-    "Advanced",
-    "Advanced options for power users")
-advanced.add_argument(
-    '--tesseract-config', default=[], type=list, action='append',
-    help="additional Tesseract configuration files")
-advanced.add_argument(
-    '--pdf-renderer', choices=['tesseract', 'hocr'], default='hocr',
-    help='choose OCR PDF renderer')
-advanced.add_argument(
-    '--tesseract-timeout', default=180.0, type=float,
-    help='give up on OCR after the timeout, but copy the preprocessed page '
-         'into the final output')
-
-debugging = parser.add_argument_group(
-    "Debugging",
-    "Arguments to help with troubleshooting and debugging")
-debugging.add_argument(
-    '-k', '--keep-temporary-files', action='store_true',
-    help="keep temporary files (helpful for debugging)")
-debugging.add_argument(
-    '-g', '--debug-rendering', action='store_true',
-    help="render each page twice with debug information on second page")
-
-options = parser.parse_args()
-
-
-# ----------
-# Languages
-
-if not options.language:
-    options.language = ['eng']  # Enforce English hegemony
-
-# Support v2.x "eng+deu" language syntax
-if '+' in options.language[0]:
-    options.language = options.language[0].split('+')
-
-if not set(options.language).issubset(tesseract.languages()):
-    complain(
-        "The installed version of tesseract does not have language "
-        "data for the following requested languages: ")
-    for lang in (set(options.language) - tesseract.languages()):
-        complain(lang)
-    sys.exit(ExitCode.bad_args)
-
-
-# ----------
-# Arguments
-
-
-if any((options.deskew, options.clean, options.clean_final)):
-    try:
-        from . import unpaper
-    except ImportError:
-        complain(
-            "Install the 'unpaper' program to use --deskew or --clean.")
-        sys.exit(ExitCode.bad_args)
-else:
-    unpaper = None
-
-if options.debug_rendering and options.pdf_renderer == 'tesseract':
-    complain(
-        "Ignoring --debug-rendering because it is not supported with"
-        "--pdf-renderer=tesseract.")
-
-if options.force_ocr and options.skip_text:
-    complain(
-        "Error: --force-ocr and --skip-text are mutually incompatible.")
-    sys.exit(ExitCode.bad_args)
-
-if options.clean and not options.clean_final \
-        and options.pdf_renderer == 'tesseract':
-    complain(
-        "Tesseract PDF renderer cannot render --clean pages without "
-        "also performing --clean-final, so --clean-final is assumed.")
-
-
-# ----------
-# Logging
-
-
-_logger, _logger_mutex = cmdline.setup_logging(__name__, options.log_file,
-                                               options.verbose)
-
-
-class WrappedLogger:
-
-    def __init__(self, my_logger, my_mutex):
-        self.logger = my_logger
-        self.mutex = my_mutex
-
-    def log(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.log(*args, **kwargs)
-
-    def debug(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.debug(*args, **kwargs)
-
-    def info(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.info(*args, **kwargs)
-
-    def warning(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.warning(*args, **kwargs)
-
-    def error(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.error(*args, **kwargs)
-
-    def critical(self, *args, **kwargs):
-        with self.mutex:
-            self.logger.critical(*args, **kwargs)
-
-_log = WrappedLogger(_logger, _logger_mutex)
-
-
-def re_symlink(input_file, soft_link_name, log=_log):
-    """
-    Helper function: relinks soft symbolic link if necessary
-    """
-    # Guard against soft linking to oneself
-    if input_file == soft_link_name:
-        log.debug("Warning: No symbolic link made. You are using " +
-                  "the original data directory as the working directory.")
-        return
-
-    # Soft link already exists: delete for relink?
-    if os.path.lexists(soft_link_name):
-        # do not delete or overwrite real (non-soft link) file
-        if not os.path.islink(soft_link_name):
-            raise Exception("%s exists and is not a link" % soft_link_name)
-        try:
-            os.unlink(soft_link_name)
-        except:
-            log.debug("Can't unlink %s" % (soft_link_name))
-
-    if not os.path.exists(input_file):
-        raise Exception("trying to create a broken symlink to %s" % input_file)
-
-    log.debug("os.symlink(%s, %s)" % (input_file, soft_link_name))
-
-    # Create symbolic link using absolute path
-    os.symlink(
-        os.path.abspath(input_file),
-        soft_link_name
-    )
-
-
-# -------------
-# The Pipeline
-
-manager = multiprocessing.Manager()
-_pdfinfo = manager.list()
-_pdfinfo_lock = manager.Lock()
-
-work_folder = mkdtemp(prefix="com.github.ocrmypdf.")
-
-
-@atexit.register
-def cleanup_working_files(*args):
-    if options.keep_temporary_files:
-        print("Temporary working files saved at:")
-        print(work_folder)
-    else:
-        with suppress(FileNotFoundError):
-            shutil.rmtree(work_folder)
-
-
-@transform(
-    input=options.input_file,
-    filter=suffix('.pdf'),
-    output='.repaired.pdf',
-    output_dir=work_folder,
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def repair_pdf(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-    args_qpdf = [
-        'qpdf', input_file, output_file
-    ]
-    try:
-        out = check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 2:
-            print("{0}: not a valid PDF, and could not repair it.".format(
-                    options.input_file))
-            print("Details:")
-            print(e.output)
-        else:
-            print(e.output)
-        sys.exit(ExitCode.input_file)
-
-    log.debug(out)
-
-    with pdfinfo_lock:
-        pdfinfo.extend(pdf_get_all_pageinfo(output_file))
-        log.info(pdfinfo)
-
-
-def get_pageinfo(input_file, pdfinfo, pdfinfo_lock):
-    pageno = int(os.path.basename(input_file)[0:6]) - 1
-    with pdfinfo_lock:
-        pageinfo = pdfinfo[pageno].copy()
-    return pageinfo
-
-
-def is_ocr_required(pageinfo, log):
-    page = pageinfo['pageno'] + 1
-    ocr_required = True
-    if not pageinfo['images']:
-        # If the page has no images, then it contains vector content or text
-        # or both. It seems quite unlikely that one would find meaningful text
-        # from rasterizing vector content. So skip the page.
-        log.info(
-            "Page {0} has no images - skipping OCR".format(page)
-        )
-        ocr_required = False
-    elif pageinfo['has_text']:
-        s = "Page {0} already has text! – {1}"
-
-        if not options.force_ocr and not options.skip_text:
-            log.error(s.format(page,
-                               "aborting (use --force-ocr to force OCR)"))
-            sys.exit(ExitCode.already_done_ocr)
-        elif options.force_ocr:
-            log.info(s.format(page,
-                              "rasterizing text and running OCR anyway"))
-            ocr_required = True
-        elif options.skip_text:
-            log.info(s.format(page,
-                              "skipping all processing on this page"))
-            ocr_required = False
-
-    if ocr_required and options.skip_big:
-        pixel_count = pageinfo['width_pixels'] * pageinfo['height_pixels']
-        if pixel_count > (options.skip_big * 1000000):
-            ocr_required = False
-            log.info(
-                "Page {0} is very large; skipping due to -b".format(page))
-
-    return ocr_required
-
-
-@split(
-    repair_pdf,
-    os.path.join(work_folder, '*.page.pdf'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def split_pages(
-        input_file,
-        output_files,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    for oo in output_files:
-        with suppress(FileNotFoundError):
-            os.unlink(oo)
-
-    pages = check_output(['qpdf', '--show-npages', input_file],
-                         universal_newlines=True, close_fds=True)
-
-    for n in range(int(pages)):
-        args_qpdf = [
-            'qpdf', input_file,
-            '--pages', input_file, '{0}'.format(n + 1), '--',
-            os.path.join(work_folder, '{0:06d}.page.pdf'.format(n + 1))
-        ]
-        check_call(args_qpdf)
-
-    from glob import glob
-    for filename in glob(os.path.join(work_folder, '*.page.pdf')):
-        pageinfo = get_pageinfo(filename, pdfinfo, pdfinfo_lock)
-
-        alt_suffix = '.ocr.page.pdf' if is_ocr_required(pageinfo, log) \
-                     else '.skip.page.pdf'
-        re_symlink(
-            filename,
-            os.path.join(
-                work_folder,
-                os.path.basename(filename)[0:6] + alt_suffix))
-
-
-@transform(
-    input=split_pages,
-    filter=suffix('.ocr.page.pdf'),
-    output='.page.png',
-    output_dir=work_folder,
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def rasterize_with_ghostscript(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-
-    device = 'png16m'  # 24-bit
-    if all(image['comp'] == 1 for image in pageinfo['images']):
-        if all(image['bpc'] == 1 for image in pageinfo['images']):
-            device = 'pngmono'
-        elif all(image['bpc'] > 1 and image['color'] == 'index'
-                 for image in pageinfo['images']):
-            device = 'png256'
-        elif all(image['bpc'] > 1 and image['color'] == 'gray'
-                 for image in pageinfo['images']):
-            device = 'pnggray'
-
-    log.debug("Rendering {0} with {1}".format(
-            os.path.basename(input_file), device))
-    xres = max(pageinfo['xres'], options.oversample or 0)
-    yres = max(pageinfo['yres'], options.oversample or 0)
-
-    ghostscript.rasterize_pdf(input_file, output_file, xres, yres, device, log)
-
-
-@transform(
-    input=rasterize_with_ghostscript,
-    filter=suffix(".page.png"),
-    output=".pp-deskew.png",
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def preprocess_deskew(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    if not options.deskew:
-        re_symlink(input_file, output_file, log)
-        return
-
-    pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-    dpi = int(pageinfo['xres'])
-
-    unpaper.deskew(input_file, output_file, dpi, log)
-
-
-@transform(
-    input=preprocess_deskew,
-    filter=suffix(".pp-deskew.png"),
-    output=".pp-clean.png",
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def preprocess_clean(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    if not options.clean:
-        re_symlink(input_file, output_file, log)
-        return
-
-    pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-    dpi = int(pageinfo['xres'])
-
-    unpaper.clean(input_file, output_file, dpi, log)
-
-
-@active_if(options.pdf_renderer == 'hocr')
-@transform(
-    input=preprocess_clean,
-    filter=suffix(".pp-clean.png"),
-    output=".hocr",
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def ocr_tesseract_hocr(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-
-    badxml = os.path.splitext(output_file)[0] + '.badxml'
-
-    args_tesseract = [
-        'tesseract',
-        '-l', '+'.join(options.language),
-        input_file,
-        badxml,
-        'hocr'
-    ] + options.tesseract_config
-    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
-              universal_newlines=True)
-    try:
-        stdout, stderr = p.communicate(timeout=options.tesseract_timeout)
-    except TimeoutExpired:
-        p.kill()
-        stdout, stderr = p.communicate()
-        # Generate a HOCR file with no recognized text if tesseract times out
-        # Temporary workaround to hocrTransform not being able to function if
-        # it does not have a valid hOCR file.
-        with open(output_file, 'w', encoding="utf-8") as f:
-            f.write(tesseract.HOCR_TEMPLATE.format(
-                pageinfo['width_pixels'],
-                pageinfo['height_pixels']))
-    else:
-        if stdout:
-            log.info(stdout)
-        if stderr:
-            log.error(stderr)
-
-        if p.returncode != 0:
-            raise CalledProcessError(p.returncode, args_tesseract)
-
-        if os.path.exists(badxml + '.html'):
-            # Tesseract 3.02 appends suffix ".html" on its own (.badxml.html)
-            shutil.move(badxml + '.html', badxml)
-        elif os.path.exists(badxml + '.hocr'):
-            # Tesseract 3.03 appends suffix ".hocr" on its own (.badxml.hocr)
-            shutil.move(badxml + '.hocr', badxml)
-
-        # Tesseract 3.03 inserts source filename into hocr file without
-        # escaping it, creating invalid XML and breaking the parser.
-        # As a workaround, rewrite the hocr file, replacing the filename
-        # with a space.  Don't know if Tesseract 3.02 does the same.
-
-        regex_nested_single_quotes = re.compile(
-            r"""title='image "([^"]*)";""")
-        with open(badxml, mode='r', encoding='utf-8') as f_in, \
-                open(output_file, mode='w', encoding='utf-8') as f_out:
-            for line in f_in:
-                line = regex_nested_single_quotes.sub(
-                    r"""title='image " ";""", line)
-                f_out.write(line)
-
-
-@active_if(options.pdf_renderer == 'hocr')
-@collate(
-    input=[rasterize_with_ghostscript, preprocess_deskew, preprocess_clean],
-    filter=regex(r".*/(\d{6})(?:\.page|\.pp-deskew|\.pp-clean)\.png"),
-    output=os.path.join(work_folder, r'\1.image'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def select_image_for_pdf(
-        infiles,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-    if options.clean_final:
-        image_suffix = '.pp-clean.png'
-    elif options.deskew:
-        image_suffix = '.pp-deskew.png'
-    else:
-        image_suffix = '.page.png'
-    image = next(ii for ii in infiles if ii.endswith(image_suffix))
-
-    pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-    if all(image['enc'] == 'jpeg' for image in pageinfo['images']):
-        # If all images were JPEGs originally, produce a JPEG as output
-        Image.open(image).save(output_file, format='JPEG')
-    else:
-        re_symlink(image, output_file)
-
-
-@active_if(options.pdf_renderer == 'hocr')
-@collate(
-    input=[select_image_for_pdf, ocr_tesseract_hocr],
-    filter=regex(r".*/(\d{6})(?:\.image|\.hocr)"),
-    output=os.path.join(work_folder, r'\1.rendered.pdf'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def render_hocr_page(
-        infiles,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-    hocr = next(ii for ii in infiles if ii.endswith('.hocr'))
-    image = next(ii for ii in infiles if ii.endswith('.image'))
-
-    pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-    dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
-
-    hocrtransform = HocrTransform(hocr, dpi)
-    hocrtransform.to_pdf(output_file, imageFileName=image,
-                         showBoundingboxes=False, invisibleText=True)
-
-
-@active_if(options.pdf_renderer == 'hocr')
-@active_if(options.debug_rendering)
-@collate(
-    input=[select_image_for_pdf, ocr_tesseract_hocr],
-    filter=regex(r".*/(\d{6})(?:\.image|\.hocr)"),
-    output=os.path.join(work_folder, r'\1.debug.pdf'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def render_hocr_debug_page(
-        infiles,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-    hocr = next(ii for ii in infiles if ii.endswith('.hocr'))
-    image = next(ii for ii in infiles if ii.endswith('.image'))
-
-    pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-    dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
-
-    hocrtransform = HocrTransform(hocr, dpi)
-    hocrtransform.to_pdf(output_file, imageFileName=None,
-                         showBoundingboxes=True, invisibleText=False)
-
-
-@active_if(options.pdf_renderer == 'tesseract')
-@collate(
-    input=[preprocess_clean, split_pages],
-    filter=regex(r".*/(\d{6})(?:\.pp-clean\.png|\.page\.pdf)"),
-    output=os.path.join(work_folder, r'\1.rendered.pdf'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def tesseract_ocr_and_render_pdf(
-        input_files,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    input_image = next((ii for ii in input_files if ii.endswith('.png')), '')
-    input_pdf = next((ii for ii in input_files if ii.endswith('.pdf')))
-    if not input_image:
-        # Skipping this page
-        re_symlink(input_pdf, output_file)
-        return
-
-    args_tesseract = [
-        'tesseract',
-        '-l', '+'.join(options.language),
-        input_image,
-        os.path.splitext(output_file)[0],  # Tesseract appends suffix
-        'pdf'
-    ] + options.tesseract_config
-    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
-              universal_newlines=True)
-
-    try:
-        stdout, stderr = p.communicate(timeout=options.tesseract_timeout)
-        if stdout:
-            log.info(stdout)
-        if stderr:
-            log.error(stderr)
-    except TimeoutExpired:
-        p.kill()
-        log.info("Tesseract - page timed out")
-        re_symlink(input_pdf, output_file)
-
-
-@transform(
-    input=repair_pdf,
-    filter=suffix('.repaired.pdf'),
-    output='.pdfa_def.ps',
-    output_dir=work_folder,
-    extras=[_log])
-def generate_postscript_stub(
-        input_file,
-        output_file,
-        log):
-
-    pdf = pypdf.PdfFileReader(input_file)
-
-    def from_document_info(key):
-        # pdf.documentInfo.get() DOES NOT behave as expected for a dict-like
-        # object, so call with precautions.  TypeError may occur if the PDF
-        # is missing the optional document info section.
-        try:
-            s = pdf.documentInfo[key]
-            return str(s)
-        except (KeyError, TypeError):
-            return ''
-
-    pdfmark = {
-        'title': from_document_info('/Title'),
-        'author': from_document_info('/Author'),
-        'keywords': from_document_info('/Keywords'),
-        'subject': from_document_info('/Subject'),
-    }
-    if options.title:
-        pdfmark['title'] = options.title
-    if options.author:
-        pdfmark['author'] = options.author
-    if options.keywords:
-        pdfmark['keywords'] = options.keywords
-    if options.subject:
-        pdfmark['subject'] = options.subject
-
-    generate_pdfa_def(output_file, pdfmark)
-
-
-@transform(
-    input=split_pages,
-    filter=suffix('.skip.page.pdf'),
-    output='.done.pdf',
-    output_dir=work_folder,
-    extras=[_log])
-def skip_page(
-        input_file,
-        output_file,
-        log):
-    re_symlink(input_file, output_file, log)
-
-
-@merge(
-    input=[render_hocr_page, render_hocr_debug_page, skip_page,
-           tesseract_ocr_and_render_pdf, generate_postscript_stub],
-    output=os.path.join(work_folder, 'merged.pdf'),
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def merge_pages(
-        input_files,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-
-    def input_file_order(s):
-        '''Sort order: All rendered pages followed
-        by their debug page, if any, followed by Postscript stub.
-        Ghostscript documentation has the Postscript stub at the
-        beginning, but it works at the end and also gets document info
-        right that way.'''
-        if s.endswith('.ps'):
-            return 99999999
-        key = int(os.path.basename(s)[0:6]) * 10
-        if 'debug' in os.path.basename(s):
-            key += 1
-        return key
-
-    pdf_pages = sorted(input_files, key=input_file_order)
-    log.info(pdf_pages)
-    ghostscript.generate_pdfa(pdf_pages, output_file, options.jobs or 1)
-
-
-@transform(
-    input=merge_pages,
-    filter=formatter(),
-    output=options.output_file,
-    extras=[_log, _pdfinfo, _pdfinfo_lock])
-def copy_final(
-        input_file,
-        output_file,
-        log,
-        pdfinfo,
-        pdfinfo_lock):
-    shutil.copy(input_file, output_file)
-
-
-def validate_pdfa(
-        input_file,
-        log):
-
-    args_qpdf = [
-        'qpdf',
-        '--check',
-        input_file
-    ]
-
-    try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 2:
-            print("{0}: not a valid PDF, and could not repair it.".format(
-                    options.input_file))
-            print("Details:")
-            print(e.output)
-        elif e.returncode == 3:
-            log.info("qpdf --check returned warnings:")
-            log.info(e.output)
-        else:
-            print(e.output)
-        return False
-
-    return True
-
-
-# @active_if(ocr_required and options.exact_image)
-# @merge([render_hocr_blank_page, extract_single_page],
-#        os.path.join(work_folder, "%04i.merged.pdf") % pageno)
-# def merge_hocr_with_original_page(infiles, output_file):
-#     with open(infiles[0], 'rb') as hocr_input, \
-#             open(infiles[1], 'rb') as page_input, \
-#             open(output_file, 'wb') as output:
-#         hocr_reader = pypdf.PdfFileReader(hocr_input)
-#         page_reader = pypdf.PdfFileReader(page_input)
-#         writer = pypdf.PdfFileWriter()
-
-#         the_page = hocr_reader.getPage(0)
-#         the_page.mergePage(page_reader.getPage(0))
-#         writer.addPage(the_page)
-#         writer.write(output)
-
-
-def available_cpu_count():
-    try:
-        return multiprocessing.cpu_count()
-    except NotImplementedError:
-        pass
-
-    try:
-        import psutil
-        return psutil.cpu_count()
-    except (ImportError, AttributeError):
-        pass
-
-    complain(
-        "Could not get CPU count.  Assuming one (1) CPU."
-        "Use -j N to set manually.")
-    return 1
-
-
-def run_pipeline():
-    if not options.jobs or options.jobs == 1:
-        options.jobs = available_cpu_count()
-
-    try:
-        cmdline.run(options)
-    except ruffus_exceptions.RethrownJobError as e:
-        if options.verbose:
-            print(e)
-
-        # Yuck. Hunt through the ruffus exception to find out what the
-        # return code is supposed to be.
-        for exc in e.args:
-            task_name, job_name, exc_name, exc_value, exc_stack = exc
-            if exc_name == 'builtins.SystemExit':
-                return eval(
-                    exc_value,
-                    {'ExitCode': ExitCode}, {'exc_value': exc_value})
-        return ExitCode.other_error
-
-    if not validate_pdfa(options.output_file, _log):
-        _log.warning('Output file: The generated PDF/A file is INVALID')
-        return ExitCode.invalid_output_pdfa
-
-    return ExitCode.ok
-
-
-if __name__ == '__main__':
-    sys.exit(run_pipeline())

ocrmypdf/pageinfo.py

@@ -1,145 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from subprocess import Popen, PIPE
-from decimal import Decimal, getcontext
-import re
-import sys
-import PyPDF2 as pypdf
-
-
-FRIENDLY_COLORSPACE = {
-    '/DeviceGray': 'gray',
-    '/CalGray': 'gray',
-    '/DeviceRGB': 'rgb',
-    '/CalRGB': 'rgb',
-    '/DeviceCMYK': 'cmyk',
-    '/Lab': 'lab',
-    '/ICCBased': 'icc',
-    '/Indexed': 'index',
-    '/Separation': 'sep',
-    '/DeviceN': 'devn',
-    '/Pattern': '-'
-}
-
-FRIENDLY_ENCODING = {
-    '/CCITTFaxDecode': 'ccitt',
-    '/DCTDecode': 'jpeg',
-    '/JPXDecode': 'jpx',
-    '/JBIG2Decode': 'jbig2',
-}
-
-FRIENDLY_COMP = {
-    'gray': 1,
-    'rgb': 3,
-    'cmyk': 4,
-    'lab': 3,
-    'index': 1
-}
-
-
-def _page_has_inline_images(page):
-    # PDF always uses \r\n for separator regardless of platform
-    # Really basic heuristic that might trigger the odd false positive
-    # This is only finds the first image and is not quite spec compliant
-    try:
-        contents = page.getContents()
-        data = contents.getData()
-    except AttributeError:
-        # If we can't access the contents or data (empty page?) then there
-        # are no inline images
-        return False
-
-    begin_image, image_data, end_image = False, False, False
-    for data in re.split(b'\s+', data):
-        if data == b'BI':
-            begin_image = True
-        elif data == b'ID':
-            image_data = True
-        elif data == b'EI':
-            end_image = True
-        if all((begin_image, image_data, end_image)):
-            return True
-    return False
-
-
-def _find_page_images(page, pageinfo):
-    try:
-        page['/Resources']['/XObject']
-    except KeyError:
-        return
-
-    # Look for XObject (out of line images)
-    for xobj in page['/Resources']['/XObject']:
-        # PyPDF2 returns the keys as an iterator
-        pdfimage = page['/Resources']['/XObject'][xobj]
-        if pdfimage['/Subtype'] != '/Image':
-            continue
-        if '/ImageMask' in pdfimage:
-            if pdfimage['/ImageMask']:
-                continue
-        image = {}
-        image['width'] = pdfimage['/Width']
-        image['height'] = pdfimage['/Height']
-        image['bpc'] = pdfimage['/BitsPerComponent']
-        if '/Filter' in pdfimage:
-            filter_ = pdfimage['/Filter']
-            if isinstance(filter_, pypdf.generic.ArrayObject):
-                filter_ = filter_[0]
-            image['enc'] = FRIENDLY_ENCODING.get(filter_, 'image')
-        else:
-            image['enc'] = 'image'
-        if '/ColorSpace' in pdfimage:
-            cs = pdfimage['/ColorSpace']
-            if isinstance(cs, pypdf.generic.ArrayObject):
-                cs = cs[0]
-            image['color'] = FRIENDLY_COLORSPACE.get(cs, '-')
-        else:
-            image['color'] = 'jpx' if image['enc'] == 'jpx' else '?'
-
-        image['comp'] = FRIENDLY_COMP.get(image['color'], '?')
-        image['dpi_w'] = image['width'] / pageinfo['width_inches']
-        image['dpi_h'] = image['height'] / pageinfo['height_inches']
-        image['dpi'] = (image['dpi_w'] * image['dpi_h']) ** Decimal(0.5)
-        yield image
-
-
-def _pdf_get_pageinfo(infile, page: int):
-    pageinfo = {}
-    pageinfo['pageno'] = page
-    pageinfo['images'] = []
-
-    pdf = pypdf.PdfFileReader(infile)
-    page = pdf.pages[page - 1]
-
-    text = page.extractText()
-    pageinfo['has_text'] = (text.strip() != '')
-
-    width_pt = page['/MediaBox'][2] - page['/MediaBox'][0]
-    height_pt = page['/MediaBox'][3] - page['/MediaBox'][1]
-    pageinfo['width_inches'] = width_pt / Decimal(72.0)
-    pageinfo['height_inches'] = height_pt / Decimal(72.0)
-
-    pageinfo['images'] = [im for im in _find_page_images(page, pageinfo)]
-
-    # Look for inline images
-    if _page_has_inline_images(page):
-        raise NotImplementedError(
-            "Warning: input PDF contains inline images - not supported")
-
-    if pageinfo['images']:
-        xres = max(image['dpi_w'] for image in pageinfo['images'])
-        yres = max(image['dpi_h'] for image in pageinfo['images'])
-        pageinfo['xres'], pageinfo['yres'] = xres, yres
-        pageinfo['width_pixels'] = \
-            int(round(xres * pageinfo['width_inches']))
-        pageinfo['height_pixels'] = \
-            int(round(yres * pageinfo['height_inches']))
-
-    return pageinfo
-
-
-def pdf_get_all_pageinfo(infile):
-    pdf = pypdf.PdfFileReader(infile)
-    getcontext().prec = 6
-    return [_pdf_get_pageinfo(infile, n) for n in range(pdf.numPages)]

ocrmypdf/pdfa.py

@@ -1,132 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-#
-# Generate a PDFA_def.ps file for Ghostscript >= 9.14
-
-from __future__ import print_function, absolute_import, division
-from string import Template
-from subprocess import Popen, PIPE
-import os
-import codecs
-
-
-# This is a template written in PostScript which is needed to create PDF/A
-# files, from the Ghostscript documentation. Lines beginning with % are
-# comments. Python substitution variables have a '$' prefix.
-pdfa_def_template = u"""%!
-% This is a sample prefix file for creating a PDF/A document.
-% Feel free to modify entries marked with "Customize".
-% This assumes an ICC profile to reside in the file (ISO Coated sb.icc),
-% unless the user modifies the corresponding line below.
-
-% Define entries in the document Info dictionary :
-/ICCProfile ($icc_profile)
-def
-
-[ /Title <$title>
-  /Author <$author>
-  /Subject <$subject>
-  /Keywords <$keywords>
-  /DOCINFO pdfmark
-
-% Define an ICC profile :
-
-[/_objdef {icc_PDFA} /type /stream /OBJ pdfmark
-[{icc_PDFA}
-<<
-  /N currentpagedevice /ProcessColorModel known {
-    currentpagedevice /ProcessColorModel get dup /DeviceGray eq
-    {pop 1} {
-      /DeviceRGB eq
-      {3}{4} ifelse
-    } ifelse
-  } {
-    (ERROR, unable to determine ProcessColorModel) == flush
-  } ifelse
->> /PUT pdfmark
-[{icc_PDFA} ICCProfile (r) file /PUT pdfmark
-
-% Define the output intent dictionary :
-
-[/_objdef {OutputIntent_PDFA} /type /dict /OBJ pdfmark
-[{OutputIntent_PDFA} <<
-  /Type /OutputIntent             % Must be so (the standard requires).
-  /S /GTS_PDFA1                   % Must be so (the standard requires).
-  /DestOutputProfile {icc_PDFA}            % Must be so (see above).
-  /OutputConditionIdentifier ($icc_identifier)
->> /PUT pdfmark
-[{Catalog} <</OutputIntents [ {OutputIntent_PDFA} ]>> /PUT pdfmark
-"""
-
-
-def encode_text_string(s: str) -> str:
-    '''Encode text string to hex string for use in a PDF
-
-    From PDF 32000-1:2008 a string object may be included in hexademical form
-    if it is enclosed in angle brackets.  For general Unicode the string should
-    be UTF-16 (big endian) with byte order marks.  A non-hexademical
-    representation is doable but this is preferable since it allows the output
-    Postscript file to be completely ASCII and no escaping of Postscript
-    characters is necessary.
-    '''
-    if s == '':
-        return ''
-    utf16_bytes = s.encode('utf-16be')
-    ascii_hex_bytes = codecs.encode(b'\xfe\xff' + utf16_bytes, 'hex')
-    ascii_hex_str = ascii_hex_bytes.decode('ascii').lower()
-    return ascii_hex_str
-
-
-def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
-    pdfmark_utf16 = {k: encode_text_string(v) for k, v in pdfmark.items()}
-
-    t = Template(pdfa_def_template)
-    result = t.substitute(icc_profile=icc_profile,
-                          icc_identifier=icc_identifier,
-                          title=pdfmark_utf16.get('title', ''),
-                          author=pdfmark_utf16.get('author', ''),
-                          subject=pdfmark_utf16.get('subject', ''),
-                          keywords=pdfmark_utf16.get('keywords', ''))
-    return result
-
-
-def _get_postscript_icc_path():
-    "Parse Ghostscript's help message to find where iccprofiles are stored"
-
-    p_gs = Popen(['gs', '--help'], close_fds=True, universal_newlines=True,
-                 stdout=PIPE, stderr=PIPE)
-    out, _ = p_gs.communicate()
-    lines = out.splitlines()
-
-    def search_paths(lines):
-        seeking = True
-        for line in lines:
-            if seeking:
-                if line.startswith('Search path'):
-                    seeking = False
-                    continue
-            else:
-                if line.strip().startswith('/'):
-                    yield from (
-                        path.strip() for path in line.split(':')
-                        if path.strip() != '')
-    for root in search_paths(lines):
-        path = os.path.realpath(os.path.join(root, '../iccprofiles'))
-        if os.path.exists(path):
-            return path
-
-    raise FileNotFoundError("Could not find Ghostscript's iccprofiles")
-
-
-def generate_pdfa_def(target_filename, pdfmark, icc='sRGB'):
-    if icc == 'sRGB':
-        icc_profile = os.path.join(_get_postscript_icc_path(), 'srgb.icc')
-    else:
-        raise NotImplementedError("Only supporting sRGB")
-
-    ps = _get_pdfa_def(icc_profile, icc, pdfmark)
-
-    # Since PostScript might not handle UTF-8 (it's hard to get a clear
-    # answer), insist on ascii
-    with open(target_filename, 'w', encoding='ascii') as f:
-        f.write(ps)

ocrmypdf/tesseract.py

@@ -1,69 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from subprocess import STDOUT, CalledProcessError, check_output
-import sys
-import os
-import re
-from functools import lru_cache
-from . import ExitCode
-
-
-@lru_cache(maxsize=1)
-def version():
-    args_tess = [
-        'tesseract',
-        '--version'
-    ]
-    try:
-        versions = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
-    except CalledProcessError:
-        print("Could not find Tesseract executable on system PATH.")
-        sys.exit(ExitCode.missing_dependency)
-
-    tesseract_version = re.match(r'tesseract\s(.+)', versions).group(1)
-    return tesseract_version
-
-
-@lru_cache(maxsize=1)
-def languages():
-    args_tess = [
-        'tesseract',
-        '--list-langs'
-    ]
-    try:
-        langs = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
-    except CalledProcessError as e:
-        print("Tesseract failed to report available languages.")
-        print("Output from Tesseract:")
-        print("-" * 40)
-        print(e.output)
-        sys.exit(ExitCode.missing_dependency)
-    return set(lang.strip() for lang in langs.splitlines()[1:])
-
-
-HOCR_TEMPLATE = '''<?xml version="1.0" encoding="UTF-8"?>
-<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
-    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
-<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
- <head>
-  <title></title>
-  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
-  <meta name='ocr-system' content='tesseract 3.02.02' />
-  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
- </head>
- <body>
-  <div class='ocr_page' id='page_1' title='image "x.tif"; bbox 0 0 {0} {1}; ppageno 0'>
-   <div class='ocr_carea' id='block_1_1' title="bbox 0 1 {0} {1}">
-    <p class='ocr_par' dir='ltr' id='par_1' title="bbox 0 1 {0} {1}">
-     <span class='ocr_line' id='line_1' title="bbox 0 1 {0} {1}"><span class='ocrx_word' id='word_1' title="bbox 0 1 {0} {1}"> </span>
-     </span>
-    </p>
-   </div>
-  </div>
- </body>
-</html>'''

ocrmypdf/test/test_pageinfo.py

@@ -1,109 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from ocrmypdf import pageinfo
-from reportlab.pdfgen.canvas import Canvas
-from PIL import Image
-from tempfile import NamedTemporaryFile
-from contextlib import suppress
-import os
-import shutil
-import pytest
-import img2pdf
-from pkg_resources import Requirement, resource_filename
-
-req = Requirement.parse('ocrmypdf')
-
-TEST_OUTPUT = os.environ.get(
-    'OCRMYPDF_TEST_OUTPUT',
-    default=os.path.join(os.path.dirname(__file__), 'output'))
-
-
-def setup_module():
-    with suppress(FileNotFoundError):
-        shutil.rmtree(TEST_OUTPUT)
-    with suppress(FileExistsError):
-        os.mkdir(TEST_OUTPUT)
-
-
-def test_single_page_text():
-    filename = os.path.join(TEST_OUTPUT, 'text.pdf')
-    pdf = Canvas(filename, pagesize=(8*72, 6*72))
-    text = pdf.beginText()
-    text.setFont('Helvetica', 12)
-    text.setTextOrigin(1*72, 3*72)
-    text.textLine("Methink'st thou art a general offence and every"
-                  " man should beat thee.")
-    pdf.drawText(text)
-    pdf.showPage()
-    pdf.save()
-
-    pdfinfo = pageinfo.pdf_get_all_pageinfo(filename)
-
-    assert len(pdfinfo) == 1
-    page = pdfinfo[0]
-
-    assert page['has_text']
-    assert len(page['images']) == 0
-
-
-def test_single_page_image():
-    filename = os.path.join(TEST_OUTPUT, 'image-mono.pdf')
-
-    with NamedTemporaryFile() as im_tmp:
-        im = Image.new('1', (8, 8), 0)
-        for n in range(8):
-            im.putpixel((n, n), 1)
-        im.save(im_tmp.name, format='PNG')
-
-        pdf_bytes = img2pdf.convert([im_tmp.name], dpi=8)
-        with open(filename, 'wb') as pdf:
-            pdf.write(pdf_bytes)
-
-    pdfinfo = pageinfo.pdf_get_all_pageinfo(filename)
-
-    assert len(pdfinfo) == 1
-    page = pdfinfo[0]
-
-    assert not page['has_text']
-    assert len(page['images']) == 1
-
-    pdfimage = page['images'][0]
-    assert pdfimage['width'] == 8
-    assert pdfimage['color'] == 'gray'
-
-    # While unexpected, this is correct
-    # PDF spec says /FlateDecode image must have /BitsPerComponent 8
-    # So mono images get upgraded to 8-bit
-    assert pdfimage['bpc'] == 8
-
-    # DPI in a 1"x1" is the image width
-    assert pdfimage['dpi_w'] == 8
-    assert pdfimage['dpi_h'] == 8
-
-
-def test_single_page_inline_image():
-    filename = os.path.join(TEST_OUTPUT, 'image-mono-inline.pdf')
-    pdf = Canvas(filename, pagesize=(8*72, 6*72))
-    with NamedTemporaryFile() as im_tmp:
-        im = Image.new('1', (8, 8), 0)
-        for n in range(8):
-            im.putpixel((n, n), 1)
-        im.save(im_tmp.name, format='PNG')
-        # Draw image in a 72x72 pt or 1"x1" area
-        pdf.drawInlineImage(im_tmp.name, 0, 0, width=72, height=72)
-        pdf.showPage()
-        pdf.save()
-
-    with pytest.raises(NotImplementedError):
-        pageinfo.pdf_get_all_pageinfo(filename)
-
-
-def test_jpeg():
-    filename = resource_filename(req, 'tests/resources/c02-22.pdf')
-
-    pdfinfo = pageinfo.pdf_get_all_pageinfo(filename)
-
-    pdfimage = pdfinfo[0]['images'][0]
-    assert pdfimage['enc'] == 'jpeg'
-

pipeline.svg

@@ -1,235 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?>
-<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN"
- "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
-<!-- Generated by graphviz version 2.38.0 (20140413.2041)
- -->
-<!-- Title: Pipeline: Pages: 1 -->
-<svg width="728pt" height="651pt"
- viewBox="0.00 0.00 728.00 650.53" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
-<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 646.53)">
-<title>Pipeline:</title>
-<polygon fill="white" stroke="none" points="-4,4 -4,-646.53 724,-646.53 724,4 -4,4"/>
-<g id="clust1" class="cluster"><title>clustertasks</title>
-<polygon fill="none" stroke="black" points="8,-8 8,-634.53 712,-634.53 712,-8 8,-8"/>
-<text text-anchor="middle" x="360" y="-606.53" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
-</g>
-<!-- t0 -->
-<g id="node1" class="node"><title>t0</title>
-<polygon fill="#efa03b" stroke="#006000" points="481.791,-588.53 386.209,-588.53 382.209,-584.53 382.209,-552.53 477.791,-552.53 481.791,-556.53 481.791,-588.53"/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 382.209,-584.53 "/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 477.791,-552.53 "/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 481.791,-588.53 "/>
-<text text-anchor="middle" x="432" y="-564.53" font-family="Times,serif" font-size="20.00" fill="#006000">repair_pdf</text>
-</g>
-<!-- t1 -->
-<g id="node2" class="node"><title>t1</title>
-<polygon fill="#efa03b" stroke="black" points="466.782,-509.564 374,-526.497 281.218,-509.564 281.304,-482.165 466.696,-482.165 466.782,-509.564"/>
-<polygon fill="none" stroke="black" points="470.799,-512.902 374,-530.569 277.201,-512.902 277.311,-478.159 470.689,-478.159 470.799,-512.902"/>
-<text text-anchor="middle" x="374" y="-495.991" font-family="Times,serif" font-size="20.00">split_pages</text>
-</g>
-<!-- t0&#45;&gt;t1 -->
-<g id="edge1" class="edge"><title>t0&#45;&gt;t1</title>
-<path fill="none" stroke="gray" d="M417.064,-552.394C412.296,-546.925 406.86,-540.689 401.493,-534.532"/>
-<polygon fill="gray" stroke="gray" points="403.996,-532.077 394.787,-526.838 398.719,-536.676 403.996,-532.077"/>
-</g>
-<!-- t10 -->
-<g id="node12" class="node"><title>t10</title>
-<polygon fill="#efa03b" stroke="#006000" points="704.338,-451.452 493.662,-451.452 489.662,-447.452 489.662,-415.452 700.338,-415.452 704.338,-419.452 704.338,-451.452"/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 489.662,-447.452 "/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 700.338,-415.452 "/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 704.338,-451.452 "/>
-<text text-anchor="middle" x="597" y="-427.452" font-family="Times,serif" font-size="20.00" fill="#006000">generate_postscript_stub</text>
-</g>
-<!-- t0&#45;&gt;t10 -->
-<g id="edge16" class="edge"><title>t0&#45;&gt;t10</title>
-<path fill="none" stroke="gray" d="M453.048,-552.468C461.441,-545.654 471.185,-537.73 480,-530.53 510.227,-505.84 544.753,-477.467 568.421,-457.99"/>
-<polygon fill="gray" stroke="gray" points="570.818,-460.55 576.315,-451.493 566.369,-455.146 570.818,-460.55"/>
-</g>
-<!-- t2 -->
-<g id="node3" class="node"><title>t2</title>
-<polygon fill="#efa03b" stroke="black" points="451.555,-451.452 228.445,-451.452 224.445,-447.452 224.445,-415.452 447.555,-415.452 451.555,-419.452 451.555,-451.452"/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 224.445,-447.452 "/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 447.555,-415.452 "/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 451.555,-451.452 "/>
-<text text-anchor="middle" x="338" y="-427.452" font-family="Times,serif" font-size="20.00">rasterize_with_ghostscript</text>
-</g>
-<!-- t1&#45;&gt;t2 -->
-<g id="edge2" class="edge"><title>t1&#45;&gt;t2</title>
-<path fill="none" stroke="#0044a0" d="M361.611,-478.092C358.514,-472.369 355.171,-466.19 352.003,-460.333"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="355.064,-458.636 347.227,-451.506 348.907,-461.967 355.064,-458.636"/>
-</g>
-<!-- t11 -->
-<g id="node10" class="node"><title>t11</title>
-<polygon fill="#efa03b" stroke="black" points="644.594,-393.452 551.406,-393.452 547.406,-389.452 547.406,-357.452 640.594,-357.452 644.594,-361.452 644.594,-393.452"/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 547.406,-389.452 "/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 640.594,-357.452 "/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 644.594,-393.452 "/>
-<text text-anchor="middle" x="596" y="-369.452" font-family="Times,serif" font-size="20.00">skip_page</text>
-</g>
-<!-- t1&#45;&gt;t11 -->
-<g id="edge13" class="edge"><title>t1&#45;&gt;t11</title>
-<path fill="none" stroke="#0044a0" d="M425.139,-478.007C437.834,-470.731 450.731,-461.83 461,-451.452 473.874,-438.442 466.946,-427.178 481,-415.452 490.212,-407.766 513.965,-399.205 537.508,-392.067"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="538.592,-395.396 547.189,-389.203 536.607,-388.683 538.592,-395.396"/>
-</g>
-<!-- t9 -->
-<g id="node11" class="node"><title>t9</title>
-<polygon fill="#efa03b" stroke="black" points="272.496,-277.452 19.5039,-277.452 15.5039,-273.452 15.5039,-241.452 268.496,-241.452 272.496,-245.452 272.496,-277.452"/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 15.5039,-273.452 "/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 268.496,-241.452 "/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 272.496,-277.452 "/>
-<text text-anchor="middle" x="144" y="-253.452" font-family="Times,serif" font-size="20.00">tesseract_ocr_and_render_pdf</text>
-</g>
-<!-- t1&#45;&gt;t9 -->
-<g id="edge15" class="edge"><title>t1&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M277.341,-486.044C254.708,-478.702 232.21,-467.759 215,-451.452 168.207,-407.114 152.056,-329.127 146.633,-287.855"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="150.084,-287.235 145.422,-277.721 143.133,-288.065 150.084,-287.235"/>
-</g>
-<!-- t3 -->
-<g id="node4" class="node"><title>t3</title>
-<polygon fill="#efa03b" stroke="black" points="458.999,-393.452 291.001,-393.452 287.001,-389.452 287.001,-357.452 454.999,-357.452 458.999,-361.452 458.999,-393.452"/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 287.001,-389.452 "/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 454.999,-357.452 "/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 458.999,-393.452 "/>
-<text text-anchor="middle" x="373" y="-369.452" font-family="Times,serif" font-size="20.00">preprocess_deskew</text>
-</g>
-<!-- t2&#45;&gt;t3 -->
-<g id="edge3" class="edge"><title>t2&#45;&gt;t3</title>
-<path fill="none" stroke="#0044a0" d="M348.691,-415.346C351.337,-411.112 354.23,-406.485 357.066,-401.946"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="360.042,-403.788 362.374,-393.453 354.106,-400.078 360.042,-403.788"/>
-</g>
-<!-- t6 -->
-<g id="node6" class="node"><title>t6</title>
-<polygon fill="#efa03b" stroke="black" points="664.375,-277.452 477.625,-277.452 473.625,-273.452 473.625,-241.452 660.375,-241.452 664.375,-245.452 664.375,-277.452"/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 473.625,-273.452 "/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 660.375,-241.452 "/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 664.375,-277.452 "/>
-<text text-anchor="middle" x="569" y="-253.452" font-family="Times,serif" font-size="20.00">select_image_for_pdf</text>
-</g>
-<!-- t2&#45;&gt;t6 -->
-<g id="edge7" class="edge"><title>t2&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M423.032,-415.374C438.812,-409.925 454.543,-402.785 468,-393.452 508.201,-365.571 539.287,-316.662 555.808,-286.558"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="559.024,-287.965 560.653,-277.496 552.851,-284.664 559.024,-287.965"/>
-</g>
-<!-- t4 -->
-<g id="node5" class="node"><title>t4</title>
-<polygon fill="#efa03b" stroke="black" points="449.705,-335.452 300.295,-335.452 296.295,-331.452 296.295,-299.452 445.705,-299.452 449.705,-303.452 449.705,-335.452"/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 296.295,-331.452 "/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 445.705,-299.452 "/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 449.705,-335.452 "/>
-<text text-anchor="middle" x="373" y="-311.452" font-family="Times,serif" font-size="20.00">preprocess_clean</text>
-</g>
-<!-- t3&#45;&gt;t4 -->
-<g id="edge4" class="edge"><title>t3&#45;&gt;t4</title>
-<path fill="none" stroke="#0044a0" d="M373,-357.346C373,-353.655 373,-349.665 373,-345.695"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="376.5,-345.453 373,-335.453 369.5,-345.453 376.5,-345.453"/>
-</g>
-<!-- t3&#45;&gt;t6 -->
-<g id="edge6" class="edge"><title>t3&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M414.399,-357.412C428.785,-351.039 444.852,-343.405 459,-335.452 466.909,-331.007 506.49,-303.805 535.921,-283.431"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="538.073,-286.198 544.299,-277.625 534.086,-280.444 538.073,-286.198"/>
-</g>
-<!-- t4&#45;&gt;t6 -->
-<g id="edge5" class="edge"><title>t4&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M432.605,-299.422C453.701,-293.395 477.617,-286.561 499.465,-280.319"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="500.687,-283.61 509.341,-277.497 498.764,-276.879 500.687,-283.61"/>
-</g>
-<!-- t5 -->
-<g id="node7" class="node"><title>t5</title>
-<polygon fill="#efa03b" stroke="black" points="455.922,-277.452 294.078,-277.452 290.078,-273.452 290.078,-241.452 451.922,-241.452 455.922,-245.452 455.922,-277.452"/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 290.078,-273.452 "/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 451.922,-241.452 "/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 455.922,-277.452 "/>
-<text text-anchor="middle" x="373" y="-253.452" font-family="Times,serif" font-size="20.00">ocr_tesseract_hocr</text>
-</g>
-<!-- t4&#45;&gt;t5 -->
-<g id="edge8" class="edge"><title>t4&#45;&gt;t5</title>
-<path fill="none" stroke="#0044a0" d="M373,-299.346C373,-295.655 373,-291.665 373,-287.695"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="376.5,-287.453 373,-277.453 369.5,-287.453 376.5,-287.453"/>
-</g>
-<!-- t4&#45;&gt;t9 -->
-<g id="edge14" class="edge"><title>t4&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M303.36,-299.422C278.157,-293.259 249.509,-286.253 223.521,-279.898"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="224.249,-276.473 213.704,-277.497 222.586,-283.273 224.249,-276.473"/>
-</g>
-<!-- t7 -->
-<g id="node8" class="node"><title>t7</title>
-<polygon fill="#efa03b" stroke="black" points="426.366,-219.452 269.634,-219.452 265.634,-215.452 265.634,-183.452 422.366,-183.452 426.366,-187.452 426.366,-219.452"/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 265.634,-215.452 "/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 422.366,-183.452 "/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 426.366,-219.452 "/>
-<text text-anchor="middle" x="346" y="-195.452" font-family="Times,serif" font-size="20.00">render_hocr_page</text>
-</g>
-<!-- t6&#45;&gt;t7 -->
-<g id="edge9" class="edge"><title>t6&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M501.184,-241.422C476.75,-235.286 448.99,-228.315 423.772,-221.982"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="424.429,-218.538 413.877,-219.497 422.724,-225.328 424.429,-218.538"/>
-</g>
-<!-- t8 -->
-<g id="node9" class="node"><title>t8</title>
-<polygon fill="#efa03b" stroke="black" points="663.742,-219.452 448.258,-219.452 444.258,-215.452 444.258,-183.452 659.742,-183.452 663.742,-187.452 663.742,-219.452"/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 444.258,-215.452 "/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 659.742,-183.452 "/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 663.742,-219.452 "/>
-<text text-anchor="middle" x="554" y="-195.452" font-family="Times,serif" font-size="20.00">render_hocr_debug_page</text>
-</g>
-<!-- t6&#45;&gt;t8 -->
-<g id="edge11" class="edge"><title>t6&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M564.418,-241.346C563.4,-237.546 562.298,-233.43 561.203,-229.345"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="564.522,-228.207 558.554,-219.453 557.761,-230.018 564.522,-228.207"/>
-</g>
-<!-- t5&#45;&gt;t7 -->
-<g id="edge10" class="edge"><title>t5&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M364.752,-241.346C362.816,-237.329 360.708,-232.958 358.629,-228.645"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="361.693,-226.941 354.197,-219.453 355.387,-229.981 361.693,-226.941"/>
-</g>
-<!-- t5&#45;&gt;t8 -->
-<g id="edge12" class="edge"><title>t5&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M428.289,-241.346C447.587,-235.375 469.416,-228.622 489.404,-222.437"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="490.53,-225.753 499.049,-219.453 488.461,-219.065 490.53,-225.753"/>
-</g>
-<!-- t12 -->
-<g id="node13" class="node"><title>t12</title>
-<polygon fill="#efa03b" stroke="black" points="473.845,-105.456 554,-78.0208 634.155,-105.456 634.08,-149.848 473.92,-149.848 473.845,-105.456"/>
-<polygon fill="none" stroke="black" points="469.836,-102.581 554,-73.7729 638.164,-102.581 638.078,-153.869 469.922,-153.869 469.836,-102.581"/>
-<text text-anchor="middle" x="554" y="-111.726" font-family="Times,serif" font-size="20.00">merge_pages</text>
-</g>
-<!-- t7&#45;&gt;t12 -->
-<g id="edge20" class="edge"><title>t7&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M389.35,-183.419C409.937,-175.33 435.398,-165.326 460.092,-155.624"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="461.479,-158.839 469.507,-151.924 458.919,-152.324 461.479,-158.839"/>
-</g>
-<!-- t8&#45;&gt;t12 -->
-<g id="edge19" class="edge"><title>t8&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M554,-183.12C554,-177.585 554,-171.177 554,-164.592"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="557.5,-164.201 554,-154.201 550.5,-164.201 557.5,-164.201"/>
-</g>
-<!-- t11&#45;&gt;t12 -->
-<g id="edge17" class="edge"><title>t11&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M615.939,-357.13C634.767,-339.333 661.745,-309.733 673,-277.452 686.754,-238.003 694.347,-219.364 673,-183.452 666.345,-172.256 656.999,-162.875 646.44,-155.047"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="648.378,-152.132 638.152,-149.36 644.417,-157.904 648.378,-152.132"/>
-</g>
-<!-- t9&#45;&gt;t12 -->
-<g id="edge18" class="edge"><title>t9&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M164.711,-241.235C186.394,-224.08 222.088,-198.224 257,-183.452 321.895,-155.994 399.893,-139.571 459.67,-130.141"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="460.504,-133.554 469.855,-128.574 459.439,-126.635 460.504,-133.554"/>
-</g>
-<!-- t10&#45;&gt;t12 -->
-<g id="edge22" class="edge"><title>t10&#45;&gt;t12</title>
-<path fill="none" stroke="gray" d="M629.379,-415.337C660.552,-396.255 703,-362.233 703,-318.452 703,-318.452 703,-318.452 703,-258.452 703,-224.066 705.581,-210.223 684,-183.452 673.976,-171.017 660.952,-160.773 647.026,-152.396"/>
-<polygon fill="gray" stroke="gray" points="648.682,-149.312 638.258,-147.419 645.227,-155.4 648.682,-149.312"/>
-</g>
-<!-- t13 -->
-<g id="node14" class="node"><title>t13</title>
-<polygon fill="#efa03b" stroke="black" points="616.338,-52 495.662,-52 491.662,-48 491.662,-16 612.338,-16 616.338,-20 616.338,-52"/>
-<polyline fill="none" stroke="black" points="612.338,-48 491.662,-48 "/>
-<polyline fill="none" stroke="black" points="612.338,-48 612.338,-16 "/>
-<polyline fill="none" stroke="black" points="612.338,-48 616.338,-52 "/>
-<text text-anchor="middle" x="554" y="-28" font-family="Times,serif" font-size="20.00">validate_pdfa</text>
-</g>
-<!-- t12&#45;&gt;t13 -->
-<g id="edge21" class="edge"><title>t12&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M554,-73.9482C554,-69.9654 554,-66.007 554,-62.2247"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="557.5,-62.1573 554,-52.1573 550.5,-62.1574 557.5,-62.1573"/>
-</g>
-</g>
-</svg>

requirements.txt

@@ -1,4 +1,11 @@
-ruffus>=2.6.3
-Pillow>=2.4.0
-reportlab>=3.1.44
-PyPDF2>=1.25.1
+# requirements.txt can be used to replicate the developer's build environment
+# setup.py lists a separate set of requirements that are looser to simplify
+# installation
+ruffus == 2.7.0
+Pillow == 5.2.0
+reportlab == 3.4.0
+PyPDF2 == 1.26.0
+img2pdf == 0.2.4
+cffi == 1.11.5
+PyMuPDF == 1.12.5
+defusedxml == 0.5.0

setup.cfg

@@ -1,2 +1,16 @@
 [bdist_wheel]
-python-tag = py34
+python-tag = py35
+
+[aliases]
+test=pytest
+
+[check-manifest]
+ignore =
+	.github
+
+[tool:pytest]
+norecursedirs = lib .pc .git output cache resources
+testpaths = tests
+
+[metadata]
+license_file = LICENSE

setup.py

@@ -1,12 +1,33 @@
 #!/usr/bin/env python3
+# -*- coding: utf-8 -*-
 # © 2015 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from __future__ import print_function, unicode_literals
 
-from setuptools import setup
-from subprocess import Popen, STDOUT, check_output, CalledProcessError
-from string import Template
-from collections.abc import Mapping
-import re
 import sys
+if sys.version_info < (3, 5):
+    print("Python 3.5 or newer is required", file=sys.stderr)
+    sys.exit(1)
+
+from setuptools import setup, find_packages  # nopep8
+from subprocess import STDOUT, check_output, CalledProcessError  # nopep8
+from collections.abc import Mapping  # nopep8
+import re  # nopep8
 
 
 missing_program = '''
@@ -142,59 +163,65 @@ def check_external_program(
 
 
 command = next((arg for arg in sys.argv[1:] if not arg.startswith('-')), '')
+forced = ('--force' in sys.argv)
 
 
-if command.startswith('install') or \
-        command in ['check', 'test', 'nosetests', 'easy_install', 'egg_info']:
+if not forced and command.startswith('install') or \
+        command in ['check', 'test', 'nosetests', 'easy_install']:
     check_external_program(
         program='tesseract',
-        need_version='3.02.02',
+        need_version='3.04',  # using backport for Travis CI
         package={'darwin': 'tesseract', 'linux': 'tesseract-ocr'}
     )
     check_external_program(
         program='gs',
-        need_version='9.14',
+        need_version='9.15',  # limited by Travis CI / Ubuntu 14.04 backports
         package='ghostscript'
     )
     check_external_program(
         program='unpaper',
-        need_version='6.1',
+        need_version='6.1',   # latest sane version
         package='unpaper',
         optional=True
     )
     check_external_program(
         program='qpdf',
-        need_version='5.0.0',
+        need_version='7.0.0', # test suite known to fail on 5.1.1
         package='qpdf',
         version_check_args=['--version']
     )
 
-
 if 'upload' in sys.argv[1:]:
     print('Use twine to upload the package - setup.py upload is insecure')
     sys.exit(1)
 
-install_requires = open('requirements.txt').read().splitlines()
 tests_require = open('test_requirements.txt').read().splitlines()
 
+
+def readme():
+    with open('README.rst') as f:
+        return f.read()
+
 setup(
     name='ocrmypdf',
-    version='3.0',  # also update: release notes, main.py
     description='OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to be searched',
+    long_description=readme(),
     url='https://github.com/jbarlow83/OCRmyPDF',
     author='James R. Barlow',
     author_email='jim@purplerock.ca',
-    license='Public Domain',
-    packages=['ocrmypdf'],
+    packages=find_packages('src', exclude=["tests", "tests.*"]),
+    package_dir={'': 'src'},
     keywords=['PDF', 'OCR', 'optical character recognition', 'PDF/A', 'scanning'],
     classifiers=[
-        "Programming Language :: Python :: 3",
+        "Programming Language :: Python :: 3.5",
+        "Programming Language :: Python :: 3.6",
+        "Programming Language :: Python :: 3.7",
         "Development Status :: 5 - Production/Stable",
         "Environment :: Console",
         "Intended Audience :: End Users/Desktop",
         "Intended Audience :: Science/Research",
         "Intended Audience :: System Administrators",
-        "License :: Public Domain",
+        "License :: OSI Approved :: GNU General Public License v3 (GPLv3)",
         "Operating System :: MacOS :: MacOS X",
         "Operating System :: POSIX",
         "Operating System :: POSIX :: BSD",
@@ -203,12 +230,36 @@ setup(
         "Topic :: Text Processing :: Indexing",
         "Topic :: Text Processing :: Linguistic",
         ],
-    install_requires=install_requires,
+    python_requires=' >= 3.5',
+    setup_requires=[
+        'cffi >= 1.9.1',        # to build the leptonica module
+        'pytest-runner',        # to enable python setup.py test
+        'setuptools_scm',       # so that version will work
+    ],
+    use_scm_version={'version_scheme': 'post-release'},
+    cffi_modules=[
+        'src/ocrmypdf/lib/compile_leptonica.py:ffibuilder'
+    ],
+    install_requires=[
+        'cffi >= 1.9.1',          # must be a setup and install requirement
+        'defusedxml >= 0.5.0',    # pure Python, so track HEAD closely
+        'img2pdf >= 0.2.4',       # pure Python, so track HEAD closely
+        'Pillow >= 4.0.0, != 5.1.0 ; sys_platform == "darwin"',
+                                  # Pillow < 4 has BytesIO/TIFF bug w/img2pdf 0.2.3
+                                  # block 5.1.0, broken wheels
+        'PyPDF2 >= 1.26',         # pure Python, so track HEAD closely
+        'reportlab >= 3.3.0',     # oldest released version with sane image handling
+        'ruffus >= 2.7.0',
+    ],
+    extras_require={
+        'fitz': ['PyMuPDF >= 1.12.5']     # for table of contents bug
+    },
     tests_require=tests_require,
     entry_points={
         'console_scripts': [
-            'ocrmypdf = ocrmypdf.main:run_pipeline'
+            'ocrmypdf = ocrmypdf.__main__:run_pipeline'
         ],
     },
+    package_data={'ocrmypdf': ['data/sRGB.icc']},
     include_package_data=True,
     zip_safe=False)

src/ocrmypdf/__init__.py

@@ -0,0 +1,25 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+import pkg_resources
+
+PROGRAM_NAME = 'ocrmypdf'
+
+# Official PEP 396
+__version__ = pkg_resources.get_distribution('ocrmypdf').version
+
+VERSION = __version__

src/ocrmypdf/__main__.py

@@ -0,0 +1,961 @@
+#!/usr/bin/env python3
+# © 2015-17 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from tempfile import mkdtemp
+from collections.abc import Sequence
+from pathlib import Path
+import sys
+import os
+import re
+import warnings
+import multiprocessing
+import atexit
+import textwrap
+import logging
+import argparse
+
+import PyPDF2 as pypdf
+import PIL
+
+import ruffus.ruffus_exceptions as ruffus_exceptions
+import ruffus.cmdline as cmdline
+import ruffus.proxy_logger as proxy_logger
+
+from .pipeline import JobContext, JobContextManager, \
+    cleanup_working_files, build_pipeline
+from .pdfa import file_claims_pdfa
+from .helpers import is_iterable_notstr, re_symlink, is_file_writable, \
+    available_cpu_count
+from .exec import tesseract, qpdf, ghostscript
+from .lib import fitz
+from . import PROGRAM_NAME, VERSION
+
+from .exceptions import ExitCode, ExitCodeException, MissingDependencyError, \
+    InputFileError, BadArgsError, OutputFileAccessError
+from . import exceptions as ocrmypdf_exceptions
+from ._unicodefun import verify_python3_env
+
+warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
+
+
+# -------------
+# External dependencies
+
+MINIMUM_TESS_VERSION = '3.04'
+
+HOCR_OK_LANGS = frozenset([
+    'eng', 'deu', 'spa', 'ita', 'por'
+])
+
+def complain(message):
+    print(*textwrap.wrap(message), file=sys.stderr)
+
+
+# Hack to help debugger context find /usr/local/bin
+if 'IDE_PROJECT_ROOTS' in os.environ:
+    os.environ['PATH'] = '/usr/local/bin:' + os.environ['PATH']
+
+# --------
+# Critical environment tests
+
+verify_python3_env()
+
+if tesseract.version() < MINIMUM_TESS_VERSION:
+    complain(
+        "Please install tesseract {0} or newer "
+        "(currently installed version is {1})".format(
+            MINIMUM_TESS_VERSION, tesseract.version()))
+    sys.exit(ExitCode.missing_dependency)
+
+# -------------
+# Parser
+
+def numeric(basetype, min_=None, max_=None):
+    "Validator for numeric params"
+    min_ = basetype(min_) if min_ is not None else None
+    max_ = basetype(max_) if max_ is not None else None
+    def _numeric(string):
+        value = basetype(string)
+        if (min_ is not None and value < min_
+                or max_ is not None and value > max_):
+            msg = "%r not in valid range %r" % (string, (min_, max_))
+            raise argparse.ArgumentTypeError(msg)
+        return value
+    return _numeric
+
+
+parser = argparse.ArgumentParser(
+    prog=PROGRAM_NAME,
+    fromfile_prefix_chars='@',
+    formatter_class=argparse.RawDescriptionHelpFormatter,
+    description="""\
+Generates a searchable PDF or PDF/A from a regular PDF.
+
+OCRmyPDF rasterizes each page of the input PDF, optionally corrects page
+rotation and performs image processing, runs the Tesseract OCR engine on the
+image, and then creates a PDF from the OCR information.
+""",
+    epilog="""\
+OCRmyPDF attempts to keep the output file at about the same size.  If a file
+contains losslessly compressed images, and output file will be losslessly
+compressed as well.
+
+PDF is a page description file that attempts to preserve a layout exactly.
+A PDF can contain vector objects (such as text or lines) and raster objects
+(images).  A page might have multiple images.  OCRmyPDF is prepared to deal
+with the wide variety of PDFs that exist in the wild.
+
+When a PDF page contains text, OCRmyPDF assumes that the page has already
+been OCRed or is a "born digital" page that should not be OCRed.  The default
+behavior is to exit in this case without producing a file.  You can use the
+option --skip-text to ignore pages with text, or --force-ocr to rasterize
+all objects on the page and produce an image-only PDF as output.
+
+    ocrmypdf --skip-text file_with_some_text_pages.pdf output.pdf
+
+    ocrmypdf --force-ocr word_document.pdf output.pdf
+
+If you are concerned about long-term archiving of PDFs, use the default option
+--output-type pdfa which converts the PDF to a standardized PDF/A-2b.  This
+converts images to sRGB colorspace, removes some features from the PDF such
+as Javascript or forms. If you want to minimize the number of changes made to
+your PDF, use --output-type pdf.
+
+If OCRmyPDF is given an image file as input, it will attempt to convert the
+image to a PDF before processing.  For more control over the conversion of
+images to PDF, use the Python package img2pdf or other image to PDF software.
+
+For example, this command uses img2pdf to convert all .png files beginning
+with the 'page' prefix to a PDF, fitting each image on A4-sized paper, and
+sending the result to OCRmyPDF through a pipe.  img2pdf is a dependency of
+ocrmypdf so it is already installed.
+
+    img2pdf --pagesize A4 page*.png | ocrmypdf - myfile.pdf
+
+Online documentation is located at:
+    https://ocrmypdf.readthedocs.io/en/latest/introduction.html
+
+""")
+
+parser.add_argument(
+    'input_file', metavar="input_pdf_or_image",
+    help="PDF file containing the images to be OCRed (or '-' to read from "
+         "standard input)")
+parser.add_argument(
+    'output_file', metavar="output_pdf",
+    help="Output searchable PDF file (or '-' to write to standard output). "
+         "Existing files will be ovewritten. If same as input file, the "
+         "input file will be updated only if processing is successful.")
+parser.add_argument(
+    '-l', '--language', action='append',
+    help="Language(s) of the file to be OCRed (see tesseract --list-langs for "
+         "all language packs installed in your system). Use -l eng+deu for "
+         "multiple languages.")
+parser.add_argument(
+    '--image-dpi', metavar='DPI', type=int,
+    help="For input image instead of PDF, use this DPI instead of file's.")
+parser.add_argument(
+    '--output-type', choices=['pdfa', 'pdf', 'pdfa-1', 'pdfa-2', 'pdfa-3'],
+    default='pdfa',
+    help="Choose output type. 'pdfa' creates a PDF/A-2b compliant file for "
+         "long term archiving (default, recommended) but may not suitable "
+         "for users who want their file altered as little as possible. 'pdfa' "
+         "also has problems with full Unicode text. 'pdf' attempts to "
+         "preserve file contents as much as possible. 'pdf-a1' creates a "
+         "PDF/A1-b file. 'pdf-a2' is equivalent to 'pdfa'. 'pdf-a3' creates a "
+         "PDF/A3-b file."
+         )
+
+# Use null string '\0' as sentinel to indicate the user supplied no argument,
+# since that is the only invalid character for filepaths on all platforms
+# bool('\0') is True in Python
+parser.add_argument(
+    '--sidecar', nargs='?', const='\0', default=None, metavar='FILE',
+    help="Generate sidecar text files that contain the same text recognized "
+         "by Tesseract. This may be useful for building a OCR text database. "
+         "If FILE is omitted, the sidecar file be named {output_file}.txt "
+         "If FILE is set to '-', the sidecar is written to stdout (a "
+         "convenient way to preview OCR quality). The output file and sidecar "
+         "may not both use stdout at the same time.")
+
+parser.add_argument(
+    '--version', action='version', version=VERSION,
+    help="Print program version and exit")
+
+jobcontrol = parser.add_argument_group(
+    "Job control options")
+jobcontrol.add_argument(
+    '-j', '--jobs', metavar='N', type=int,
+    help="Use up to N CPU cores simultaneously (default: use all).")
+jobcontrol.add_argument(
+    '-q', '--quiet', action='store_true', help="Suppress INFO messages")
+jobcontrol.add_argument(
+    '-v', '--verbose', const="+", default=[], nargs='?', action="append",
+    help="Print more verbose messages for each additional verbose level")
+
+metadata = parser.add_argument_group(
+    "Metadata options",
+    "Set output PDF/A metadata (default: copy input document's metadata)")
+metadata.add_argument(
+    '--title', type=str,
+    help="Set document title (place multiple words in quotes)")
+metadata.add_argument(
+    '--author', type=str,
+    help="Set document author")
+metadata.add_argument(
+    '--subject', type=str,
+    help="Set document subject description")
+metadata.add_argument(
+    '--keywords', type=str,
+    help="Set document keywords")
+
+preprocessing = parser.add_argument_group(
+    "Image preprocessing options",
+    "Options to improve the quality of the final PDF and OCR")
+preprocessing.add_argument(
+    '-r', '--rotate-pages', action='store_true',
+    help="Automatically rotate pages based on detected text orientation")
+preprocessing.add_argument(
+    '--remove-background', action='store_true',
+    help="Attempt to remove background from gray or color pages, setting it "
+         "to white ")
+preprocessing.add_argument(
+    '-d', '--deskew', action='store_true',
+    help="Deskew each page before performing OCR")
+preprocessing.add_argument(
+    '-c', '--clean', action='store_true',
+    help="Clean pages from scanning artifacts before performing OCR, and send "
+         "the cleaned page to OCR, but do not include the cleaned page in "
+         "the output")
+preprocessing.add_argument(
+    '-i', '--clean-final', action='store_true',
+    help="Clean page as above, and incorporate the cleaned image in the final "
+         "PDF.  Might remove desired content.")
+preprocessing.add_argument(
+    '--oversample', metavar='DPI', type=numeric(int, 0, 5000), default=0,
+    help="Oversample images to at least the specified DPI, to improve OCR "
+         "results slightly")
+
+ocrsettings = parser.add_argument_group(
+    "OCR options",
+    "Control how OCR is applied")
+ocrsettings.add_argument(
+    '-f', '--force-ocr', action='store_true',
+    help="Rasterize any fonts or vector objects on each page, apply OCR, and "
+         "save the rastered output (this rewrites the PDF)")
+ocrsettings.add_argument(
+    '-s', '--skip-text', action='store_true',
+    help="Skip OCR on any pages that already contain text, but include the "
+         "page in final output; useful for PDFs that contain a mix of "
+         "images, text pages, and/or previously OCRed pages")
+# ocrsettings.add_argument(
+#     '--redo-ocr', action='store_true',
+#     help="removing any existing OCR text, but otherwise preserve mixed PDF "
+#          "pages")
+
+ocrsettings.add_argument(
+    '--skip-big', type=numeric(float, 0, 5000), metavar='MPixels',
+    help="Skip OCR on pages larger than the specified amount of megapixels, "
+         "but include skipped pages in final output")
+
+advanced = parser.add_argument_group(
+    "Advanced",
+    "Advanced options to control Tesseract's OCR behavior")
+advanced.add_argument(
+    '--max-image-mpixels', action='store', type=numeric(float, 0), metavar='MPixels',
+    help="Set maximum number of pixels to unpack before treating an image as a "
+         "decompression bomb",
+    default=128.0)
+advanced.add_argument(
+    '--tesseract-config', action='append', metavar='CFG', default=[],
+    help="Additional Tesseract configuration files -- see documentation")
+advanced.add_argument(
+    '--tesseract-pagesegmode', action='store', type=int, metavar='PSM',
+    choices=range(0, 14),
+    help="Set Tesseract page segmentation mode (see tesseract --help)")
+advanced.add_argument(
+    '--tesseract-oem', action='store', type=int, metavar='MODE',
+    choices=range(0, 4),
+    help=("Set Tesseract 4.0 OCR engine mode: "
+         "0 - original Tesseract only; "
+         "1 - neural nets LSTM only; "
+         "2 - Tesseract + LSTM; "
+         "3 - default.")
+    )
+advanced.add_argument(
+    '--pdf-renderer',
+    choices=['auto', 'tesseract', 'hocr', 'sandwich'], default='auto',
+    help="Choose OCR PDF renderer - the default option is to let OCRmyPDF "
+         "choose."
+         "auto - let OCRmyPDF choose; "
+         "sandwich - default renderer for Tesseract 3.05.01 and newer; "
+         "hocr - default renderer for older versions of Tesseract; "
+         "tesseract - gives better results for non-Latin languages and "
+         "Tesseract older than 3.05.01 but has problems with some versions "
+         " of Ghostscript; deprecated"
+    )
+advanced.add_argument(
+    '--tesseract-timeout', default=180.0, type=numeric(float, 0), metavar='SECONDS',
+    help='Give up on OCR after the timeout, but copy the preprocessed page '
+         'into the final output')
+advanced.add_argument(
+    '--rotate-pages-threshold', default=14.0, type=numeric(float, max_=1000), metavar='CONFIDENCE',
+    help="Only rotate pages when confidence is above this value (arbitrary "
+         "units reported by tesseract)")
+advanced.add_argument(
+    '--pdfa-image-compression', choices=['auto', 'jpeg', 'lossless'],
+    default='auto',
+    help="Specify how to compress images in the output PDF/A. 'auto' lets "
+         "OCRmyPDF decide.  'jpeg' changes all grayscale and color images to "
+         "JPEG compression.  'lossless' uses PNG-style lossless compression "
+         "for all images.  Monochrome images are always compressed using a "
+         "lossless codec.  Compression settings "
+         "are applied to all pages, including those for which OCR was "
+         "skipped.  Not supported for --output-type=pdf ; that setting "
+         "preserves the original compression of all images.")
+advanced.add_argument(
+    '--user-words', metavar='FILE',
+    help="Specify the location of the Tesseract user words file. This is a "
+         "list of words Tesseract should consider while performing OCR in "
+         "addition to its standard language dictionaries. This can improve "
+         "OCR quality especially for specialized and technical documents.")
+advanced.add_argument(
+    '--user-patterns', metavar='FILE',
+    help="Specify the location of the Tesseract user patterns file.")
+advanced.add_argument(
+    '--skip-repair', action='store_true',
+    help="Normally OCRmyPDF automatically repairs PDFs using qpdf before "
+         "processing.  If you have already run qpdf or a similar program "
+         "that repairs PDF errors, you can tell OCRmyPDF to skip repair with "
+         "this option.  This may be helpful in batch processing where all "
+         "files are repaired prior to OCR occurs, since repair is single "
+         "threaded and time consuming for large files.")
+
+debugging = parser.add_argument_group(
+    "Debugging",
+    "Arguments to help with troubleshooting and debugging")
+debugging.add_argument(
+    '-k', '--keep-temporary-files', action='store_true',
+    help="Keep temporary files (helpful for debugging)")
+debugging.add_argument(
+    '-g', '--debug-rendering', action='store_true',
+    help="Render each page twice with debug information on second page")
+debugging.add_argument(
+    '--flowchart', type=str,
+    help="Generate the pipeline execution flowchart")
+
+
+def check_options_languages(options, _log):
+    if not options.language:
+        options.language = ['eng']  # Enforce English hegemony
+
+    # Support v2.x "eng+deu" language syntax
+    if '+' in options.language[0]:
+        options.language = options.language[0].split('+')
+
+    languages = set(options.language)
+    if not languages.issubset(tesseract.languages()):
+        msg = (
+            "The installed version of tesseract does not have language "
+            "data for the following requested languages: \n")
+        for lang in (languages - tesseract.languages()):
+            msg += lang + '\n'
+        raise MissingDependencyError(msg)
+
+
+def check_options_output(options, log):
+    if options.pdf_renderer == 'auto':
+        if tesseract.has_textonly_pdf():
+            options.pdf_renderer = 'sandwich'
+        else:
+            options.pdf_renderer = 'hocr'
+
+    if options.pdf_renderer == 'sandwich' and not tesseract.has_textonly_pdf():
+        raise MissingDependencyError(
+            "The 'sandwich' renderer requires Tesseract 3.05.01 or newer; "
+            "or Tesseract 4.00 alpha newer than February 2017.")
+
+    if options.pdf_renderer == 'tesseract':
+        if tesseract.version() < '3.05' and \
+                options.output_type.startswith('pdfa'):
+            log.warning(
+                "For best results use --pdf-renderer=tesseract "
+                "--output-type=pdf to disable PDF/A generation via "
+                "Ghostscript, which is known to corrupt the OCR text of "
+                "some PDFs produced your version of Tesseract.")
+        elif tesseract.has_textonly_pdf():
+            log.warning(
+                "The argument --pdf-renderer=tesseract provides support for "
+                "versions of tesseract older than your version. For best "
+                "results omit this argument and let OCRmyPDF choose the "
+                "best available renderer.")
+
+    if options.debug_rendering and options.pdf_renderer != 'hocr':
+        log.info(
+            "Ignoring --debug-rendering because it requires --pdf-renderer=hocr")
+
+    if options.output_type == 'pdfa':
+        options.output_type = 'pdfa-2'
+
+    if options.output_type == 'pdfa-3' and ghostscript.version() < '9.19':
+        raise MissingDependencyError(
+            "--output-type pdfa-3 requires Ghostscript 9.19 or later"
+        )
+
+    lossless_reconstruction = False
+    if options.pdf_renderer in ('hocr', 'sandwich'):
+        if not any((options.deskew, options.clean_final, options.force_ocr,
+                   options.remove_background)):
+            lossless_reconstruction = True
+    options.lossless_reconstruction = lossless_reconstruction
+
+
+def check_options_sidecar(options, log):
+    if options.sidecar == '\0':
+        if options.output_file == '-':
+            raise argparse.ArgumentError(
+                None,
+                "--sidecar filename must be specified when output file is "
+                "stdout.")
+        options.sidecar = options.output_file + '.txt'
+
+
+def check_options_preprocessing(options, log):
+    if any((options.clean, options.clean_final)):
+        from .exec import unpaper
+        try:
+            if unpaper.version() < '6.1':
+                raise MissingDependencyError(
+                    "The installed 'unpaper' is not supported. "
+                    "Install version 6.1 or newer.")
+        except FileNotFoundError:
+            raise MissingDependencyError(
+                "Install the 'unpaper' program to use --clean, --clean-final.")
+
+    if options.clean and \
+            not options.clean_final and \
+            options.pdf_renderer == 'tesseract':
+        log.info(
+            "Tesseract PDF renderer cannot render --clean pages without "
+            "also performing --clean-final, so --clean-final is assumed.")
+
+
+def check_options_ocr_behavior(options, log):
+    if options.force_ocr and options.skip_text:
+        raise argparse.ArgumentError(
+            None,
+            "Error: --force-ocr and --skip-text are mutually incompatible.")
+
+    # if options.redo_ocr and (options.skip_text or options.force_ocr):
+    #     raise argparse.ArgumentError(
+    #         "Error: --redo-ocr and other OCR options are incompatible.")
+    languages = set(options.language)
+    if options.pdf_renderer == 'hocr' and \
+            not languages.issubset(HOCR_OK_LANGS):
+        msg = (
+            "The 'hocr' PDF renderer is known to cause problems with one "
+            "or more of the languages in your document. ")
+
+        if tesseract.has_textonly_pdf():
+            msg += (
+                "Use --pdf-renderer auto (the default) to avoid this issue.")
+        else:
+            msg += (
+                "Use --pdf-renderer tesseract --output-type pdf to avoid "
+                "this issue")
+        log.warning(msg)
+    elif ghostscript.version() < '9.20' and \
+            not languages.issubset(HOCR_OK_LANGS) \
+            and options.output_type != 'pdf':
+        msg = (
+            "The installed version of Ghostscript does not work correctly "
+            "with the OCR languages you specified. Use --output-type pdf or "
+            "upgrade to Ghostscript 9.20 or later to avoid this issue.")
+        msg += "Found Ghostscript {}".format(ghostscript.version())
+        log.warning(msg)
+
+
+def check_options_advanced(options, log):
+    if options.tesseract_oem and not tesseract.v4():
+        log.warning(
+            "--tesseract-oem requires Tesseract 4.x -- argument ignored")
+    if options.pdf_renderer == 'sandwich' and not tesseract.has_textonly_pdf():
+        raise MissingDependencyError(
+            "--pdf-renderer sandwich requires Tesseract 4.x "
+            "commit 3d9fb3b or later")
+    if options.pdfa_image_compression != 'auto' and \
+            options.output_type.startswith('pdfa'):
+        log.warning(
+            "--pdfa-image-compression argument has no effect when "
+            "--output-type is not 'pdfa', 'pdfa-1', or 'pdfa-2'"
+        )
+
+    if tesseract.v4() and (options.user_words or options.user_patterns):
+        log.warning(
+            'Tesseract 4.x ignores --user-words, so this has no effect')
+
+
+def check_options_metadata(options, log):
+    import unicodedata
+    docinfo = [options.title, options.author, options.keywords,
+               options.subject]
+    for s in (m for m in docinfo if m):
+        for c in s:
+            if unicodedata.category(c) == 'Co' or ord(c) >= 0x10000:
+                raise ValueError(
+                    "One of the metadata strings contains "
+                    "an unsupported Unicode character: '{}' (U+{})".format(
+                        c, hex(ord(c))[2:].upper()
+                ))
+
+
+def check_options_pillow(options, log):
+    PIL.Image.MAX_IMAGE_PIXELS = int(options.max_image_mpixels * 1000000)
+    if PIL.Image.MAX_IMAGE_PIXELS == 0:
+        PIL.Image.MAX_IMAGE_PIXELS = None
+
+
+def check_options(options, log):
+    try:
+        check_options_languages(options, log)
+        check_options_metadata(options, log)
+        check_options_output(options, log)
+        check_options_sidecar(options, log)
+        check_options_preprocessing(options, log)
+        check_options_ocr_behavior(options, log)
+        check_options_advanced(options, log)
+        check_options_pillow(options, log)
+    except ValueError as e:
+        log.error(e)
+        sys.exit(ExitCode.bad_args)
+    except argparse.ArgumentError as e:
+        log.error(e)
+        sys.exit(ExitCode.bad_args)
+    except MissingDependencyError as e:
+        log.error(e)
+        sys.exit(ExitCode.missing_dependency)
+
+
+# ----------
+# Logging
+
+
+def logging_factory(logger_name, logger_args):
+    verbose = logger_args['verbose']
+    quiet = logger_args['quiet']
+
+    root_logger = logging.getLogger(logger_name)
+    root_logger.setLevel(logging.DEBUG)
+
+    handler = logging.StreamHandler(sys.stderr)
+    formatter_ = logging.Formatter("%(levelname)7s - %(message)s")
+    handler.setFormatter(formatter_)
+    if verbose:
+        handler.setLevel(logging.DEBUG)
+    elif quiet:
+        handler.setLevel(logging.WARNING)
+    else:
+        handler.setLevel(logging.INFO)
+    root_logger.addHandler(handler)
+    return root_logger
+
+
+def cleanup_ruffus_error_message(msg):
+    msg = re.sub(r'\s+', r' ', msg)
+    msg = re.sub(r"\((.+?)\)", r'\1', msg)
+    msg = msg.strip()
+    return msg
+
+
+def do_ruffus_exception(ruffus_five_tuple, options, log):
+    """Replace the elaborate ruffus stack trace with a user friendly
+    description of the error message that occurred."""
+    exit_code = None
+
+    task_name, job_name, exc_name, exc_value, exc_stack = ruffus_five_tuple
+    job_name = job_name  # unused
+    if exc_name == 'builtins.SystemExit':
+        match = re.search(r"\.(.+?)\)", exc_value)
+        exit_code_name = match.groups()[0]
+        exit_code = getattr(ExitCode, exit_code_name, 'other_error')
+    elif exc_name == 'ruffus.ruffus_exceptions.MissingInputFileError':
+        log.error(cleanup_ruffus_error_message(exc_value))
+        exit_code = ExitCode.input_file
+    elif exc_name == 'builtins.TypeError':
+        # Even though repair_pdf will fail, ruffus will still try
+        # to call split_pages with no input files, likely due to a bug
+        if task_name == 'split_pages':
+            log.error("Input file '{0}' is not a valid PDF".format(
+                options.input_file))
+            exit_code = ExitCode.input_file
+    elif exc_name == 'builtins.KeyboardInterrupt':
+        log.error("Interrupted by user")
+        exit_code = ExitCode.ctrl_c
+    elif exc_name == 'subprocess.CalledProcessError':
+        # It's up to the subprocess handler to report something useful
+        msg = "Error occurred while running this command:"
+        log.error(msg + '\n' + exc_value)
+        exit_code = ExitCode.child_process_error
+    elif (exc_name == 'PyPDF2.utils.PdfReadError' and \
+            'not been decrypted' in exc_value) or \
+            (exc_name == 'ocrmypdf.exceptions.EncryptedPdfError'):
+        log.error(textwrap.dedent("""\
+            Input PDF is encrypted. The encryption must be removed to
+            perform OCR.
+
+            For information about this PDF's security use
+                qpdf --show-encryption infilename
+
+            You can remove the encryption using
+                qpdf --decrypt [--password=[password]] infilename
+
+            """))
+        exit_code = ExitCode.encrypted_pdf
+    elif exc_name == 'ocrmypdf.exceptions.PdfMergeFailedError':
+        log.error(textwrap.dedent("""\
+            Failed to merge PDF image layer with OCR layer
+
+            Usually this happens because the input PDF file is mal-formed and
+            ocrmypdf cannot automatically correct the problem on its own.
+
+            Try using
+                ocrmypdf --pdf-renderer tesseract  [..other args..]
+            """))
+        exit_code = ExitCode.input_file
+    elif exc_name.startswith('ocrmypdf.exceptions.'):
+        base_exc_name = exc_name.replace('ocrmypdf.exceptions.', '')
+        exc_class = getattr(ocrmypdf_exceptions, base_exc_name)
+        exit_code = exc_class.exit_code
+    elif exc_name == 'PIL.Image.DecompressionBombError':
+        msg = cleanup_ruffus_error_message(exc_value)
+        msg += ("\nUse the --max-image-mpixels argument to set increase the "
+                "maximum number of megapixels to accept.")
+        log.error(msg)
+        exit_code = ExitCode.input_file
+
+    if exit_code is not None:
+        return exit_code
+
+    if not options.verbose:
+        log.error(exc_stack)
+    return ExitCode.other_error
+
+
+def traverse_ruffus_exception(exceptions, options, log):
+    """Traverse a RethrownJobError and output the exceptions
+
+    Ruffus presents exceptions as 5 element tuples. The RethrownJobException
+    has a list of exceptions like
+        e.job_exceptions = [(5-tuple), (5-tuple), ...]
+
+    ruffus < 2.7.0 had a bug with exception marshalling that would give
+    different output whether the main or child process raised the exception.
+    We no longer support this.
+
+    Attempting to log the exception itself will re-marshall it to the logger
+    which is normally running in another process. It's better to avoid re-
+    marshalling.
+
+    The exit code will be based on this, even if multiple exceptions occurred
+    at the same time."""
+
+    exit_codes = []
+    for exc in exceptions:
+        exit_code = do_ruffus_exception(exc, options, log)
+        exit_codes.append(exit_code)
+
+    return exit_codes[0]  # Multiple codes are rare so take the first one
+
+
+
+def check_closed_streams(options):
+    """Work around Python issue with multiprocessing forking on closed streams
+
+    https://bugs.python.org/issue28326
+
+    Attempting to a fork/exec a new Python process when any of std{in,out,err}
+    are closed or not flushable for some reason may raise an exception.
+    Fix this by opening devnull if the handle seems to be closed.  Do this
+    globally to avoid tracking places all places that fork.
+
+    Seems to be specific to multiprocessing.Process not all Python process
+    forkers.
+
+    The error actually occurs when the stream object is not flushable,
+    but replacing an open stream object that is not flushable with
+    /dev/null is a bad idea since it will create a silent failure.  Replacing
+    a closed handle with /dev/null seems safe.
+
+    """
+
+    if sys.version_info[0:3] >= (3, 6, 4):
+        return True  # Issued fixed in Python 3.6.4+
+
+    if sys.stderr is None:
+        sys.stderr = open(os.devnull, 'w')
+
+    if sys.stdin is None:
+        if options.input_file == '-':
+            print("Trying to read from stdin but stdin seems closed",
+                  file=sys.stderr)
+            return False
+        sys.stdin = open(os.devnull, 'r')
+
+    if sys.stdout is None:
+        if options.output_file == '-':
+            # Can't replace stdout if the user is piping
+            # If this case can even happen, it must be some kind of weird
+            # stream.
+            print(textwrap.dedent("""\
+                Output was set to stdout '-' but the stream attached to
+                stdout does not support the flush() system call.  This
+                will fail."""), file=sys.stderr)
+            return False
+        sys.stdout = open(os.devnull, 'w')
+
+    return True
+
+
+def log_page_orientations(pdfinfo, _log):
+    direction = {0: 'n', 90: 'e',
+                180: 's', 270: 'w'}
+    orientations = []
+    for n, page in enumerate(pdfinfo):
+        angle = page.rotation or 0
+        if angle != 0:
+            orientations.append('{0}{1}'.format(
+                n + 1,
+                direction.get(angle, '')))
+    if orientations:
+        _log.info('Page orientations detected: ' + ' '.join(orientations))
+
+
+def preamble(_log):
+    _log.debug('ocrmypdf ' + VERSION)
+    _log.debug('tesseract ' + tesseract.version())
+    _log.debug('qpdf ' + qpdf.version())
+    if fitz:
+        _log.debug('PyMuPDF ' + fitz.version[0])
+        _log.debug('libmupdf ' + fitz.version[1])
+    else:
+        _log.debug('PyMuPDF not installed')
+
+
+def check_environ(options, _log):
+    old_envvars = (
+        'OCRMYPDF_TESSERACT',
+        'OCRMYPDF_QPDF',
+        'OCRMYPDF_GS',
+        'OCRMYPDF_UNPAPER')
+    for k in old_envvars:
+        if k in os.environ:
+            _log.warning(textwrap.dedent("""\
+                OCRmyPDF no longer uses the environment variable {}.
+                Change PATH to select alternate programs.""".format(k)))
+
+
+def check_input_file(options, _log, start_input_file):
+    if options.input_file == '-':
+        # stdin
+        _log.info('reading file from standard input')
+        with open(start_input_file, 'wb') as stream_buffer:
+            from shutil import copyfileobj
+            copyfileobj(sys.stdin.buffer, stream_buffer)
+    else:
+        try:
+            re_symlink(options.input_file, start_input_file, _log)
+        except FileNotFoundError:
+            _log.error("File not found - " + options.input_file)
+            raise InputFileError()
+
+
+def check_requested_output_file(options, _log):
+    if options.output_file == '-':
+        if sys.stdout.isatty():
+            _log.error(textwrap.dedent("""\
+                Output was set to stdout '-' but it looks like stdout
+                is connected to a terminal.  Please redirect stdout to a
+                file."""))
+            raise BadArgsError()
+    elif not is_file_writable(options.output_file):
+        _log.error(
+            "Output file location (" + options.output_file + ") " +
+            "is not a writable file.")
+        raise OutputFileAccessError()
+
+
+def report_output_file_size(options, _log, input_file, output_file):
+    try:
+        output_size = Path(output_file).stat().st_size
+        input_size = Path(input_file).stat().st_size
+    except FileNotFoundError:
+        return  # Outputting to stream or something
+    ratio = output_size / input_size
+    if ratio < 1.35 or input_size < 25000:
+        return  # Seems fine
+
+    reasons = []
+    if not fitz:
+        reasons.append("The optional dependency PyMuPDF is not installed.")
+    image_preproc = {
+        'deskew',
+        'clean_final',
+        'remove_background',
+        'oversample',
+        'force_ocr'
+    }
+    for arg in image_preproc:
+        attr = getattr(options, arg, None)
+        if not attr:
+            continue
+        reasons.append(
+                "The argument --{} was issued, causing transcoding.".format(
+                        arg.replace('_', '-')))
+
+    if reasons:
+        explanation = (
+            "Possible reasons for this include:\n" + '\n'.join(reasons) + "\n")
+    else:
+        explanation = (
+            "No reason for this increase is known.  Please report this issue.")
+
+    _log.warning(textwrap.dedent("""\
+        The output file size is {:.2f}× larger than the input file.
+        {}
+        """.format(ratio, explanation)))
+
+
+def run_pipeline():
+    options = parser.parse_args()
+    options.verbose_abbreviated_path = 1
+
+    if not check_closed_streams(options):
+        return ExitCode.bad_args
+
+    logger_args = {'verbose': options.verbose, 'quiet': options.quiet}
+
+    _log, _log_mutex = proxy_logger.make_shared_logger_and_proxy(
+        logging_factory, __name__, logger_args)
+    preamble(_log)
+    check_options(options, _log)
+
+    # Complain about qpdf version < 7.0.0
+    # Suppress the warning if in the test suite, since there are no PPAs
+    # for qpdf 7.0.0 for Ubuntu trusty (i.e. Travis)
+    if qpdf.version() < '7.0.0' and not os.environ.get('PYTEST_CURRENT_TEST'):
+        complain(
+            "You are using qpdf version {0} which has known issues including "
+            "security vulnerabilities with certain malformed PDFs. Consider "
+            "upgrading to version 7.0.0 or newer.".format(qpdf.version()))
+
+    if ghostscript.version() == '9.24':
+        complain(
+            "Ghostscript 9.24 contains serious regressions and is not "
+            "supported. Please upgrade to Ghostscript 9.25 or use an older "
+            "version."
+        )
+        return ExitCode.missing_dependency
+
+    # Any changes to options will not take effect for options that are already
+    # bound to function parameters in the pipeline. (For example
+    # options.input_file, options.pdf_renderer are already bound.)
+    if not options.jobs:
+        options.jobs = available_cpu_count()
+
+    # Performance is improved by setting Tesseract to single threaded. In tests
+    # this gives better throughput than letting a smaller number of Tesseract
+    # jobs run multithreaded.
+    if tesseract.v4():
+        os.environ.setdefault('OMP_THREAD_LIMIT', '1')
+    check_environ(options, _log)
+    if os.environ.get('PYTEST_CURRENT_TEST'):
+        os.environ['_OCRMYPDF_TEST_INFILE'] = options.input_file
+
+    try:
+        work_folder = mkdtemp(prefix="com.github.ocrmypdf.")
+        options.history_file = os.path.join(
+            work_folder, 'ruffus_history.sqlite')
+        start_input_file = os.path.join(
+            work_folder, 'origin')
+
+        check_input_file(options, _log, start_input_file)
+        check_requested_output_file(options, _log)
+
+        manager = JobContextManager()
+        manager.register('JobContext', JobContext)  # pylint: disable=no-member
+        manager.start()
+
+        context = manager.JobContext()  # pylint: disable=no-member
+        context.set_options(options)
+        context.set_work_folder(work_folder)
+
+        build_pipeline(options, work_folder, _log, context)
+        atexit.register(cleanup_working_files, work_folder, options)
+        cmdline.run(options)
+    except ruffus_exceptions.RethrownJobError as e:
+        if options.verbose:
+            _log.debug(str(e))  # stringify exception so logger doesn't have to
+        exceptions = e.job_exceptions
+        exitcode = traverse_ruffus_exception(exceptions, options, _log)
+        if exitcode is None:
+            _log.error("Unexpected ruffus exception: " + str(e))
+            _log.error(repr(e))
+            return ExitCode.other_error
+        return exitcode
+    except ExitCodeException as e:
+        return e.exit_code
+    except Exception as e:
+        _log.error(e)
+        return ExitCode.other_error
+
+    if options.flowchart:
+        _log.info("Flowchart saved to {}".format(options.flowchart))
+        return ExitCode.ok
+    elif options.output_file == '-':
+        _log.info("Output sent to stdout")
+    elif os.path.samefile(options.output_file, os.devnull):
+        pass  # Say nothing when sending to dev null
+    else:
+        if options.output_type.startswith('pdfa'):
+            pdfa_info = file_claims_pdfa(options.output_file)
+            if pdfa_info['pass']:
+                msg = 'Output file is a {} (as expected)'
+                _log.info(msg.format(pdfa_info['conformance']))
+            else:
+                msg = 'Output file is okay but is not PDF/A (seems to be {})'
+                _log.warning(msg.format(pdfa_info['conformance']))
+                return ExitCode.invalid_output_pdf
+        if not qpdf.check(options.output_file, _log):
+            _log.warning('Output file: The generated PDF is INVALID')
+            return ExitCode.invalid_output_pdf
+
+        report_output_file_size(options, _log, start_input_file,
+                                options.output_file)
+
+    pdfinfo = context.get_pdfinfo()
+    if options.verbose:
+        from pprint import pformat
+        _log.debug(pformat(pdfinfo))
+
+    log_page_orientations(pdfinfo, _log)
+
+    return ExitCode.ok
+
+
+if __name__ == '__main__':
+    sys.exit(run_pipeline())

src/ocrmypdf/_unicodefun.py

@@ -0,0 +1,113 @@
+# Copyright (c) 2014, Armin Ronacher
+#
+# Copyright (c) 2017, James R Barlow
+#
+# Some rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are
+# met:
+#
+#     * Redistributions of source code must retain the above copyright
+#       notice, this list of conditions and the following disclaimer.
+#
+#     * Redistributions in binary form must reproduce the above
+#       copyright notice, this list of conditions and the following
+#       disclaimer in the documentation and/or other materials provided
+#       with the distribution.
+#
+#     * The names of the contributors may not be used to endorse or
+#       promote products derived from this software without specific
+#       prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+import os
+import sys
+import codecs
+
+
+def verify_python3_env():
+    """Ensures that the environment is good for unicode on Python 3."""
+
+    # PEP 538 changes in Python 3.7 should make this wrangling unnecessary
+    if sys.version_info[0:3] >= (3, 7, 0):
+        return
+
+    try:
+        import locale
+        fs_enc = codecs.lookup(locale.getpreferredencoding()).name
+    except Exception:
+        fs_enc = 'ascii'
+    if fs_enc != 'ascii':
+        return
+
+    extra = ''
+    if os.name == 'posix':
+        import subprocess
+        rv = subprocess.Popen(['locale', '-a'], stdout=subprocess.PIPE,
+                              stderr=subprocess.PIPE).communicate()[0]
+        good_locales = set()
+        has_c_utf8 = False
+
+        # Make sure we're operating on text here.
+        if isinstance(rv, bytes):
+            rv = rv.decode('ascii', 'replace')
+
+        for line in rv.splitlines():
+            locale = line.strip()
+            if locale.lower().endswith(('.utf-8', '.utf8')):
+                good_locales.add(locale)
+                if locale.lower() in ('c.utf8', 'c.utf-8'):
+                    has_c_utf8 = True
+
+        extra += '\n\n'
+        if not good_locales:
+            extra += (
+                'Additional information: on this system no suitable UTF-8\n'
+                'locales were discovered.  This most likely requires resolving\n'
+                'by reconfiguring the locale system.'
+            )
+        elif has_c_utf8:
+            extra += (
+                'This system supports the C.UTF-8 locale which is recommended.\n'
+                'You might be able to resolve your issue by exporting the\n'
+                'following environment variables:\n\n'
+                '    export LC_ALL=C.UTF-8\n'
+                '    export LANG=C.UTF-8'
+            )
+        else:
+            extra += (
+                'This system lists a couple of UTF-8 supporting locales that\n'
+                'you can pick from.  The following suitable locales were\n'
+                'discovered: %s'
+            ) % ', '.join(sorted(good_locales))
+
+        bad_locale = None
+        for locale in os.environ.get('LC_ALL'), os.environ.get('LANG'):
+            if locale and locale.lower().endswith(('.utf-8', '.utf8')):
+                bad_locale = locale
+            if locale is not None:
+                break
+        if bad_locale is not None:
+            extra += (
+                '\nocrmypdf discovered that you exported a UTF-8 locale\n'
+                'but the locale system could not pick up from it because\n'
+                'it does not exist.  The exported locale is "%s" but it\n'
+                'is not supported'
+            ) % bad_locale
+
+    raise RuntimeError('ocrmypdf will abort further execution because Python 3 '
+                       'was configured to use ASCII as encoding for the '
+                       'environment.' + extra)

src/ocrmypdf/exceptions.py

@@ -0,0 +1,81 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+
+from enum import IntEnum
+
+class ExitCode(IntEnum):
+    ok = 0
+    bad_args = 1
+    input_file = 2
+    missing_dependency = 3
+    invalid_output_pdf = 4
+    file_access_error = 5
+    already_done_ocr = 6
+    child_process_error = 7
+    encrypted_pdf = 8
+    invalid_config = 9
+    other_error = 15
+    ctrl_c = 130
+
+
+class ExitCodeException(Exception):
+    exit_code = ExitCode.other_error
+
+
+class BadArgsError(ExitCodeException):
+    exit_code = ExitCode.bad_args
+
+
+class PdfMergeFailedError(ExitCodeException):
+    exit_code = ExitCode.input_file
+
+
+class MissingDependencyError(ExitCodeException):
+    exit_code = ExitCode.missing_dependency
+
+
+class UnsupportedImageFormatError(ExitCodeException):
+    exit_code = ExitCode.input_file
+
+
+class DpiError(ExitCodeException):
+    exit_code = ExitCode.input_file
+
+
+class OutputFileAccessError(ExitCodeException):
+    exit_code = ExitCode.file_access_error
+
+
+class PriorOcrFoundError(ExitCodeException):
+    exit_code = ExitCode.already_done_ocr
+
+
+class InputFileError(ExitCodeException):
+    exit_code = ExitCode.input_file
+
+
+class SubprocessOutputError(ExitCodeException):
+    exit_code = ExitCode.child_process_error
+
+
+class EncryptedPdfError(ExitCodeException):
+    exit_code = ExitCode.encrypted_pdf
+
+
+class TesseractConfigError(ExitCodeException):
+    exit_code = ExitCode.invalid_config

src/ocrmypdf/exec/__init__.py

@@ -0,0 +1,60 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+"""Wrappers to manage subprocess calls"""
+
+import os
+import re
+import sys
+from subprocess import run, STDOUT, PIPE, CalledProcessError
+from ..exceptions import MissingDependencyError
+
+
+
+def get_version(program, *,
+        version_arg='--version', regex=r'(\d+(\.\d+)*)'):
+    "Get the version of the specified program"
+    args_prog = [
+        program,
+        version_arg
+    ]
+    try:
+        proc = run(
+            args_prog, close_fds=True, universal_newlines=True,
+            stdout=PIPE, stderr=STDOUT, check=True)
+        output = proc.stdout
+    except FileNotFoundError as e:
+        raise MissingDependencyError(
+            "Could not find program '{}' on the PATH".format(
+                program)) from e
+    except CalledProcessError as e:
+        if e.returncode < 0:
+            raise MissingDependencyError(
+                "Ran program '{}' but it exited with an error:\n{}".format(
+                    program, e.output)) from e
+        raise MissingDependencyError(
+            "Could not find program '{}' on the PATH".format(
+                program)) from e
+    try:
+        version = re.match(regex, output.strip()).group(1)
+    except AttributeError as e:
+        raise MissingDependencyError(
+            ("The program '{}' did not report its version. "
+            "Message was:\n{}").format(program, output)
+        )
+
+    return version

src/ocrmypdf/exec/ghostscript.py

@@ -0,0 +1,204 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from tempfile import NamedTemporaryFile
+from subprocess import run, PIPE, STDOUT, CalledProcessError
+from shutil import copy
+from functools import lru_cache
+import re
+import sys
+from PIL import Image
+from . import get_version
+from ..exceptions import SubprocessOutputError, MissingDependencyError
+from ..helpers import fspath
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('gs')
+
+
+def jpeg_passthrough_available():
+    """
+    Returns True if the installed version of Ghostscript supports JPEG passthru
+
+    Prior to 9.23, Ghostscript decode and re-encoded JPEGs internally. In 9.23
+    it gained the ability to keep JPEGs unmodified. However, the 9.23
+    implementation was buggy and would deletes the last two bytes of images in
+    some cases, as reported here.
+    https://bugs.ghostscript.com/show_bug.cgi?id=699216
+
+    The issue was fixed for 9.24, hence that is the first version we consider
+    the feature available. (However, we don't use 9.24 at all, so the first
+    version that allows JPEG passthrough is 9.25.
+
+    Regardless, in ocrmypdf 6.x we are ignoring this new feature entirely to
+    avoid new behavior.
+    """
+    return False
+
+
+def _gs_error_reported(stream):
+    return re.search(r'error', stream, flags=re.IGNORECASE)
+
+
+def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
+                  pageno=1, page_dpi=None):
+    """
+    Rasterize one page of a PDF at resolution (xres, yres) in canvas units.
+
+    The image is sized to match the integer pixels dimensions implied by
+    (xres, yres) even if those numbers are noninteger. The image's DPI will
+     be overridden with the values in page_dpi.
+
+    :param input_file: pathlike
+    :param output_file: pathlike
+    :param xres: resolution at which to rasterize page
+    :param yres:
+    :param raster_device:
+    :param log:
+    :param pageno: page number to rasterize (beginning at page 1)
+    :param page_dpi: resolution tuple (x, y) overriding output image DPI
+    :return:
+    """
+    res = xres, yres
+    int_res = round(xres), round(yres)
+    if not page_dpi:
+        page_dpi = res
+    with NamedTemporaryFile(delete=True) as tmp:
+        args_gs = [
+            'gs',
+            '-dQUIET',
+            '-dSAFER',
+            '-dBATCH',
+            '-dNOPAUSE',
+            '-sDEVICE=%s' % raster_device,
+            '-dFirstPage=%i' % pageno,
+            '-dLastPage=%i' % pageno,
+            '-r{0}x{1}'.format(str(int_res[0]), str(int_res[1])),
+            '-o', tmp.name,
+            fspath(input_file)
+        ]
+
+        p = run(args_gs, stdout=PIPE, stderr=STDOUT,
+                universal_newlines=True)
+        if _gs_error_reported(p.stdout):
+            log.error(p.stdout)
+        else:
+            log.debug(p.stdout)
+
+        if p.returncode != 0:
+            log.error('Ghostscript rasterizing failed')
+            raise SubprocessOutputError()
+
+        # Ghostscript only accepts integers for output resolution
+        # if the resolution happens to be fractional, then the discrepancy
+        # would change the size of the output page, especially if the DPI
+        # is quite low. Resize the image to the expected size
+
+        tmp.seek(0)
+        with Image.open(tmp) as im:
+            expected_size = round(im.size[0] / int_res[0] * res[0]), \
+                            round(im.size[1] / int_res[1] * res[1])
+            if expected_size != im.size or page_dpi != (xres, yres):
+                log.debug(
+                    "Ghostscript: resize output image {} -> {}".format(
+                        im.size, expected_size))
+                im.resize(expected_size).save(
+                    fspath(output_file), dpi=page_dpi)
+            else:
+                copy(tmp.name, fspath(output_file))
+
+
+def generate_pdfa(pdf_pages, output_file, compression, log,
+                  threads=1, pdf_version='1.5', pdfa_part='2'):
+    compression_args = []
+    if compression == 'jpeg':
+        compression_args = [
+            "-dAutoFilterColorImages=false",
+            "-dColorImageFilter=/DCTEncode",
+            "-dAutoFilterGrayImages=false",
+            "-dGrayImageFilter=/DCTEncode",
+        ]
+    elif compression == 'lossless':
+        compression_args = [
+            "-dAutoFilterColorImages=false",
+            "-dColorImageFilter=/FlateEncode",
+            "-dAutoFilterGrayImages=false",
+            "-dGrayImageFilter=/FlateEncode",
+        ]
+    else:
+        compression_args = [
+            "-dAutoFilterColorImages=true",
+            "-dAutoFilterGrayImages=true",
+        ]
+
+    # Older versions of Ghostscript expect a leading slash in
+    # sColorConversionStrategy, newer ones should not have it. See Ghostscript
+    # git commit fe1c025d.
+    strategy = 'RGB' if version() >= '9.19' else '/RGB'
+
+    if version() >= '9.23':
+        # 9.23: new feature JPEG passthrough is broken in some cases, best to
+        # disable it always
+        # https://bugs.ghostscript.com/show_bug.cgi?id=699216
+        # fixed in 9.24, but to avoid changing expected behavior we disable it
+        # for ocrmypdf 6.x
+        compression_args.append('-dPassThroughJPEGImages=false')
+
+    with NamedTemporaryFile(delete=True) as gs_pdf:
+        args_gs = [
+            "gs",
+            "-dQUIET",
+            "-dBATCH",
+            "-dNOPAUSE",
+            "-dCompatibilityLevel=" + str(pdf_version),
+            "-dNumRenderingThreads=" + str(threads),
+            "-sDEVICE=pdfwrite",
+            "-dAutoRotatePages=/None",
+            "-sColorConversionStrategy=" + strategy,
+            "-sProcessColorModel=DeviceRGB"
+        ] + compression_args + [
+            "-dJPEGQ=95",
+            "-dPDFA=" + pdfa_part,
+            "-dPDFACompatibilityPolicy=1",
+            "-sOutputFile=" + gs_pdf.name,
+        ]
+        args_gs.extend(pdf_pages)
+        p = run(args_gs, stdout=PIPE, stderr=STDOUT,
+                universal_newlines=True)
+
+        if _gs_error_reported(p.stdout):
+            log.error(p.stdout)
+        elif 'overprint mode not set' in p.stdout:
+            # Unless someone is going to print PDF/A documents on a
+            # magical sRGB printer I can't see the removal of overprinting
+            # being a problem....
+            log.debug(
+                "Ghostscript had to remove PDF 'overprinting' from the "
+                "input file to complete PDF/A conversion. "
+                )
+        else:
+            log.debug(p.stdout)
+
+        if p.returncode == 0:
+            # Ghostscript does not change return code when it fails to create
+            # PDF/A - check PDF/A status elsewhere
+            copy(gs_pdf.name, output_file)
+        else:
+            log.error('Ghostscript PDF/A rendering failed')
+            raise SubprocessOutputError()

src/ocrmypdf/exec/qpdf.py

@@ -0,0 +1,227 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from subprocess import CalledProcessError, STDOUT, PIPE, run
+from functools import lru_cache
+import sys
+import os
+import re
+import resource
+
+from ..exceptions import InputFileError, SubprocessOutputError, \
+    MissingDependencyError, EncryptedPdfError
+from . import  get_version
+from ..helpers import re_symlink
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('qpdf', regex=r'qpdf version (.+)')    
+
+
+def check(input_file, log=None):
+    args_qpdf = [
+        'qpdf',
+        '--check',
+        input_file
+    ]
+
+    if log is None:
+        import logging as log
+
+    try:
+        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
+            check=True)
+    except CalledProcessError as e:
+        if e.returncode == 2:
+            log.error("{0}: not a valid PDF, and could not repair it.".format(
+                input_file))
+            log.error("Details:")
+            log.error(e.output)
+        elif e.returncode == 3:
+            log.info("qpdf --check returned warnings:")
+            log.info(e.output)
+        else:
+            log.warning(e.output)
+        return False
+    return True
+
+
+def _probably_encrypted(e):
+    """qpdf can report a false positive "file is encrypted" message for damaged
+    files - suppress this"""
+    return e.returncode == 2 and \
+        'invalid password' in e.output and \
+        'file is damaged' not in e.output
+
+
+def repair(input_file, output_file, log):
+    args_qpdf = [
+        'qpdf', input_file, output_file
+    ]
+    try:
+        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
+            check=True)
+    except CalledProcessError as e:
+        if e.returncode == 3 and e.output.find("operation succeeded"):
+            log.debug('qpdf found and fixed errors: ' + e.output)
+            return
+
+        if _probably_encrypted(e):
+            raise EncryptedPdfError() from e
+        elif e.returncode == 2:
+            log.error("{0}: not a valid PDF, and could not repair it.".format(
+                      input_file))
+            log.error("Details: " + e.output)
+            raise InputFileError() from e
+        else:
+            log.error("{0}: unknown error".format(
+                      input_file))
+            log.error(e.output)
+            raise SubprocessOutputError() from e
+
+
+def get_npages(input_file, log):
+    try:
+        pages = run(
+            ['qpdf', '--show-npages', input_file],
+            universal_newlines=True, check=True, stdout=PIPE, stderr=STDOUT)
+    except CalledProcessError as e:
+        if e.returncode == 2 and e.output.find('No such file'):
+            log.error(e.output)
+            raise InputFileError() from e
+    return int(pages)
+
+
+def split_pages(input_file, work_folder, npages):
+    """Split multipage PDF into individual pages.
+
+    Incredibly enough, this multiple process approach is about 70 times
+    faster than using Ghostscript.
+    """
+    for n in range(int(npages)):
+        args_qpdf = [
+            'qpdf', input_file,
+            '--pages', input_file, '{0}'.format(n + 1), '--',
+            os.path.join(work_folder, '{0:06d}.page.pdf'.format(n + 1))
+        ]
+        run(args_qpdf, check=True)
+
+
+def extract_page(input_file, output_file, pageno):
+    args_qpdf = [
+        'qpdf', input_file,
+        '--pages', input_file, '{0}'.format(pageno + 1), '--',
+        output_file
+    ]
+    run(args_qpdf, check=True)
+    
+
+def _merge_inner(input_files, output_file, min_version=None, log=None):
+    """Merge the list of input files (all filenames) into the output file.
+
+    The input files may contain one or more pages.
+    """
+
+    # Single page 'merges' should still be attempted to that the same error
+    # checking is applied to single page case
+
+    version_arg = ['--min-version={}'.format(min_version)] \
+                  if min_version else []
+
+    if log is None:
+        import logging as log
+
+    args_qpdf = [
+        'qpdf'
+    ] + version_arg + [
+        input_files[0], '--pages'
+    ] + input_files + ['--', output_file]
+
+    try:
+        run(args_qpdf, check=True, stderr=PIPE, universal_newlines=True)
+    except CalledProcessError as e:
+        if e.returncode == 3 and \
+                e.stderr.find("unknown token while reading object") and \
+                e.stderr.find("operation succeeded"):
+            # Only whitelist the 'unknown token' problem (decimal/string issue)
+            # qpdf issue #165
+            log.warning('qpdf found and fixed errors: ' + e.stderr)
+            return
+        raise e from e
+
+
+def merge(input_files, output_file, min_version=None, log=None, max_files=None):
+    """Merge the list of input files (all filenames) into the output file.
+
+    The input files may contain one or more pages.
+
+    """
+    # qpdf requires that every file that contributes to the output has a file 
+    # descriptor that remains open. That means, given our approach of one 
+    # intermediate PDF per, we can practically hit the number of file 
+    # descriptors. 
+
+    if max_files is None or max_files < 2:
+        # Find out how many open file descriptors we can get away with
+        ulimits = resource.getrlimit(resource.RLIMIT_NOFILE)
+        max_open_files = ulimits[0]
+        max_files = max_open_files // 2  # Conservative guess
+
+    # We'll write things alongside the output file
+    output_dir = os.path.dirname(output_file)
+
+    import random
+    import string    
+
+    def randstr():
+        return ''.join(random.sample(string.ascii_lowercase, 6))
+
+    # How many files to grab at once, merging all their contents
+    step_size = max_files
+
+    workqueue = input_files.copy()
+    counter = 1
+    next_workqueue = []
+    while len(workqueue) > 1 or len(next_workqueue) > 0:
+        # Take n files out of the queue
+        n = min(step_size, len(workqueue))
+        job = workqueue[0:n]
+        del workqueue[0:n]
+        log.debug('merging ' + repr(job))
+
+        # Merge them into 1 file, which will contain n^depth pages
+        merge_file = os.path.join(
+            output_dir, "merge-{:06d}-{}.pdf".format(counter, randstr()))
+        counter += 1
+        _merge_inner(job, merge_file, min_version=min_version, log=log)
+
+        # On the next 
+        next_workqueue.append(merge_file)
+        log.debug('next_workqueue ' + repr(next_workqueue))
+
+        # If we're out of things to do in this queue, move on to the next
+        # queue. On the counter-th pass of the workqueue we can chew through
+        # (step_size)**N pages, so on most systems the second pass finishes
+        # the job.
+        if len(workqueue) == 0:
+            workqueue = next_workqueue
+            next_workqueue = []
+
+    re_symlink(workqueue.pop(), output_file, log)
+
+

src/ocrmypdf/exec/tesseract.py

@@ -0,0 +1,373 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+import sys
+import os
+import re
+import shutil
+from functools import lru_cache
+from collections import namedtuple
+from textwrap import dedent
+from subprocess import PIPE, CalledProcessError, \
+    TimeoutExpired, check_output, STDOUT
+from contextlib import suppress
+
+import PyPDF2 as pypdf
+
+from ..exceptions import MissingDependencyError, TesseractConfigError
+from ..helpers import page_number
+from . import get_version
+
+OrientationConfidence = namedtuple(
+    'OrientationConfidence',
+    ('angle', 'confidence'))
+
+HOCR_TEMPLATE = """<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ <head>
+  <title></title>
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
+  <meta name='ocr-system' content='tesseract 3.05.00' />
+  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
+</head>
+<body>
+  <div class='ocr_page' id='page_1' title='image "_blank.png"; bbox 0 0 {0} {1}; ppageno 0'>
+  </div>
+ </body>
+</html>
+"""
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('tesseract', regex=r'tesseract\s(.+)')
+
+
+def v4():
+    "Is this Tesseract v4.0?"
+    return version() >= '4'
+
+
+@lru_cache(maxsize=1)
+def has_textonly_pdf():
+    """Does Tesseract have textonly_pdf capability?
+    
+    Available in 3.05.01, and v4.00.00alpha since January 2017. Best to 
+    parse the parameter list
+    """
+    args_tess = [
+        'tesseract',
+        '--print-parameters',
+        'pdf'
+    ]
+    params = ''
+    try:
+        params = check_output(
+                args_tess, universal_newlines=True, stderr=STDOUT)
+    except CalledProcessError as e:
+        print("Could not --print-parameters from tesseract",
+              file=sys.stderr)
+        raise MissingDependencyError from e
+    if 'textonly_pdf' in params:
+        return True
+    return False
+
+
+def psm():
+    "If Tesseract 4.0, use argument --psm instead of -psm"
+    return '--psm' if v4() else '-psm'
+
+
+@lru_cache(maxsize=1)
+def languages():
+    args_tess = [
+        'tesseract',
+        '--list-langs'
+    ]
+    try:
+        langs = check_output(
+                args_tess, universal_newlines=True, stderr=STDOUT)
+    except CalledProcessError as e:
+        msg = dedent("""Tesseract failed to report available languages.
+        Output from Tesseract:
+        -----------
+        """)
+        msg += e.output
+        print(msg, file=sys.stderr)
+        raise MissingDependencyError from e
+    return set(lang.strip() for lang in langs.splitlines()[1:])
+
+
+def tess_base_args(langs, engine_mode):
+    args = [
+        'tesseract',
+    ]
+    if langs:
+        args.extend(['-l', '+'.join(langs)])
+    if engine_mode is not None and v4():
+        args.extend(['--oem', str(engine_mode)])
+    return args
+
+
+def get_orientation(input_file, language: list, engine_mode, timeout: float,
+                    log):
+    args_tesseract = tess_base_args(['osd'], engine_mode) + [
+        psm(), '0',
+        input_file,
+        'stdout'
+    ]
+
+    try:
+        stdout = check_output(
+            args_tesseract, stderr=STDOUT, timeout=timeout)
+    except TimeoutExpired:
+        return OrientationConfidence(angle=0, confidence=0.0)
+    except CalledProcessError as e:
+        tesseract_log_output(log, e.output, input_file)
+        if (b'Too few characters. Skipping this page' in e.output or
+                b'Image too large' in e.output):
+            return OrientationConfidence(0, 0)
+        raise e from e
+    else:
+        osd = {}
+        for line in stdout.decode().splitlines():
+            line = line.strip()
+            parts = line.split(':', maxsplit=2)
+            if len(parts) == 2:
+                osd[parts[0].strip()] = parts[1].strip()
+
+        angle = int(osd.get('Orientation in degrees', 0))
+        if 'Orientation' in osd:
+            # Tesseract < 3.04.01
+            # reports "Orientation in degrees" as a counterclockwise angle
+            # We keep it clockwise
+            assert 'Rotate' not in osd
+            angle = -angle % 360
+        else:
+            # Tesseract >= 3.04.01
+            # reports "Orientation in degrees" as a clockwise angle
+            assert 'Rotate' in osd
+
+        oc = OrientationConfidence(
+            angle=angle,
+            confidence=float(osd.get('Orientation confidence', 0)))
+        return oc
+
+
+def tesseract_log_output(log, stdout, input_file):
+    prefix = "{0:4d}: [tesseract] ".format(page_number(input_file))
+
+    try:
+        text = stdout.decode()
+    except UnicodeDecodeError:
+        log.error(prefix + "command line output was not utf-8. " +
+            "This usually means Tesseract's language packs do not match "
+            "the installed version of Tesseract.")
+        text = stdout.decode('utf-8', 'backslashreplace')
+
+    lines = text.splitlines()
+    for line in lines:
+        if line.startswith("Tesseract Open Source"):
+            continue
+        elif line.startswith("Warning in pixReadMem"):
+            continue
+        elif 'diacritics' in line:
+            log.warning(prefix + "lots of diacritics - possibly poor OCR")
+        elif line.startswith('OSD: Weak margin'):
+            log.warning(prefix + "unsure about page orientation")
+        elif 'error' in line.lower() or 'exception' in line.lower():
+            log.error(prefix + line.strip())
+        elif 'warning' in line.lower():
+            log.warning(prefix + line.strip())
+        elif 'read_params_file' in line.lower():
+            log.error(prefix + line.strip())
+        else:
+            log.info(prefix + line.strip())
+
+
+def page_timedout(log, input_file):
+    prefix = "{0:4d}: [tesseract] ".format(page_number(input_file))
+    log.warning(prefix + " took too long to OCR - skipping")
+
+
+def _generate_null_hocr(output_hocr, output_sidecar, image):
+    """Produce a .hocr file that reports no text detected on a page that is
+    the same size as the input image."""
+    from PIL import Image
+
+    im = Image.open(image)
+    w, h = im.size
+
+    with open(output_hocr, 'w', encoding="utf-8") as f:
+        f.write(HOCR_TEMPLATE.format(w, h))
+    with open(output_sidecar, 'w', encoding='utf-8') as f:
+        f.write('[skipped page]')
+
+
+def generate_hocr(input_file, output_files, language: list, engine_mode,
+                  tessconfig: list,
+                  timeout: float, pagesegmode: int, user_words, user_patterns,
+                  log):
+
+    output_hocr = next(o for o in output_files if o.endswith('.hocr'))
+    output_sidecar = next(o for o in output_files if o.endswith('.txt'))
+    prefix = os.path.splitext(output_hocr)[0]
+
+    args_tesseract = tess_base_args(language, engine_mode)
+
+    if pagesegmode is not None:
+        args_tesseract.extend([psm(), str(pagesegmode)])
+
+    if user_words:
+        args_tesseract.extend(['--user-words', user_words])
+
+    if user_patterns:
+        args_tesseract.extend(['--user-patterns', user_patterns])
+
+    # Reminder: test suite tesseract spoofers will break after any changes
+    # to the number of order parameters here
+    # Tesseract 3.04 requires the order here to be "hocr txt" and will fail
+    # on "txt hocr"
+
+    args_tesseract.extend([
+        input_file,
+        prefix,
+        'hocr',
+        'txt'
+    ] + tessconfig)
+    try:
+        log.debug(args_tesseract)
+        stdout = check_output(
+            args_tesseract, stderr=STDOUT,
+            timeout=timeout)
+    except TimeoutExpired:
+        # Generate a HOCR file with no recognized text if tesseract times out
+        # Temporary workaround to hocrTransform not being able to function if
+        # it does not have a valid hOCR file.
+        page_timedout(log, input_file)
+        _generate_null_hocr(output_hocr, output_sidecar, input_file)
+    except CalledProcessError as e:
+        tesseract_log_output(log, e.output, input_file)
+        if b'read_params_file: parameter not found' in e.output:
+            raise TesseractConfigError() from e
+        if b'Image too large' in e.output:
+            _generate_null_hocr(output_hocr, output_sidecar, input_file)
+            return
+
+        raise e from e
+    else:
+        tesseract_log_output(log, stdout, input_file)
+        # The sidecar text file will get the suffix .txt; rename it to
+        # whatever caller wants it named
+        if os.path.exists(prefix + '.txt'):
+            shutil.move(prefix + '.txt', output_sidecar)
+
+
+def use_skip_page(text_only, skip_pdf, output_pdf, output_text):
+    with open(output_text, 'w') as f:
+        f.write('[skipped page]')
+
+    if not text_only:
+        with suppress(FileNotFoundError):
+            os.remove(output_pdf)  # In case it was partially created
+        os.symlink(skip_pdf, output_pdf)
+        return
+
+    # For text only we must create a blank page with dimensions identical
+    # to the skip page because this is equivalent to a page with no text
+
+    pdf_in = pypdf.PdfFileReader(skip_pdf)
+    page0 = pdf_in.pages[0]
+
+    with open(output_pdf, 'wb') as out:
+        pdf_out = pypdf.PdfFileWriter()
+        w, h = page0.mediaBox.getWidth(), page0.mediaBox.getHeight()
+        # If skip page has a /Rotate key, replicate the rotation
+        rotation = int(page0.get('/Rotate', 0))
+        if rotation % 180 == 90:
+            w, h = h, w
+        pdf_out.addBlankPage(w, h)
+        pdf_out.write(out)
+
+
+def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
+                 language: list, engine_mode, text_only: bool,
+                 tessconfig: list, timeout: float, pagesegmode: int,
+                 user_words, user_patterns, log):
+    '''Use Tesseract to render a PDF.
+
+    input_image -- image to analyze
+    skip_pdf -- if we time out, use this file as output
+    output_pdf -- file to generate
+    output_text -- OCR text file
+    language -- list of languages to consider
+    engine_mode -- engine mode argument for tess v4
+    text_only -- enable tesseract text only mode?
+    tessconfig -- tesseract configuration
+    timeout -- timeout (seconds)
+    log -- logger object
+    '''
+
+    args_tesseract = tess_base_args(language, engine_mode)
+
+    if pagesegmode is not None:
+        args_tesseract.extend([psm(), str(pagesegmode)])
+
+    if text_only:
+        args_tesseract.extend(['-c', 'textonly_pdf=1'])
+
+    if user_words:
+        args_tesseract.extend(['--user-words', user_words])
+
+    if user_patterns:
+        args_tesseract.extend(['--user-patterns', user_patterns])
+
+    prefix = os.path.splitext(output_pdf)[0]  # Tesseract appends suffixes
+
+    # Reminder: test suite tesseract spoofers might break after any changes
+    # to the number of order parameters here
+
+    args_tesseract.extend([
+        input_image,
+        prefix,
+        'pdf',
+        'txt'
+    ] + tessconfig)
+
+    try:
+        log.debug(args_tesseract)
+        stdout = check_output(
+            args_tesseract, stderr=STDOUT,
+            timeout=timeout)
+        if os.path.exists(prefix + '.txt'):
+            shutil.move(prefix + '.txt', output_text)
+    except TimeoutExpired:
+        page_timedout(log, input_image)
+        use_skip_page(text_only, skip_pdf, output_pdf, output_text)
+    except CalledProcessError as e:
+        tesseract_log_output(log, e.output, input_image)
+        if b'read_params_file: parameter not found' in e.output:
+            raise TesseractConfigError() from e
+
+        if b'Image too large' in e.output:
+            use_skip_page(text_only, skip_pdf, output_pdf, output_text)
+            return
+        raise e from e
+    else:
+        tesseract_log_output(log, stdout, input_image)

src/ocrmypdf/exec/unpaper.py

@@ -1,27 +1,30 @@
-#!/usr/bin/env python3
 # © 2015 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
 # unpaper documentation:
 # https://github.com/Flameeyes/unpaper/blob/master/doc/basic-concepts.md
 
-from subprocess import Popen, PIPE
+from subprocess import CalledProcessError, STDOUT, check_output
 from tempfile import NamedTemporaryFile
 import sys
 import os
 from functools import lru_cache
-from . import ExitCode
-
-
-@lru_cache(maxsize=1)
-def version():
-    args_unpaper = [
-        'unpaper',
-        '--version'
-    ]
-    p_unpaper = Popen(args_unpaper, close_fds=True, universal_newlines=True,
-                      stdout=PIPE, stderr=PIPE)
-    version, _ = p_unpaper.communicate(timeout=5)
-
-    return version.strip()
+from ..exceptions import MissingDependencyError
+from . import get_version
 
 
 try:
@@ -31,6 +34,11 @@ except ImportError:
     raise
 
 
+@lru_cache(maxsize=1)
+def version():
+    return get_version('unpaper')
+
+
 def run(input_file, output_file, dpi, log, mode_args):
     args_unpaper = [
         'unpaper',
@@ -48,17 +56,17 @@ def run(input_file, output_file, dpi, log, mode_args):
                 im = im.convert(mode='1')
             else:
                 im = im.convert(mode='RGB')
-        except IOError:
-            log.error(
-                    "Could not convert image with type " + im.mode)
-            sys.exit(ExitCode.missing_dependency)
+        except IOError as e:
+            log.error("Could not convert image with type " + im.mode)
+            im.close()
+            raise MissingDependencyError() from e
 
     try:
         suffix = SUFFIXES[im.mode]
     except KeyError:
-        log.error(
-                "Failed to convert image to a supported format.")
-        sys.exit(ExitCode.missing_dependency)
+        log.error("Failed to convert image to a supported format.")
+        im.close()
+        raise MissingDependencyError() from e
 
     with NamedTemporaryFile(suffix=suffix) as input_pnm, \
             NamedTemporaryFile(suffix=suffix, mode="r+b") as output_pnm:
@@ -68,31 +76,23 @@ def run(input_file, output_file, dpi, log, mode_args):
         os.unlink(output_pnm.name)
 
         args_unpaper.extend([input_pnm.name, output_pnm.name])
-        p_unpaper = Popen(
-            args_unpaper, close_fds=True,
-            universal_newlines=True, stdout=PIPE, stderr=PIPE
-            )
-        out, err = p_unpaper.communicate()
-        log.debug(out)
-        log.debug(err)
-
-        Image.open(output_pnm.name).save(output_file)
-
-
-def deskew(input_file, output_file, dpi, log):
-    run(input_file, output_file, dpi, log, [
-        '--mask-scan-size', '100',  # don't blank out narrow columns
-        '--no-border-align',  # don't align visible content to borders
-        '--no-mask-center',   # don't center visible content within page
-        '--no-grayfilter',    # don't remove light gray areas
-        '--no-blackfilter',   # don't remove solid black areas
-        '--no-noisefilter',   # don't remove salt and pepper noise
-        '--no-blurfilter'     # don't remove blurry objects/debris
-    ])
+        try:
+            stdout = check_output(
+                args_unpaper, close_fds=True,
+                universal_newlines=True, stderr=STDOUT,
+                )
+        except CalledProcessError as e:
+            log.debug(e.output)
+            raise e from e
+        else:
+            log.debug(stdout)
+            # unpaper sets dpi to 72
+            Image.open(output_pnm.name).save(output_file, dpi=(dpi, dpi))
 
 
 def clean(input_file, output_file, dpi, log):
     run(input_file, output_file, dpi, log, [
+        '--layout', 'none',
         '--mask-scan-size', '100',  # don't blank out narrow columns
         '--no-border-align',  # don't align visible content to borders
         '--no-mask-center',   # don't center visible content within page

src/ocrmypdf/helpers.py

@@ -0,0 +1,172 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from PIL import Image
+
+from functools import partial
+from collections.abc import Iterable
+from contextlib import suppress, contextmanager
+from pathlib import Path
+import sys
+import os
+import multiprocessing
+import warnings
+
+
+def re_symlink(input_file, soft_link_name, log=None):
+    """
+    Helper function: relinks soft symbolic link if necessary
+    """
+
+    if log is None:
+        prdebug = partial(print, file=sys.stderr)
+    else:
+        prdebug = log.debug
+
+    # Guard against soft linking to oneself
+    if input_file == soft_link_name:
+        prdebug("Warning: No symbolic link made. You are using " +
+                "the original data directory as the working directory.")
+        return
+
+    # Soft link already exists: delete for relink?
+    if os.path.lexists(soft_link_name):
+        # do not delete or overwrite real (non-soft link) file
+        if not os.path.islink(soft_link_name):
+            raise FileExistsError(
+                "%s exists and is not a link" % soft_link_name)
+        try:
+            os.unlink(soft_link_name)
+        except OSError:
+            prdebug("Can't unlink %s" % (soft_link_name))
+
+    if not os.path.exists(input_file):
+        raise FileNotFoundError(
+            "trying to create a broken symlink to %s" % input_file)
+
+    prdebug("os.symlink(%s, %s)" % (input_file, soft_link_name))
+
+    # Create symbolic link using absolute path
+    os.symlink(
+        os.path.abspath(input_file),
+        soft_link_name
+    )
+
+
+def is_iterable_notstr(thing):
+    return isinstance(thing, Iterable) and not isinstance(thing, str)
+
+
+def page_number(input_file):
+    "Get one-based page number implied by filename (000002.pdf -> 2)"
+    return int(os.path.basename(fspath(input_file))[0:6])
+
+
+def available_cpu_count():
+    try:
+        return multiprocessing.cpu_count()
+    except NotImplementedError:
+        pass
+
+    try:
+        import psutil
+        return psutil.cpu_count()
+    except (ImportError, AttributeError):
+        pass
+
+    warnings.warn(
+        "Could not get CPU count.  Assuming one (1) CPU."
+        "Use -j N to set manually.")
+    return 1
+
+
+def is_file_writable(test_file):
+    """Intentionally racy test if target is writable.
+
+    We intend to write to the output file if and only if we succeed and
+    can replace it atomically. Before doing the OCR work, make sure
+    the location is writable.
+    """
+    p = Path(test_file)
+
+    if p.is_symlink():
+        # Python 3.5 does not accept parameters for Path.resolve() and behaves
+        # as if strict=True (throws an exception on failure). Python 3.6
+        # defaults to strict=False. This implements strict=False like behavior
+        # for Python 3.5.
+        if sys.version_info[0:2] <= (3, 5):
+            p = Path(os.path.realpath(str(p)))
+        else:
+            p = p.resolve(strict=False)
+
+    # p.is_file() throws an exception in some cases
+    if p.exists() and p.is_file():
+        return os.access(
+            str(p), os.W_OK,
+            effective_ids=(os.access in os.supports_effective_ids))
+    else:
+        try:
+            fp = p.open('wb')
+        except OSError:
+            return False
+        else:
+            fp.close()
+            with suppress(OSError):
+                p.unlink()
+        return True
+
+
+if sys.version_info[0:2] <= (3, 5):
+    def universal_open(p, *args, **kwargs):
+        "Work around Python 3.5's inability to open(pathlib.Path())"
+        try:
+            return p.open(*args, **kwargs)
+        except AttributeError:
+            return open(p, *args, **kwargs)
+
+
+    def fspath(path):
+        import pathlib
+        '''https://www.python.org/dev/peps/pep-0519/#os'''
+        if isinstance(path, (str, bytes)):
+            return path
+
+        # Work from the object's type to match method resolution of other magic
+        # methods.
+        path_type = type(path)
+        try:
+            path = path_type.__fspath__(path)
+        except AttributeError:
+            # Added for Python 3.5 support.
+            if isinstance(path, pathlib.Path):
+                return str(path)
+            elif hasattr(path_type, '__fspath__'):
+                raise
+        else:
+            if isinstance(path, (str, bytes)):
+                return path
+            else:
+                raise TypeError("expected __fspath__() to return str or bytes, "
+                                "not " + type(path).__name__)
+
+        raise TypeError(
+            "expected str, bytes, pathlib.Path or os.PathLike object, not "
+            + path_type.__name__)
+
+else:
+    universal_open = open
+    fspath = os.fspath

src/ocrmypdf/hocrtransform.py

@@ -0,0 +1,350 @@
+#!/usr/bin/env python3
+#
+# Copyright (c) 2010, Jonathan Brinley
+#   Original version from: https://github.com/jbrinley/HocrConverter
+#
+# Copyright (c) 2013-14, Julien Pfefferkorn
+#   Modifications
+#
+# Copyright (c) 2015-16, James R. Barlow
+#   Set text to transparent
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+# 
+# The above copyright notice and this permission notice shall be included
+# in all copies or substantial portions of the Software.
+# 
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+# OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+# CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+from reportlab.pdfgen.canvas import Canvas
+from reportlab.lib.units import inch
+from xml.etree import ElementTree
+from PIL import Image
+from collections import namedtuple
+from math import atan, sin, cos
+import re
+import argparse
+
+
+Rect = namedtuple('Rect', ['x1', 'y1', 'x2', 'y2'])
+
+
+class HocrTransformError(Exception):
+    pass
+
+
+class HocrTransform():
+
+    """
+    A class for converting documents from the hOCR format.
+    For details of the hOCR format, see:
+    http://kba.cloud/hocr-spec/
+    """
+
+    box_pattern = re.compile(r'bbox((\s+\d+){4})')
+    baseline_pattern = re.compile(r'''
+        baseline \s+ 
+        ([\-\+]?\d*\.?\d*) \s+  # +/- decimal float
+        ([\-\+]?\d+)            # +/- int''', re.VERBOSE)
+    ligatures = str.maketrans({
+        'ff': 'ff',
+        'ffi': 'f‌f‌i',
+        'ffl': 'f‌f‌l',
+        'fi': 'fi',
+        'fl': 'fl',
+    })
+
+    def __init__(self, hocrFileName, dpi):
+        self.dpi = dpi
+        self.hocr = ElementTree.parse(hocrFileName)
+
+        # if the hOCR file has a namespace, ElementTree requires its use to
+        # find elements
+        matches = re.match(r'({.*})html', self.hocr.getroot().tag)
+        self.xmlns = ''
+        if matches:
+            self.xmlns = matches.group(1)
+
+        # get dimension in pt (not pixel!!!!) of the OCRed image
+        self.width, self.height = None, None
+        for div in self.hocr.findall(
+                ".//%sdiv[@class='ocr_page']" % (self.xmlns)):
+            coords = self.element_coordinates(div)
+            pt_coords = self.pt_from_pixel(coords)
+            self.width = pt_coords.x2 - pt_coords.x1
+            self.height = pt_coords.y2 - pt_coords.y1
+            # there shouldn't be more than one, and if there is, we don't want
+            # it
+            break
+        if self.width is None or self.height is None:
+            raise HocrTransformError("hocr file is missing page dimensions")
+
+    def __str__(self):
+        """
+        Return the textual content of the HTML body
+        """
+        if self.hocr is None:
+            return ''
+        body = self.hocr.find(".//%sbody" % (self.xmlns))
+        if body:
+            return self._get_element_text(body)
+        else:
+            return ''
+
+    def _get_element_text(self, element):
+        """
+        Return the textual content of the element and its children
+        """
+        text = ''
+        if element.text is not None:
+            text += element.text
+        for child in element.getchildren():
+            text += self._get_element_text(child)
+        if element.tail is not None:
+            text += element.tail
+        return text
+
+    @classmethod
+    def element_coordinates(cls, element):
+        """
+        Returns a tuple containing the coordinates of the bounding box around
+        an element
+        """
+        out = (0, 0, 0, 0)
+        if 'title' in element.attrib:
+            matches = cls.box_pattern.search(element.attrib['title'])
+            if matches:
+                coords = matches.group(1).split()
+                out = Rect._make(int(coords[n]) for n in range(4))
+        return out
+
+    @classmethod
+    def baseline(cls, element):
+        """
+        Returns a tuple containing the baseline slope and intercept.
+        """
+        if 'title' in element.attrib:
+            matches = cls.baseline_pattern.search(element.attrib['title'])
+            if matches:
+                return float(matches.group(1)), int(matches.group(2))
+        return (0, 0) 
+
+    def pt_from_pixel(self, pxl):
+        """
+        Returns the quantity in PDF units (pt) given quantity in pixels
+        """
+        return Rect._make(
+            (c / self.dpi * inch) for c in pxl)
+
+    @classmethod
+    def replace_unsupported_chars(cls, s):
+        """
+        Given an input string, returns the corresponding string that:
+        - is available in the helvetica facetype
+        - does not contain any ligature (to allow easy search in the PDF file)
+        """
+        return s.translate(cls.ligatures)
+
+    def to_pdf(self, outFileName, imageFileName=None, showBoundingboxes=False,
+               fontname="Helvetica", invisibleText=False, interwordSpaces=False):
+        """
+        Creates a PDF file with an image superimposed on top of the text.
+        Text is positioned according to the bounding box of the lines in
+        the hOCR file.
+        The image need not be identical to the image used to create the hOCR
+        file.
+        It can have a lower resolution, different color mode, etc.
+        """
+        # create the PDF file
+        # page size in points (1/72 in.)
+        pdf = Canvas(
+            outFileName, pagesize=(self.width, self.height), pageCompression=1)
+
+        # draw bounding box for each paragraph
+        # light blue for bounding box of paragraph
+        pdf.setStrokeColorRGB(0, 1, 1)
+        # light blue for bounding box of paragraph
+        pdf.setFillColorRGB(0, 1, 1)
+        pdf.setLineWidth(0)		# no line for bounding box
+        for elem in self.hocr.findall(
+                ".//%sp[@class='%s']" % (self.xmlns, "ocr_par")):
+
+            elemtxt = self._get_element_text(elem).rstrip()
+            if len(elemtxt) == 0:
+                continue
+
+            pxl_coords = self.element_coordinates(elem)
+            pt = self.pt_from_pixel(pxl_coords)
+
+            # draw the bbox border
+            if showBoundingboxes:
+                pdf.rect(
+                    pt.x1, self.height - pt.y2, pt.x2 - pt.x1, pt.y2 - pt.y1,
+                    fill=1)
+                    
+        found_lines = False
+        for line in self.hocr.findall(
+                ".//%sspan[@class='%s']" % (self.xmlns, "ocr_line")):
+            found_lines = True
+            self._do_line(pdf, line, "ocrx_word", fontname, invisibleText, 
+                          interwordSpaces, showBoundingboxes)
+
+        if not found_lines:
+            # Tesseract did not report any lines (just words)
+            root = self.hocr.find(".//%sdiv[@class='%s']" % (self.xmlns, "ocr_page"))
+            self._do_line(pdf, root, "ocrx_word", fontname, invisibleText,
+                          interwordSpaces, showBoundingboxes)
+        # put the image on the page, scaled to fill the page
+        if imageFileName is not None:
+            pdf.drawImage(imageFileName, 0, 0,
+                          width=self.width, height=self.height)
+
+        # finish up the page and save it
+        pdf.showPage()
+        pdf.save()
+
+
+    @classmethod
+    def polyval(cls, poly, x):
+        return x * poly[0] + poly[1]
+
+
+    def _do_line(self, pdf, line, elemclass, fontname, invisibleText, 
+                 interwordSpaces, showBoundingboxes):
+        pxl_line_coords = self.element_coordinates(line)
+        line_box = self.pt_from_pixel(pxl_line_coords)
+        line_height = line_box.y2 - line_box.y1
+
+        slope, pxl_intercept = self.baseline(line)
+        if abs(slope) < 0.005:
+            slope = 0.0
+        angle = atan(slope)
+        cos_a, sin_a = cos(angle), sin(angle)
+
+        text = pdf.beginText()
+        intercept = pxl_intercept / self.dpi * inch
+
+        # Don't allow the font to break out of the bounding box. Division by
+        # cos_a accounts for extra clearance between the glyph's vertical axis
+        # on a sloped baseline and the edge of the bounding box.
+        fontsize = (line_height - abs(intercept)) / cos_a
+        text.setFont(fontname, fontsize)
+        if invisibleText:
+            text.setTextRenderMode(3)  # Invisible (indicates OCR text)
+
+        # Intercept is normally negative, so this places it above the bottom
+        # of the line box
+        baseline_y2 = self.height - (line_box.y2 + intercept)
+
+        if showBoundingboxes:
+            # draw the baseline in magenta, dashed
+            pdf.setDash()
+            pdf.setStrokeColorRGB(0.95, 0.65, 0.95)
+            pdf.setLineWidth(0.5)
+            # negate slope because it is defined as a rise/run in pixel
+            # coordinates and page coordinates have the y axis flipped
+            pdf.line(line_box.x1, 
+                     baseline_y2,
+                     line_box.x2,
+                     self.polyval((-slope, baseline_y2), 
+                                  line_box.x2 - line_box.x1))
+            # light green for bounding box of word/line
+            pdf.setDash(6, 3)
+            pdf.setStrokeColorRGB(1, 0, 0)
+
+        text.setTextTransform(
+            cos_a, -sin_a, sin_a, cos_a,
+            line_box.x1, baseline_y2
+        )
+        pdf.setFillColorRGB(0, 0, 0)  # text in black
+
+        elements = line.findall(
+                ".//%sspan[@class='%s']" % (self.xmlns, elemclass))
+        for elem in elements:
+            elemtxt = self._get_element_text(elem).strip()
+            elemtxt = self.replace_unsupported_chars(elemtxt)
+            if elemtxt == '':
+                continue
+
+            pxl_coords = self.element_coordinates(elem)
+            box = self.pt_from_pixel(pxl_coords)
+            if interwordSpaces:
+                # if  `--interword-spaces` is true, append a space
+                # to the end of each text element to allow simpler PDF viewers 
+                # such as PDF.js to better recognize words in search and copy 
+                # and paste. Do not remove space from last word in line, even
+                # though it would look better, because it will interfere with
+                # naive text extraction. \n does not work either.
+                elemtxt += ' '
+                box = Rect._make((
+                    box.x1, 
+                    line_box.y1,
+                    box.x2 + pdf.stringWidth(' ', fontname, line_height), 
+                    line_box.y2))
+            box_width = box.x2 - box.x1
+            font_width = pdf.stringWidth(elemtxt, fontname, fontsize)
+
+            # draw the bbox border
+            if showBoundingboxes:
+                pdf.rect(
+                    box.x1,
+                    self.height - line_box.y2,
+                    box_width,
+                    line_height,
+                    fill=0)
+
+            # Adjust relative position of cursor
+            # This is equivalent to:
+            #   text.setTextOrigin(pt.x1, self.height - line_box.y2)
+            # but the former generates a full text reposition matrix (Tm) in the
+            # content stream while this issues a "offset" (Td) command.
+            # .moveCursor() is relative to start of the text line, where the
+            # "text line" means whatever reportlab defines it as. Do not use
+            # use .getCursor(), since moveCursor() rather unintuitively plans 
+            # its moves relative to .getStartOfLine().
+            # For skewed lines, in the text transform we set up a rotated
+            # coordinate system, so we don't have to account for the 
+            # incremental offset. Surprisingly most PDF viewers can handle this.
+            cursor = text.getStartOfLine()
+            dx = box.x1 - cursor[0]
+            dy = baseline_y2 - cursor[1]            
+            text.moveCursor(dx, dy)
+
+            # If reportlab tells us this word is 0 units wide, our best seems
+            # to be to suppress this text
+            if font_width > 0:
+                text.setHorizScale(100 * box_width / font_width)
+                text.textOut(elemtxt)
+        pdf.drawText(text)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Convert hocr file to PDF')
+    parser.add_argument('-b', '--boundingboxes', action="store_true",
+                        default=False, help='Show bounding boxes borders')
+    parser.add_argument('-r', '--resolution', type=int,
+                        default=300,
+                        help='Resolution of the image that was OCRed')
+    parser.add_argument('-i', '--image', default=None,
+                        help='Path to the image to be placed above the text')
+    parser.add_argument('--interword-spaces', action='store_true',
+                         default=False, help='Add spaces between words')
+    parser.add_argument('hocrfile', help='Path to the hocr file to be parsed')
+    parser.add_argument(
+        'outputfile', help='Path to the PDF file to be generated')
+    args = parser.parse_args()
+
+    hocr = HocrTransform(args.hocrfile, args.resolution)
+    hocr.to_pdf(args.outputfile, args.image, args.boundingboxes, interwordSpaces=args.interword_spaces)

src/ocrmypdf/leptonica.py

@@ -0,0 +1,583 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+#
+# © 2013-16: jbarlow83 from Github (https://github.com/jbarlow83)
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+#
+# Python FFI wrapper for Leptonica library
+
+import argparse
+import sys
+import os
+import logging
+from tempfile import TemporaryFile
+from ctypes.util import find_library
+from .lib._leptonica import ffi
+from functools import lru_cache
+from enum import Enum
+from .helpers import fspath
+
+lept = ffi.dlopen(find_library('lept'))
+
+logger = logging.getLogger(__name__)
+
+
+def stderr(*objs):
+    """Python 2/3 compatible print to stderr.
+    """
+    print("leptonica.py:", *objs, file=sys.stderr)
+
+
+class LeptonicaErrorTrap(object):
+    """Context manager to trap errors reported by Leptonica.
+
+    Leptonica's error return codes are unreliable to the point of being
+    almost useless.  It does, however, write errors to stderr provided that is
+    not disabled at its compile time.  Fortunately this is done using error
+    macros so it is very self-consistent.
+
+    This context manager redirects stderr to a temporary file which is then
+    read and parsed for error messages.  As a side benefit, debug messages
+    from Leptonica are also suppressed.
+
+    """
+    def __enter__(self):
+        from io import UnsupportedOperation
+        self.tmpfile = TemporaryFile()
+
+        # Save the old stderr, and redirect stderr to temporary file
+        sys.stderr.flush()
+        try:
+            self.copy_of_stderr = os.dup(sys.stderr.fileno())
+            os.dup2(self.tmpfile.fileno(), sys.stderr.fileno(),
+                    inheritable=False)
+        except UnsupportedOperation:
+            self.copy_of_stderr = None
+        return
+
+    def __exit__(self, exc_type, exc_value, traceback):
+        # Restore old stderr
+        sys.stderr.flush()
+        if self.copy_of_stderr is not None:
+            os.dup2(self.copy_of_stderr, sys.stderr.fileno())
+            os.close(self.copy_of_stderr)
+
+        # Get data from tmpfile (in with block to ensure it is closed)
+        with self.tmpfile as tmpfile:
+            tmpfile.seek(0)  # Cursor will be at end, so move back to beginning
+            leptonica_output = tmpfile.read().decode(errors='replace')
+
+        assert self.tmpfile.closed
+        assert not sys.stderr.closed
+
+        # If there are Python errors, let them bubble up
+        if exc_type:
+            logger.warning(leptonica_output)
+            return False
+
+        # If there are Leptonica errors, wrap them in Python excpetions
+        if 'Error' in leptonica_output:
+            if 'image file not found' in leptonica_output:
+                raise FileNotFoundError()
+            if 'pixWrite: stream not opened' in leptonica_output:
+                raise LeptonicaIOError()
+            raise LeptonicaError(leptonica_output)
+
+        return False
+
+
+class LeptonicaError(Exception):
+    pass
+
+
+class LeptonicaIOError(LeptonicaError):
+    pass
+
+
+class Pix:
+    """Wrapper around leptonica's PIX object.
+
+    Leptonica uses referencing counting on PIX objects. Also, many Leptonica
+    functions return the original object with an increased reference count
+    if the operation had no effect (for example, image skew was found to be 0).
+    This has complications for memory management in Python. Whenever Leptonica
+    returns a PIX object (new or old), we wrap it in this class, which
+    registers it with the FFI garbage collector. pixDestroy() decrements the
+    reference count and only destroys when the last reference is removed.
+
+    Leptonica's reference counting is not threadsafe. This class can be used
+    in a threadsafe manner if a Python threading.Lock protects the data.
+
+    This class treats Pix objects as immutable.  All methods return new
+    modified objects.  This allows convenient chaining:
+
+    >>>   Pix.read('filename.jpg').scale((0.5, 0.5)).deskew().show()
+
+    """
+
+    def __init__(self, pix):
+        self._pix = ffi.gc(pix, Pix._pix_destroy)
+
+    def __repr__(self):
+        if self._pix:
+            s = "<leptonica.Pix image size={0}x{1} depth={2}{4} at 0x{3:x}>"
+            return s.format(self._pix.w, self._pix.h, self._pix.d,
+                            int(ffi.cast('intptr_t', self._pix)),
+                            '(colormapped)' if self._pix.colormap else '')
+        else:
+            return "<leptonica.Pix image NULL>"
+
+    def _repr_png_(self):
+        """iPython display hook
+
+        returns png version of image
+        """
+
+        data = ffi.new('l_uint8 **')
+        size = ffi.new('size_t *')
+
+        err = lept.pixWriteMemPng(data, size, self._pix, 0)
+        if err != 0:
+            raise LeptonicaIOError("pixWriteMemPng")
+
+        char_data = ffi.cast('char *', data[0])
+        return ffi.buffer(char_data, size[0])[:]
+
+    def __getstate__(self):
+        data = ffi.new('l_uint32 **')
+        size = ffi.new('size_t *')
+
+        err = lept.pixSerializeToMemory(self._pix, data, size)
+        if err != 0:
+            raise LeptonicaIOError("pixSerializeToMemory")
+
+        char_data = ffi.cast('char *', data[0])
+
+        # Copy from C bytes to python bytes()
+        data_bytes = ffi.buffer(char_data, size[0])[:]
+
+        # Can now free C bytes
+        lept.lept_free(char_data)
+        return dict(data=data_bytes)
+
+    def __setstate__(self, state):
+        cdata_bytes = ffi.new('char[]', state['data'])
+        cdata_uint32 = ffi.cast('l_uint32 *', cdata_bytes)
+
+        pix = lept.pixDeserializeFromMemory(
+            cdata_uint32, len(state['data']))
+        Pix.__init__(self, pix)
+
+    def __eq__(self, other):
+        return self.__getstate__() == other.__getstate__()
+
+    @property
+    def width(self):
+        return self._pix.w
+
+    @property
+    def height(self):
+        return self._pix.h
+
+    @property
+    def depth(self):
+        return self._pix.d
+
+    @property
+    def size(self):
+        return (self._pix.w, self._pix.h)
+
+    @property
+    def info(self):
+        return {'dpi': (self._pix.xres, self._pix.yres)}
+
+    @property
+    def mode(self):
+        "Return mode like PIL.Image"
+        if self.depth == 1:
+            return '1'
+        elif self.depth >= 16:
+            return 'RGB'
+        elif not self._pix.colormap:
+            return 'L'
+        else:
+            return 'P'
+
+    @classmethod
+    def read(cls, path):
+        """Load an image file into a PIX object.
+
+        Leptonica can load TIFF, PNM (PBM, PGM, PPM), PNG, and JPEG.  If
+        loading fails then the object will wrap a C null pointer.
+        """
+        filename = fspath(path)
+        with LeptonicaErrorTrap():
+            return cls(lept.pixRead(os.fsencode(filename)))
+
+    def write_implied_format(
+            self, path, jpeg_quality=0, jpeg_progressive=0):
+        """Write pix to the filename, with the extension indicating format.
+
+        jpeg_quality -- quality (iff JPEG; 1 - 100, 0 for default)
+        jpeg_progressive -- (iff JPEG; 0 for baseline seq., 1 for progressive)
+        """
+        filename = fspath(path)
+        with LeptonicaErrorTrap():
+            lept.pixWriteImpliedFormat(
+                os.fsencode(filename),
+                self._pix, jpeg_quality, jpeg_progressive)
+
+    def topil(self):
+        "Returns a PIL.Image version of this Pix"
+        from PIL import Image
+
+        # Leptonica manages data in words, so it implicitly does an endian
+        # swap.  Tell Pillow about this when it reads the data.
+        pix = self
+        if sys.byteorder == 'little':
+            if self.mode == 'RGB':
+                raw_mode = 'XBGR'
+            elif self.mode == 'RGBA':
+                raw_mode = 'ABGR'
+            elif self.mode == '1':
+                raw_mode = '1;I'
+                pix = Pix(lept.pixEndianByteSwapNew(pix._pix))
+            else:
+                raw_mode = self.mode
+                pix = Pix(lept.pixEndianByteSwapNew(pix._pix))
+        else:
+            raw_mode = self.mode  # no endian swap needed
+
+        size = (pix._pix.w, pix._pix.h)
+        bytecount = pix._pix.wpl * 4 * pix._pix.h
+        buf = ffi.buffer(pix._pix.data, bytecount)
+        stride = pix._pix.wpl * 4
+
+        im = Image.frombytes(self.mode, size, buf, 'raw', raw_mode, stride)
+
+        return im
+
+    def show(self):
+        return self.topil().show()
+
+    def deskew(self, reduction_factor=0):
+        """Returns the deskewed pix object.
+
+        A clone of the original is returned when the algorithm cannot find a
+        skew angle with sufficient confidence.
+
+        reduction_factor -- amount to downsample (0 for default) when searching
+            for skew angle
+        """
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixDeskew(self._pix, reduction_factor))
+
+    def scale(self, scale_xy):
+        "Returns the pix object rescaled according to the proportions given."
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixScale(self._pix, scale_xy[0], scale_xy[1]))
+
+    def rotate180(self):
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixRotate180(ffi.NULL, self._pix))
+
+    def rotate_orth(self, quads):
+        "Orthographic rotation, quads: 0-3, number of clockwise rotations"
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixRotateOrth(self._pix, quads))
+
+    def find_skew(self):
+        """Returns a tuple (deskew angle in degrees, confidence value).
+
+        Returns (None, None) if no angle is available.
+        """
+        with LeptonicaErrorTrap():
+            angle = ffi.new('float *', 0.0)
+            confidence = ffi.new('float *', 0.0)
+            result = lept.pixFindSkew(self._pix, angle, confidence)
+            if result == 0:
+                return (angle[0], confidence[0])
+            else:
+                return (None, None)
+
+    def convert_rgb_to_luminance(self):
+        with LeptonicaErrorTrap():
+            gray_pix = lept.pixConvertRGBToLuminance(self._pix)
+            if gray_pix:
+                return Pix(gray_pix)
+            return None
+
+    def remove_colormap(self, removal_type):
+        """Remove a palette (colormap); if no colormap, returns a copy of this 
+        image
+
+            removal_type - any of lept.REMOVE_CMAP_*
+
+        """
+        return self
+        # with LeptonicaErrorTrap():
+        #     return Pix(lept.pixRemoveColormapGeneral(
+        #             self._pix, removal_type, lept.L_COPY))
+
+    def otsu_adaptive_threshold(
+            self, tile_size=(300, 300), kernel_size=(4, 4), scorefract=0.1):
+        with LeptonicaErrorTrap():
+            sx, sy = tile_size
+            smoothx, smoothy = kernel_size
+            p_pix = ffi.new('PIX **')
+
+            result = lept.pixOtsuAdaptiveThreshold(
+                self._pix,
+                sx, sy,
+                smoothx, smoothy,
+                scorefract,
+                ffi.NULL,
+                p_pix)
+            if result == 0:
+                return Pix(p_pix[0])
+            else:
+                return None
+
+    def otsu_threshold_on_background_norm(
+            self, mask=None, tile_size=(10, 15), thresh=100, mincount=50,
+            bgval=255, kernel_size=(2, 2), scorefract=0.1):
+        with LeptonicaErrorTrap():
+            sx, sy = tile_size
+            smoothx, smoothy = kernel_size
+            if mask is None:
+                mask = ffi.NULL
+            if isinstance(mask, Pix):
+                mask = mask._pix
+
+            thresh_pix = lept.pixOtsuThreshOnBackgroundNorm(
+                self._pix,
+                mask,
+                sx, sy,
+                thresh, mincount, bgval,
+                smoothx, smoothy,
+                scorefract,
+                ffi.NULL
+                )
+            if thresh_pix == ffi.NULL:
+                return None
+            return Pix(thresh_pix)
+
+    def crop_to_foreground(
+            self, threshold=128, mindist=70, erasedist=30, pagenum=0,
+            showmorph=0, display=0, pdfdir=ffi.NULL):
+        with LeptonicaErrorTrap():
+            cropbox = Box(lept.pixFindPageForeground(
+                self._pix,
+                threshold,
+                mindist,
+                erasedist,
+                pagenum,
+                showmorph,
+                display,
+                pdfdir))
+
+            print(repr(cropbox))
+
+            cropped_pix = lept.pixClipRectangle(
+                self._pix,
+                cropbox._box,
+                ffi.NULL)
+
+            return Pix(cropped_pix)
+
+    def clean_background_to_white(
+            self, mask=None, grayscale=None, gamma=1.0, black=0, white=255):
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixCleanBackgroundToWhite(
+                self._pix,
+                mask or ffi.NULL,
+                grayscale or ffi.NULL,
+                gamma,
+                black,
+                white))
+
+    def gamma_trc(self, gamma=1.0, minval=0, maxval=255):
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixGammaTRC(
+                ffi.NULL,
+                self._pix,
+                gamma,
+                minval,
+                maxval
+                ))
+
+    def background_norm(
+            self, mask=None, grayscale=None, tile_size=(10, 15), fg_threshold=60,
+            min_count=40, bg_val=200, smooth_kernel=(2, 1)):
+        # Background norm doesn't work on color mapped Pix, so remove colormap
+        target_pix = self.remove_colormap(lept.REMOVE_CMAP_BASED_ON_SRC)
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixBackgroundNorm(
+                target_pix._pix,
+                mask or ffi.NULL,
+                grayscale or ffi.NULL,
+                tile_size[0],
+                tile_size[1],
+                fg_threshold,
+                min_count,
+                bg_val,
+                smooth_kernel[0],
+                smooth_kernel[1]
+                ))
+
+    @staticmethod
+    @lru_cache(maxsize=1)
+    def make_pixel_sum_tab8():
+        return lept.makePixelSumTab8()
+
+    @staticmethod
+    def correlation_binary(pix1, pix2):
+        if get_leptonica_version() < 'leptonica-1.72':
+            # Older versions of Leptonica (pre-1.72) have a buggy
+            # implementation of pixCorrelationBinary that overflows on larger
+            # images.  Ubuntu 14.04/trusty has 1.70. Ubuntu PPA
+            # ppa:alex-p/tesseract-ocr has leptonlib 1.75.
+            pix1_count = ffi.new('l_int32 *')
+            pix2_count = ffi.new('l_int32 *')
+            pixn_count = ffi.new('l_int32 *')
+            tab8 = Pix.make_pixel_sum_tab8()
+
+            lept.pixCountPixels(pix1._pix, pix1_count, tab8)
+            lept.pixCountPixels(pix2._pix, pix2_count, tab8)
+            pixn = Pix(lept.pixAnd(ffi.NULL, pix1._pix, pix2._pix))
+            lept.pixCountPixels(pixn._pix, pixn_count, tab8)
+
+            # Python converts these int32s to larger units as needed
+            # to avoid overflow. Overflow happens easily here.
+            correlation = (
+                    (pixn_count[0] * pixn_count[0]) /
+                    (pix1_count[0] * pix2_count[0])
+                    )
+            return correlation
+        else:
+            correlation = ffi.new('float *', 0.0)
+            result = lept.pixCorrelationBinary(pix1._pix, pix2._pix,
+                                               correlation)
+            if result != 0:
+                raise LeptonicaError("Correlation failed")
+            return correlation[0]
+
+    @staticmethod
+    def _pix_destroy(pix):
+        p_pix = ffi.new('PIX **', pix)
+        lept.pixDestroy(p_pix)
+        # print('pix destroy ' + repr(pix))
+
+
+class Box:
+    """Wrapper around Leptonica's BOX objects.
+
+    See class Pix for notes about reference counting.
+    """
+
+    def __init__(self, box):
+        self._box = ffi.gc(box, Box._box_destroy)
+
+    def __repr__(self):
+        if self._box:
+            return '<leptonica.Box x={0} y={1} w={2} h={3}>'.format(
+                self.x, self.y, self.w, self.h)
+        return '<leptonica.Box NULL>'
+
+    @property
+    def x(self):
+        return self._box.x
+
+    @property
+    def y(self):
+        return self._box.y
+
+    @property
+    def w(self):
+        return self._box.w
+
+    @property
+    def h(self):
+        return self._box.h
+
+    @staticmethod
+    def _box_destroy(box):
+        p_box = ffi.new('BOX **', box)
+        lept.boxDestroy(p_box)
+
+
+@lru_cache(maxsize=1)
+def get_leptonica_version():
+    """Get Leptonica version string.
+
+    Caveat: Leptonica expects the caller to free this memory.  We don't,
+    since that would involve binding to libc to access libc.free(),
+    a pointless effort to reclaim 100 bytes of memory.
+    """
+    return ffi.string(lept.getLeptonicaVersion()).decode()
+
+
+def deskew(infile, outfile, dpi):
+    try:
+        pix_source = Pix.read(infile)
+    except LeptonicaIOError:
+        raise LeptonicaIOError("Failed to open file: %s" % infile)
+
+    if dpi < 150:
+        reduction_factor = 1  # Don't downsample too much if DPI is already low
+    else:
+        reduction_factor = 0  # Use default
+    pix_deskewed = pix_source.deskew(reduction_factor)
+
+    try:
+        pix_deskewed.write_implied_format(outfile)
+    except LeptonicaIOError:
+        raise LeptonicaIOError("Failed to open destination file: %s" % outfile)
+
+
+def remove_background(infile, outfile, tile_size=(40, 60), gamma=1.0,
+                      black_threshold=70, white_threshold=190):
+    try:
+        pix = Pix.read(infile)
+    except LeptonicaIOError:
+        raise LeptonicaIOError("Failed to open file: %s" % infile)
+
+    pix = pix.background_norm(tile_size=tile_size).gamma_trc(
+            gamma, black_threshold, white_threshold)
+
+    try:
+        pix.write_implied_format(outfile)
+    except LeptonicaIOError:
+        raise LeptonicaIOError("Failed to open destination file: %s" % outfile)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(
+        description="Python wrapper to access Leptonica")
+
+    subparsers = parser.add_subparsers(title='commands',
+                                       description='supported operations')
+
+    parser_deskew = subparsers.add_parser('deskew')
+    parser_deskew.add_argument('-r', '--dpi', dest='dpi', action='store',
+                               type=int, default=300, help='input resolution')
+    parser_deskew.add_argument('infile', help='image to deskew')
+    parser_deskew.add_argument('outfile', help='deskewed output image')
+    parser_deskew.set_defaults(func=deskew)
+
+    args = parser.parse_args()
+    args.func(args)

src/ocrmypdf/lib/__init__.py

@@ -0,0 +1,28 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+"""Bindings to external libraries"""
+
+import os as _os
+
+try:
+    import fitz
+except ImportError:
+    fitz = None
+
+if _os.environ.get('_OCRMYPDF_NO_FITZ'):
+    fitz = None

src/ocrmypdf/lib/_leptonica.py

@@ -0,0 +1,11 @@
+# auto-generated file
+import _cffi_backend
+
+ffi = _cffi_backend.FFI('ocrmypdf.lib._leptonica',
+    _version = 0x2601,
+    _types = b'\x00\x00\x0F\x0D\x00\x00\xBB\x03\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\xBE\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\xBA\x03\x00\x00\x0F\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x02\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x08\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\xC9\x03\x00\x00\x1C\x01\x00\x00\x00\x0F\x00\x00\xBD\x0D\x00\x00\x00\x0F\x00\x00\x49\x0D\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x1F\x03\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x49\x11\x00\x00\x49\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x49\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x01\x03\x00\x00\x9D\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\xC8\x03\x00\x00\x65\x03\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x08\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\xC6\x03\x00\x00\xA3\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x00\xCC\x0D\x00\x00\x10\x11\x00\x00\x00\x0F\x00\x00\xCC\x0D\x00\x00\x9D\x11\x00\x00\x00\x0F\x00\x00\xCC\x0D\x00\x00\xCC\x03\x00\x00\x00\x0F\x00\x00\x00\x09\x00\x00\x01\x09\x00\x00\x02\x09\x00\x00\xBE\x03\x00\x00\x02\x01\x00\x00\x0E\x01\x00\x00\x00\x0B\x00\x00\x01\x0B\x00\x00\x0B\x01\x00\x00\x05\x01\x00\x00\x03\x01\x00\x00\xBC\x03\x00\x00\xC7\x03\x00\x00\x04\x01\x00\x00\xC9\x03\x00\x00\x08\x01\x00\x00\x0C\x01\x00\x00\x06\x01\x00\x00\x00\x01',
+    _globals = (b'\xFF\xFF\xFF\x0BL_CLONE',2,b'\xFF\xFF\xFF\x0BL_COPY',1,b'\xFF\xFF\xFF\x0BL_COPY_CLONE',3,b'\xFF\xFF\xFF\x0BL_INSERT',0,b'\xFF\xFF\xFF\x0BL_NOCOPY',0,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_BASED_ON_SRC',4,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_BINARY',0,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_FULL_COLOR',2,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_GRAYSCALE',1,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_WITH_ALPHA',3,b'\x00\x00\xB1\x23boxDestroy',0,b'\x00\x00\x67\x23getLeptonicaVersion',0,b'\x00\x00\xB7\x23lept_free',0,b'\x00\x00\x69\x23makePixelSumTab8',0,b'\x00\x00\x16\x23pixAnd',0,b'\x00\x00\x23\x23pixBackgroundNorm',0,b'\x00\x00\x1B\x23pixCleanBackgroundToWhite',0,b'\x00\x00\x0D\x23pixClipRectangle',0,b'\x00\x00\x8D\x23pixColorFraction',0,b'\x00\x00\x59\x23pixColorMagnitude',0,b'\x00\x00\x0A\x23pixConvertRGBToLuminance',0,b'\x00\x00\x6B\x23pixCorrelationBinary',0,b'\x00\x00\x80\x23pixCountPixels',0,b'\x00\x00\x63\x23pixDeserializeFromMemory',0,b'\x00\x00\x50\x23pixDeskew',0,b'\x00\x00\xB4\x23pixDestroy',0,b'\x00\x00\x0A\x23pixEndianByteSwapNew',0,b'\x00\x00\x00\x23pixFindPageForeground',0,b'\x00\x00\x7B\x23pixFindSkew',0,b'\x00\x00\x2F\x23pixGammaTRC',0,b'\x00\x00\x70\x23pixGetAverageMaskedRGB',0,b'\x00\x00\x36\x23pixGlobalNormRGB',0,b'\x00\x00\x12\x23pixInvert',0,b'\x00\x00\x54\x23pixMaskOverColorPixels',0,b'\x00\x00\x85\x23pixNumSignificantGrayColors',0,b'\x00\x00\x96\x23pixOtsuAdaptiveThreshold',0,b'\x00\x00\x3E\x23pixOtsuThreshOnBackgroundNorm',0,b'\x00\x00\x60\x23pixRead',0,b'\x00\x00\x50\x23pixRemoveColormap',0,b'\x00\x00\x54\x23pixRemoveColormapGeneral',0,b'\x00\x00\x12\x23pixRotate180',0,b'\x00\x00\x50\x23pixRotateOrth',0,b'\x00\x00\x4B\x23pixScale',0,b'\x00\x00\xA0\x23pixSerializeToMemory',0,b'\x00\x00\xA5\x23pixWriteImpliedFormat',0,b'\x00\x00\xAB\x23pixWriteMemPng',0),
+    _struct_unions = ((b'\x00\x00\x00\xBA\x00\x00\x00\x02Box',b'\x00\x00\x02\x11x',b'\x00\x00\x02\x11y',b'\x00\x00\x02\x11w',b'\x00\x00\x02\x11h',b'\x00\x00\xC9\x11refcount'),(b'\x00\x00\x00\xBB\x00\x00\x00\x02Pix',b'\x00\x00\xC9\x11w',b'\x00\x00\xC9\x11h',b'\x00\x00\xC9\x11d',b'\x00\x00\xC9\x11spp',b'\x00\x00\xC9\x11wpl',b'\x00\x00\xC9\x11refcount',b'\x00\x00\x02\x11xres',b'\x00\x00\x02\x11yres',b'\x00\x00\x02\x11informat',b'\x00\x00\x02\x11special',b'\x00\x00\xBD\x11text',b'\x00\x00\xC5\x11colormap',b'\x00\x00\xC8\x11data'),(b'\x00\x00\x00\xBC\x00\x00\x00\x02PixColormap',b'\x00\x00\xB8\x11array',b'\x00\x00\x02\x11depth',b'\x00\x00\x02\x11nalloc',b'\x00\x00\x02\x11n')),
+    _enums = (b'\x00\x00\x00\xC0\x00\x00\x00\x16$1\x00REMOVE_CMAP_TO_BINARY,REMOVE_CMAP_TO_GRAYSCALE,REMOVE_CMAP_TO_FULL_COLOR,REMOVE_CMAP_WITH_ALPHA,REMOVE_CMAP_BASED_ON_SRC',b'\x00\x00\x00\xC1\x00\x00\x00\x16$2\x00L_NOCOPY,L_INSERT,L_COPY,L_CLONE,L_COPY_CLONE'),
+    _typenames = (b'\x00\x00\x00\xBABOX',b'\x00\x00\x00\xBBPIX',b'\x00\x00\x00\xBCPIXCMAP',b'\x00\x00\x00\x1Fl_float32',b'\x00\x00\x00\xBFl_float64',b'\x00\x00\x00\xC3l_int16',b'\x00\x00\x00\x02l_int32',b'\x00\x00\x00\xC2l_int64',b'\x00\x00\x00\xC4l_int8',b'\x00\x00\x00\xCBl_uint16',b'\x00\x00\x00\xC9l_uint32',b'\x00\x00\x00\xCAl_uint64',b'\x00\x00\x00\xC7l_uint8'),
+)

src/ocrmypdf/lib/compile_leptonica.py

@@ -0,0 +1,262 @@
+#!/usr/bin/env python3
+# © 2017 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from cffi import FFI
+
+ffibuilder = FFI()
+ffibuilder.set_source("ocrmypdf.lib._leptonica", None)
+ffibuilder.cdef("""
+typedef signed char             l_int8;
+typedef unsigned char           l_uint8;
+typedef short                   l_int16;
+typedef unsigned short          l_uint16;
+typedef int                     l_int32;
+typedef unsigned int            l_uint32;
+typedef float                   l_float32;
+typedef double                  l_float64;
+typedef long long               l_int64;
+typedef unsigned long long      l_uint64;
+
+struct Pix
+{
+    l_uint32             w;           /* width in pixels                   */
+    l_uint32             h;           /* height in pixels                  */
+    l_uint32             d;           /* depth in bits (bpp)               */
+    l_uint32             spp;         /* number of samples per pixel       */
+    l_uint32             wpl;         /* 32-bit words/line                 */
+    l_uint32             refcount;    /* reference count (1 if no clones)  */
+    l_int32              xres;        /* image res (ppi) in x direction    */
+                                      /* (use 0 if unknown)                */
+    l_int32              yres;        /* image res (ppi) in y direction    */
+                                      /* (use 0 if unknown)                */
+    l_int32              informat;    /* input file format, IFF_*          */
+    l_int32              special;     /* special instructions for I/O, etc */
+    char                *text;        /* text string associated with pix   */
+    struct PixColormap  *colormap;    /* colormap (may be null)            */
+    l_uint32            *data;        /* the image data                    */
+};
+typedef struct Pix PIX;
+
+struct PixColormap
+{
+    void            *array;     /* colormap table (array of RGBA_QUAD)     */
+    l_int32          depth;     /* of pix (1, 2, 4 or 8 bpp)               */
+    l_int32          nalloc;    /* number of color entries allocated       */
+    l_int32          n;         /* number of color entries used            */
+};
+typedef struct PixColormap  PIXCMAP;
+
+struct Box
+{
+    l_int32            x;
+    l_int32            y;
+    l_int32            w;
+    l_int32            h;
+    l_uint32           refcount;      /* reference count (1 if no clones)  */
+
+};
+typedef struct Box    BOX;
+
+enum {
+    REMOVE_CMAP_TO_BINARY = 0,     /*!< remove colormap for conv to 1 bpp  */
+    REMOVE_CMAP_TO_GRAYSCALE = 1,  /*!< remove colormap for conv to 8 bpp  */
+    REMOVE_CMAP_TO_FULL_COLOR = 2, /*!< remove colormap for conv to 32 bpp */
+    REMOVE_CMAP_WITH_ALPHA = 3,    /*!< remove colormap and alpha          */
+    REMOVE_CMAP_BASED_ON_SRC = 4   /*!< remove depending on src format     */
+};
+
+/*! Access and storage flags */
+enum {
+    L_NOCOPY = 0,     /*!< do not copy the object; do not delete the ptr  */
+    L_INSERT = L_NOCOPY,    /*!< stuff it in; do not copy or clone        */
+    L_COPY = 1,       /*!< make/use a copy of the object                  */
+    L_CLONE = 2,      /*!< make/use clone (ref count) of the object       */
+    L_COPY_CLONE = 3  /*!< make a new array object (e.g., pixa) and fill  */
+                      /*!< the array with clones (e.g., pix)              */
+};
+
+""")
+
+ffibuilder.cdef("""
+PIX * pixRead ( const char *filename );
+PIX * pixScale ( PIX *pixs, l_float32 scalex, l_float32 scaley );
+l_int32 pixFindSkew ( PIX *pixs, l_float32 *pangle, l_float32 *pconf );
+l_int32 pixWriteImpliedFormat ( const char *filename, PIX *pix, l_int32 quality, l_int32 progressive );
+l_int32
+pixWriteMemPng(l_uint8  **pdata,
+               size_t    *psize,
+               PIX       *pix,
+               l_float32  gamma);
+
+void pixDestroy ( PIX **ppix );
+
+PIX *
+pixEndianByteSwapNew(PIX  *pixs);
+
+PIX * pixDeskew ( PIX *pixs, l_int32 redsearch );
+char * getLeptonicaVersion (  );
+l_int32 pixCorrelationBinary(PIX *pix1, PIX *pix2, l_float32 *pval);
+PIX *pixRotate180(PIX *pixd, PIX *pixs);
+PIX *
+pixRotateOrth(PIX     *pixs,
+              l_int32  quads);
+
+l_int32 pixCountPixels ( PIX *pix, l_int32 *pcount, l_int32 *tab8 );
+PIX * pixAnd ( PIX *pixd, PIX *pixs1, PIX *pixs2 );
+l_int32 * makePixelSumTab8 ( void );
+
+PIX * pixDeserializeFromMemory ( const l_uint32 *data, size_t nbytes );
+l_int32 pixSerializeToMemory ( PIX *pixs, l_uint32 **pdata, size_t *pnbytes );
+
+PIX * pixConvertRGBToLuminance(PIX *pixs);
+
+PIX * pixRemoveColormap(PIX *pixs, l_int32  type);
+
+l_int32
+pixOtsuAdaptiveThreshold(PIX       *pixs,
+                         l_int32    sx,
+                         l_int32    sy,
+                         l_int32    smoothx,
+                         l_int32    smoothy,
+                         l_float32  scorefract,
+                         PIX      **ppixth,
+                         PIX      **ppixd);
+
+PIX *
+pixOtsuThreshOnBackgroundNorm(PIX       *pixs,
+                              PIX       *pixim,
+                              l_int32    sx,
+                              l_int32    sy,
+                              l_int32    thresh,
+                              l_int32    mincount,
+                              l_int32    bgval,
+                              l_int32    smoothx,
+                              l_int32    smoothy,
+                              l_float32  scorefract,
+                              l_int32   *pthresh);
+
+PIX *
+pixCleanBackgroundToWhite(PIX       *pixs,
+                          PIX       *pixim,
+                          PIX       *pixg,
+                          l_float32  gamma,
+                          l_int32    blackval,
+                          l_int32    whiteval);
+
+BOX *
+pixFindPageForeground(PIX         *pixs,
+                      l_int32      threshold,
+                      l_int32      mindist,
+                      l_int32      erasedist,
+                      l_int32      pagenum,
+                      l_int32      showmorph,
+                      l_int32      display,
+                      const char  *pdfdir);
+
+PIX *
+pixClipRectangle(PIX   *pixs,
+                 BOX   *box,
+                 BOX  **pboxc);
+
+PIX *
+pixBackgroundNorm(PIX     *pixs,
+                  PIX     *pixim,
+                  PIX     *pixg,
+                  l_int32  sx,
+                  l_int32  sy,
+                  l_int32  thresh,
+                  l_int32  mincount,
+                  l_int32  bgval,
+                  l_int32  smoothx,
+                  l_int32  smoothy);
+
+PIX *
+pixGammaTRC(PIX       *pixd,
+            PIX       *pixs,
+            l_float32  gamma,
+            l_int32    minval,
+            l_int32    maxval);
+
+
+l_int32
+pixNumSignificantGrayColors(PIX       *pixs,
+                            l_int32    darkthresh,
+                            l_int32    lightthresh,
+                            l_float32  minfract,
+                            l_int32    factor,
+                            l_int32   *pncolors);
+
+l_int32
+pixColorFraction(PIX        *pixs,
+                 l_int32     darkthresh,
+                 l_int32     lightthresh,
+                 l_int32     diffthresh,
+                 l_int32     factor,
+                 l_float32  *ppixfract,
+                 l_float32  *pcolorfract);
+
+PIX *
+pixColorMagnitude(PIX     *pixs,
+                  l_int32  rwhite,
+                  l_int32  gwhite,
+                  l_int32  bwhite,
+                  l_int32  type);
+
+PIX *
+pixMaskOverColorPixels(PIX     *pixs,
+                       l_int32  threshdiff,
+                       l_int32  mindist);
+
+l_int32
+pixGetAverageMaskedRGB(PIX        *pixs,
+                       PIX        *pixm,
+                       l_int32     x,
+                       l_int32     y,
+                       l_int32     factor,
+                       l_int32     type,
+                       l_float32  *prval,
+                       l_float32  *pgval,
+                       l_float32  *pbval);
+
+PIX * 
+pixGlobalNormRGB(PIX * 	pixd,
+                 PIX * 	pixs,
+                 l_int32 	rval,
+                 l_int32 	gval,
+                 l_int32 	bval,
+                 l_int32 	mapval); 
+
+PIX *
+pixInvert(PIX * pixd,
+          PIX * pixs);
+
+PIX *
+pixRemoveColormapGeneral(PIX     *pixs,
+                         l_int32  type,
+                         l_int32  ifnocmap);
+
+void                 
+boxDestroy(BOX  **pbox);
+
+void
+lept_free(void *ptr);  
+""")
+
+
+if __name__ == '__main__':
+    ffibuilder.compile(verbose=True)

src/ocrmypdf/pdfa.py

@@ -0,0 +1,324 @@
+# © 2015 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+"""
+Generate a PDFMARK file for Ghostscript >= 9.14, for PDF/A conversion
+
+pdfmark is an extension to the Postscript language that describes some PDF
+features like bookmarks and annotations. It was originally specified Adobe
+Distiller, for Postscript to PDF conversion:
+https://www.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/pdfmark_reference.pdf
+
+Ghostscript uses pdfmark for PDF to PDF/A conversion as well. To use Ghostscript
+to create a PDF/A, we need to create a pdfmark file with the necessary metadata.
+
+This takes care of the many version-specific bugs and pecularities in
+Ghostscript's handling of pdfmark.
+
+"""
+
+from string import Template
+from binascii import hexlify
+from datetime import datetime
+from xml.parsers.expat import ExpatError
+import pkg_resources
+import PyPDF2 as pypdf
+from defusedxml.minidom import parseString as defused_parseString
+from unittest.mock import patch
+
+ICC_PROFILE_RELPATH = 'data/sRGB.icc'
+
+SRGB_ICC_PROFILE = pkg_resources.resource_filename(
+    'ocrmypdf', ICC_PROFILE_RELPATH)
+
+
+# This is a template written in PostScript which is needed to create PDF/A
+# files, from the Ghostscript documentation. Lines beginning with % are
+# comments. Python substitution variables have a '$' prefix.
+pdfa_def_template = u"""%!
+% This is derived from Ghostscript's template for creating a PDF/A document.
+% This is a small PostScript program that includes some necessary information
+% to create a PDF/A compliant file.
+
+% Define entries in the document Info dictionary :
+/ICCProfile ($icc_profile)
+def
+
+[$docinfo
+  /DOCINFO pdfmark
+
+% Define an ICC profile :
+
+[/_objdef {icc_PDFA} /type /stream /OBJ pdfmark
+[{icc_PDFA}
+<<
+  /N currentpagedevice /ProcessColorModel known {
+    currentpagedevice /ProcessColorModel get dup /DeviceGray eq
+    {pop 1} {
+      /DeviceRGB eq
+      {3}{4} ifelse
+    } ifelse
+  } {
+    (ERROR, unable to determine ProcessColorModel) == flush
+  } ifelse
+>> /PUT pdfmark
+[{icc_PDFA} ICCProfile (r) file /PUT pdfmark
+
+% Define the output intent dictionary :
+
+[/_objdef {OutputIntent_PDFA} /type /dict /OBJ pdfmark
+[{OutputIntent_PDFA} <<
+  /Type /OutputIntent             % Must be so (the standard requires).
+  /S /GTS_PDFA1                   % Must be so (the standard requires).
+  /DestOutputProfile {icc_PDFA}            % Must be so (see above).
+  /OutputConditionIdentifier ($icc_identifier)
+>> /PUT pdfmark
+[{Catalog} <</OutputIntents [ {OutputIntent_PDFA} ]>> /PUT pdfmark
+"""
+
+
+def encode_text_string(s: str) -> str:
+    """
+    Encode text string to hex string for use in a PDF
+
+    From PDF 32000-1:2008 a string object may be included in hexademical form
+    if it is enclosed in angle brackets.  For general Unicode the string should
+    be UTF-16 (big endian) with byte order marks.  Many strings including all
+    ASCII strings could be encoded as PdfDocEncoding literals provided
+    that certain Postscript sequences are escaped.  But it's far simpler to
+    encode everything as UTF-16.
+    """
+
+    # Sometimes lazy C programmers leave their NULs at the end of strings they
+    # insert into PDFs
+    # tests/resources/aspect.pdf is one example (created by ImageMagick)
+    s = s.replace('\x00', '')
+
+    if s == '':
+        return ''
+
+    utf16_bytes = s.encode('utf-16be')
+    ascii_hex_bytes = hexlify(b'\xfe\xff' + utf16_bytes)
+    ascii_hex_str = ascii_hex_bytes.decode('ascii').lower()
+    return ascii_hex_str
+
+
+def _encode_ascii(s: str) -> str:
+    """
+    Aggressively strip non-ASCII and PDF escape sequences
+
+    Ghostscript 9.24+ lost support for UTF-16BE in pdfmark files for reasons
+    given in GhostPDL commit e997c683. Our temporary workaround is use ASCII
+    and drop all non-ASCII characters. A slightly improved alternative would
+    be to implement PdfDocEncoding in pikepdf and encode to that, or handle
+    metadata there.
+    """
+    trans = str.maketrans({
+        '(': '',
+        ')': '',
+        '\\': '',
+        '\0': ''
+    })
+    return s.translate(trans).encode('ascii', errors='replace').decode()
+
+
+def encode_pdf_date(d: datetime) -> str:
+    """
+    Encode Python datetime object as PDF date string
+
+    From Adobe pdfmark manual:
+    (D:YYYYMMDDHHmmSSOHH'mm')
+    D: is an optional prefix. YYYY is the year. All fields after the year are
+    optional. MM is the month (01-12), DD is the day (01-31), HH is the
+    hour (00-23), mm are the minutes (00-59), and SS are the seconds
+    (00-59). The remainder of the string defines the relation of local
+    time to GMT. O is either + for a positive difference (local time is
+    later than GMT) or - (minus) for a negative difference. HH' is the
+    absolute value of the offset from GMT in hours, and mm' is the
+    absolute value of the offset in minutes. If no GMT information is
+    specified, the relation between the specified time and GMT is
+    considered unknown. Regardless of whether or not GMT
+    information is specified, the remainder of the string should specify
+    the local time.
+    """
+
+    pdfmark_date_fmt = r'%Y%m%d%H%M%S'
+    s = d.strftime(pdfmark_date_fmt)
+
+    tz = d.strftime('%z')
+    if tz == 'Z' or tz == '':
+        # Ghostscript <= 9.23 handles missing timezones incorrectly, so if
+        # timezone is missing, move it into GMT.
+        # https://bugs.ghostscript.com/show_bug.cgi?id=699182
+        s += "+00'00'"
+    else:
+        sign, tz_hours, tz_mins = tz[0], tz[1:3], tz[3:5]
+        s += "{}{}'{}'".format(sign, tz_hours, tz_mins)
+    return s
+
+
+def decode_pdf_date(s: str) -> datetime:
+    """
+    Decode a pdfmark date to a Python datetime object
+
+    A pdfmark date is a string in a paritcular format. See the pdfmark
+    Reference for the specification.
+
+    """
+    if s.startswith('D:'):
+        s = s[2:]
+
+    # Literal Z00'00', is incorrect but found in the wild,
+    # probably made by OS X Quartz -- standardize
+    if s.endswith("Z00'00'"):
+        s = s.replace("Z00'00'", '+0000')
+    elif s.endswith('Z'):
+        s = s.replace('Z', '+0000')
+
+    s = s.replace("'", "")  # Remove apos from PDF time strings
+
+    return datetime.strptime(s, r'%Y%m%d%H%M%S%z')
+
+
+def _get_pdfmark_dates(pdfmark):
+    """Encode dates for pdfmark Postscript.  The best way to deal with a
+    missing date entry is set it to null, because if the key is omitted
+    Ghostscript will set it to now - we do not want to erase the fact that
+    the value was unknown.  Setting to an empty string breaks Ghostscript
+    9.22 as reported here:
+    https://bugs.ghostscript.com/show_bug.cgi?id=699182
+    """
+
+    for key in ('/CreationDate', '/ModDate'):
+        if key not in pdfmark:
+            continue
+        if pdfmark[key].strip() == '':
+            yield '  {} null'.format(key)
+            continue
+        date_str = pdfmark[key]
+        if date_str.startswith('D:'):
+            date_str = date_str[2:]
+        try:
+            yield '  {} (D:{})'.format(
+                    key,
+                    encode_pdf_date(decode_pdf_date(date_str)))
+        except ValueError:
+            yield '  {} null'.format(key)
+
+
+def _get_pdfa_def(icc_profile, icc_identifier, pdfmark, ascii_docinfo=False):
+    """Create a Postscript pdfmark file for Ghostscript.
+
+    pdfmark contains the various objects as strings; these must be encoded in
+    ASCII, and dates have a special format.
+
+    :param icc_profile: filename of the ICC profile to include in pdfmark
+    :param icc_identifier: ICC identifier such as 'sRGB'
+    :param pdfmark: a dictionary containing keys to include the pdfmark
+    :param ascii_docinfo: if True, the docinfo block must be encoded in pure
+        ASCII and may not contain UTF-16BE-BOM-hex encoded strings, as
+        required for Ghostscript 9.24+
+
+    :returns: a string containing the entire pdfmark
+
+    """
+
+    # Ghostscript <= 9.21 has a bug where null entries in DOCINFO might produce
+    # ERROR: VMerror (-25) on closing pdfwrite device.
+    # https://bugs.ghostscript.com/show_bug.cgi?id=697684
+    # Work around this by only adding keys that have a nontrivial value
+    docinfo_keys = ('/Title', '/Author', '/Subject', '/Creator', '/Keywords')
+
+    def docinfo_gen():
+        if not ascii_docinfo:
+            docinfo_line_template = '  {key} <{value}>'
+            encode = encode_text_string
+        else:
+            docinfo_line_template = '  {key} ({value})'
+            encode = _encode_ascii
+        yield from _get_pdfmark_dates(pdfmark)
+        for key in docinfo_keys:
+            if key in pdfmark and pdfmark[key].strip() != '':
+                line = docinfo_line_template.format(
+                    key=key, value=encode(pdfmark[key]))
+                yield line
+    docinfo = '\n'.join(docinfo_gen())
+
+    t = Template(pdfa_def_template)
+    result = t.substitute(icc_profile=icc_profile,
+                          icc_identifier=icc_identifier,
+                          docinfo=docinfo)
+    return result
+
+
+def generate_pdfa_ps(target_filename, pdfmark, icc='sRGB', ascii_docinfo=False):
+    if icc == 'sRGB':
+        icc_profile = SRGB_ICC_PROFILE
+    else:
+        raise NotImplementedError("Only supporting sRGB")
+
+    ps = _get_pdfa_def(icc_profile, icc, pdfmark, ascii_docinfo=ascii_docinfo)
+
+    # We should have encoded everything to pure ASCII by this point, and
+    # to be safe, only allow ASCII in PostScript
+    with open(target_filename, 'w', encoding='ascii') as f:
+        f.write(ps)
+
+
+def file_claims_pdfa(filename):
+    """Determines if the file claims to be PDF/A compliant
+
+    Checking if a file is a truly compliant PDF/A is a massive undertaking
+    that no open source tool does properly.  Some commercial tools are
+    generally reliable (Acrobat).
+
+    This checks if the XMP metadata contains a PDF/A marker.
+    """
+    pdf = pypdf.PdfFileReader(filename)
+    try:
+        # Monkeypatch PyPDF2 to use defusedxml as its XML parser, for safety
+        with patch('xml.dom.minidom.parseString', new=defused_parseString):
+            xmp = pdf.getXmpMetadata()
+    except ExpatError:
+        return {'pass': False, 'output': 'pdf',
+                'conformance': 'Invalid XML metadata'}
+
+    try:
+        pdfa_nodes = xmp.getNodesInNamespace(
+            aboutUri='',
+            namespace='http://www.aiim.org/pdfa/ns/id/')
+    except AttributeError:
+        return {'pass': False, 'output': 'pdf',
+                'conformance': 'No XMP metadata'}
+
+    pdfa_dict = {attr.localName: attr.value for attr in pdfa_nodes}
+    if not pdfa_dict:
+        return {'pass': False, 'output': 'pdf',
+                'conformance': 'No XMP metadata'}
+
+    part_conformance = pdfa_dict['part'] + pdfa_dict['conformance']
+    valid_part_conforms = {'1A', '1B', '2A', '2B', '2U', '3A', '3B', '3U'}
+
+    conformance = 'PDF/A-{}'.format(
+        part_conformance)
+
+    if part_conformance in valid_part_conforms:
+        pdfa_dict['pass'] = True
+        pdfa_dict['output'] = 'pdfa'
+    pdfa_dict['conformance'] = conformance
+
+    return pdfa_dict

src/ocrmypdf/pdfinfo.py

@@ -0,0 +1,783 @@
+#!/usr/bin/env python3
+# © 2015 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from subprocess import Popen, PIPE
+from decimal import Decimal
+from math import hypot, isclose
+import re
+import sys
+from collections import namedtuple
+from collections.abc import MutableMapping, Mapping
+import warnings
+from pathlib import Path
+from enum import Enum
+from contextlib import contextmanager
+
+import PyPDF2 as pypdf
+from .lib import fitz
+from .helpers import universal_open, fspath
+
+
+
+matrix_mult = pypdf.pdf.utils.matrixMultiply
+
+Colorspace = Enum('Colorspace',
+                  'gray rgb cmyk lab icc index sep devn pattern jpeg2000')
+
+Encoding = Enum('Encoding',
+                'ccitt jpeg jpeg2000 jbig2 asciihex ascii85 lzw flate ' + \
+                'runlength')
+
+
+FRIENDLY_COLORSPACE = {
+    '/DeviceGray': Colorspace.gray,
+    '/CalGray': Colorspace.gray,
+    '/DeviceRGB': Colorspace.rgb,
+    '/CalRGB': Colorspace.rgb,
+    '/DeviceCMYK': Colorspace.cmyk,
+    '/Lab': Colorspace.lab,
+    '/ICCBased': Colorspace.icc,
+    '/Indexed': Colorspace.index,
+    '/Separation': Colorspace.sep,
+    '/DeviceN': Colorspace.devn,
+    '/Pattern': Colorspace.pattern,
+    '/G': Colorspace.gray,  # Abbreviations permitted in inline images
+    '/RGB': Colorspace.rgb,
+    '/CMYK': Colorspace.cmyk,
+    '/I': Colorspace.index,
+}
+
+FRIENDLY_ENCODING = {
+    '/CCITTFaxDecode': Encoding.ccitt,
+    '/DCTDecode': Encoding.jpeg,
+    '/JPXDecode': Encoding.jpeg2000,
+    '/JBIG2Decode': Encoding.jbig2,
+    '/CCF': Encoding.ccitt,  # Abbreviations permitted in inline images
+    '/DCT': Encoding.jpeg,
+    '/AHx': Encoding.asciihex,
+    '/A85': Encoding.ascii85,
+    '/LZW': Encoding.lzw,
+    '/Fl': Encoding.flate,
+    '/RL': Encoding.runlength
+}
+
+FRIENDLY_COMP = {
+    Colorspace.gray: 1,
+    Colorspace.rgb: 3,
+    Colorspace.cmyk: 4,
+    Colorspace.lab: 3,
+    Colorspace.index: 1
+}
+
+
+UNIT_SQUARE = (1.0, 0.0, 0.0, 1.0, 0.0, 0.0)
+
+
+def _matrix_from_shorthand(shorthand):
+    """Convert from PDF matrix shorthand to full matrix
+
+    PDF 1.7 spec defines a shorthand for describing the entries of a matrix
+    since the last column is always (0, 0, 1).
+    """
+
+    a, b, c, d, e, f = map(float, shorthand)
+    return ((a, b, 0),
+            (c, d, 0),
+            (e, f, 1))
+
+
+def _shorthand_from_matrix(matrix):
+    """Convert from transformation matrix to PDF shorthand."""
+    a, b = matrix[0][0], matrix[0][1]
+    c, d = matrix[1][0], matrix[1][1]
+    e, f = matrix[2][0], matrix[2][1]
+    return tuple(map(float, (a, b, c, d, e, f)))
+
+
+def _is_unit_square(shorthand):
+    values = map(float, shorthand)
+    pairwise = zip(values, UNIT_SQUARE)
+    return all([isclose(a, b, rel_tol=1e-3) for a, b in pairwise])
+
+XobjectSettings = namedtuple('XobjectSettings',
+    ['name', 'shorthand', 'stack_depth'])
+
+InlineSettings = namedtuple('InlineSettings',
+    ['settings', 'shorthand', 'stack_depth'])
+
+ContentsInfo = namedtuple('ContentsInfo', 
+    ['xobject_settings', 'inline_images', 'found_text'])
+
+
+def _normalize_stack(operations):
+    """Fix runs of qQ's in the stack
+
+    For some reason PyPDF2 converts runs of qqq, QQ, QQQq, etc. into single
+    operations.  Break this silliness up and issue each stack operation
+    individually so we don't lose count.
+
+    """
+    for operands, command in operations:
+        if re.match(br'Q*q+$', command):   # Zero or more Q, one or more q
+            for char in command:           # Split into individual bytes
+                yield ([], bytes([char]))  # Yield individual bytes
+        else:
+            yield (operands, command)
+
+
+def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE):
+    """Interpret the PDF content stream
+
+    The stack represents the state of the PDF graphics stack.  We are only
+    interested in the current transformation matrix (CTM) so we only track
+    this object; a full implementation would need to track many other items.
+
+    The CTM is initialized to the mapping from user space to device space.
+    PDF units are 1/72".  In a PDF viewer or printer this matrix is initialized
+    to the transformation to device space.  For example if set to
+    (1/72, 0, 0, 1/72, 0, 0) then all units would be calculated in inches.
+
+    Images are always considered to be (0, 0) -> (1, 1).  Before drawing an
+    image there should be a 'cm' that sets up an image coordinate system
+    where drawing from (0, 0) -> (1, 1) will draw on the desired area of the
+    page.
+
+    PDF units suit our needs so we initialize ctm to the identity matrix.
+
+    PyPDF2 replaces inline images with a fake "INLINE IMAGE" operator.
+
+    """
+
+    operations = contentstream.operations
+    stack = []
+    ctm = _matrix_from_shorthand(initial_shorthand)
+    xobject_settings = []
+    inline_images = []
+    found_text = False
+
+    for n, op in enumerate(_normalize_stack(operations)):
+        operands, command = op
+        if command == b'q':
+            stack.append(ctm)
+            if len(stack) > 32:
+                raise RuntimeError(
+                    "PDF graphics stack overflow, command %i" % n)
+        elif command == b'Q':
+            try:
+                ctm = stack.pop()
+            except IndexError:
+                raise RuntimeError(
+                    "PDF graphics stack underflow, command %i" % n)
+        elif command == b'cm':
+            ctm = matrix_mult(
+                _matrix_from_shorthand(operands), ctm)
+        elif command == b'Do':
+            image_name = operands[0]
+            settings = XobjectSettings(
+                name=image_name, shorthand=_shorthand_from_matrix(ctm),
+                stack_depth=len(stack))
+            xobject_settings.append(settings)
+        elif command == b'INLINE IMAGE':
+            settings = operands['settings']
+            inline = InlineSettings(
+                settings=settings, shorthand=_shorthand_from_matrix(ctm),
+                stack_depth=len(stack))
+            inline_images.append(inline)
+        elif command in (b'Tj', b'TJ', b'"', b"'"):
+            found_text = True
+
+
+    return ContentsInfo(
+        xobject_settings=xobject_settings,
+        inline_images=inline_images,
+        found_text=found_text)
+
+
+def _get_dpi(ctm_shorthand, image_size):
+    """Given the transformation matrix and image size, find the image DPI.
+
+    PDFs do not include image resolution information within image data.
+    Instead, the PDF page content stream describes the location where the
+    image will be rasterized, and the effective resolution is the ratio of the
+    pixel size to raster target size.
+
+    Normally a scanned PDF has the paper size set appropriately but this is
+    not guaranteed. The most common case is a cropped image will change the
+    page size (/CropBox) without altering the page content stream. That means
+    it is not sufficient to assume that the image fills the page, even though
+    that is the most common case.
+
+    A PDF image may be scaled (always), cropped, translated, rotated in place
+    to an arbitrary angle (rarely) and skewed. Only equal area mappings can
+    be expressed, that is, it is not necessary to consider distortions where
+    the effective DPI varies with position.
+
+    To determine the image scale, transform an offset axis vector v0 (0, 0),
+    width-axis vector v0 (1, 0), height-axis vector vh (0, 1) with the matrix,
+    which gives the dimensions of the image in PDF units. From there we can
+    compare to actual image dimensions. PDF uses
+    row vector * matrix_tranposed unlike the traditional
+    matrix * column vector.
+
+    The offset, width and height vectors can be combined in a matrix and
+    multiplied by the transform matrix. Then we want to calculated
+        magnitude(width_vector - offset_vector)
+    and
+        magnitude(height_vector - offset_vector)
+
+    When the above is worked out algebraically, the effect of translation
+    cancels out, and the vector magnitudes become functions of the nonzero
+    transformation matrix indices. The results of the derivation are used
+    in this code.
+
+    pdfimages -list does calculate the DPI in some way that is not completely
+    naive, but it does not get the DPI of rotated images right, so cannot be
+    used anymore to validate this. Photoshop works, or using Acrobat to
+    rotate the image back to normal.
+
+    It does not matter if the image is partially cropped, or even out of the
+    /MediaBox.
+
+    """
+
+    a, b, c, d, _, _ = ctm_shorthand
+
+    # Calculate the width and height of the image in PDF units
+    image_drawn_width = hypot(a, b)
+    image_drawn_height = hypot(c, d)
+
+    # The scale of the image is pixels per unit of default user space (1/72")
+    scale_w = image_size[0] / image_drawn_width
+    scale_h = image_size[1] / image_drawn_height
+
+    # DPI = scale * 72
+    dpi_w = scale_w * 72.0
+    dpi_h = scale_h * 72.0
+
+    return dpi_w, dpi_h
+
+
+class ImageInfo:
+    DPI_PREC = Decimal('1.000')
+
+    def __init__(self, *, name='', pdfimage=None, inline=None,
+                 shorthand=None):
+
+        self._name = name
+        self._shorthand = shorthand
+        if inline:
+            # Fixme does not work for inline images with non abbreviated
+            # fields
+            self._origin = 'inline'
+            self._width = inline.settings['/W']
+            self._height = inline.settings['/H']
+            self._bpc = inline.settings.get('/BPC', 8)
+            try:
+                self._color = FRIENDLY_COLORSPACE[inline.settings['/CS']]
+            except Exception:
+                self._color = '-'
+            self._comp = FRIENDLY_COMP.get(self._color, '?')
+            if '/F' in inline.settings:
+                filter_ = inline.settings['/F']
+                if isinstance(filter_, pypdf.generic.ArrayObject):
+                    filter_ = filter_[0]
+                self._enc = FRIENDLY_ENCODING.get(filter_, 'image')
+            else:
+                self._enc = 'image'
+        elif pdfimage:
+            self._origin = 'xobject'
+            self._width = pdfimage['/Width']
+            self._height = pdfimage['/Height']
+            if '/BitsPerComponent' in pdfimage:
+                self._bpc = pdfimage['/BitsPerComponent']
+            else:
+                self._bpc = 8
+
+            # Fixme: this is incorrectly treats explicit masks as stencil masks,
+            # but good enough for now. Explicit masks have /ImageMask true but are
+            # never called for in content stream, instead are drawn as a /Mask on
+            # other images. For our purposes finding out the details of /Mask
+            # will seldom matter.
+            if '/ImageMask' in pdfimage:
+                self._type = 'stencil' if pdfimage['/ImageMask'].value \
+                    else 'image'
+            else:
+                self._type = 'image'
+            if '/Filter' in pdfimage:
+                filter_ = pdfimage['/Filter']
+                if isinstance(filter_, pypdf.generic.ArrayObject):
+                    filter_ = filter_[0]
+                self._enc = FRIENDLY_ENCODING.get(filter_, 'image')
+            else:
+                self._enc = 'image'
+            if '/ColorSpace' in pdfimage:
+                cs = pdfimage['/ColorSpace']
+                if isinstance(cs, pypdf.generic.ArrayObject):
+                    cs = cs[0]
+                self._color = FRIENDLY_COLORSPACE.get(cs, '-')
+            else:
+                self._color = FRIENDLY_COLORSPACE[Colorspace.jpeg2000] \
+                    if self._enc == Encoding.jpeg2000 else '?'
+
+            self._comp = FRIENDLY_COMP.get(self._color, '?')
+
+            # Bit of a hack... infer grayscale if component count is uncertain
+            # but encoding must be monochrome. This happens if a monochrome image
+            # has an ICC profile attached. Better solution would be to examine
+            # the ICC profile.
+            if self._comp == '?' and self._enc in (Encoding.ccitt, 'jbig2'):
+                self._comp = FRIENDLY_COMP[Colorspace.gray]
+
+    @property
+    def name(self):
+        return self._name
+
+    @property
+    def type_(self):
+        return self._type
+
+    @property
+    def width(self):
+        return self._width
+
+    @property
+    def height(self):
+        return self._height
+
+    @property
+    def bpc(self):
+        return self._bpc
+
+    @property
+    def color(self):
+        return self._color
+
+    @property
+    def comp(self):
+        return self._comp
+
+    @property
+    def enc(self):
+        return self._enc
+
+    @property
+    def xres(self):
+        return _get_dpi(self._shorthand, (self._width, self._height))[0]
+
+    @property
+    def yres(self):
+        return _get_dpi(self._shorthand, (self._width, self._height))[1]
+
+    def __getitem__(self, item):
+        warnings.warn("ImageInfo.__getitem__", DeprecationWarning)
+        if item in ('name', 'width', 'height', 'bpc', 'color', 'comp', 'enc'):
+            return getattr(self, item)
+        elif item == 'dpi_w':
+            return Decimal(self.xres).quantize(self.DPI_PREC)
+        elif item == 'dpi_h':
+            return Decimal(self.yres).quantize(self.DPI_PREC)
+        elif item == 'dpi':
+            return Decimal(self.xres * self.yres).sqrt().quantize(
+                self.DPI_PREC)
+        else:
+            raise KeyError(item)
+
+    def __repr__(self):
+        class_locals = {attr: getattr(self, attr, None) for attr in dir(self)
+                        if not attr.startswith('_')}
+        return (
+            "<ImageInfo '{name}' {type_} {width}x{height} {color} "
+            "{comp} {bpc} {enc} {xres}x{yres}>").format(**class_locals)
+
+
+def _find_inline_images(contentsinfo):
+    "Find inline images in the contentstream"
+
+    for n, inline in enumerate(contentsinfo.inline_images):
+        yield ImageInfo(name='inline-%02d' % n, shorthand=inline.shorthand,
+                        inline=inline)
+
+
+def _image_xobjects(container):
+    """Search for all XObject-based images in the container
+
+    Usually the container is a page, but it could also be a Form XObject
+    that contains images. Filter out the Form XObjects which are dealt with
+    elsewhere.
+
+    Generate a sequence of tuples (image, xobj container), where container,
+    where xobj is the name of the object and image is the object itself,
+    since the object does not know its own name.
+
+    """
+
+    if '/Resources' not in container:
+        return
+    resources = container['/Resources']
+    if '/XObject' not in resources:
+        return
+    for xobj in resources['/XObject']:
+        candidate = resources['/XObject'][xobj]
+        if candidate['/Subtype'] == '/Image':
+            pdfimage = candidate
+            yield (pdfimage, xobj)
+
+
+def _find_regular_images(container, contentsinfo):
+    """Find images stored in the container's /Resources /XObject
+
+    Usually the container is a page, but it could also be a Form XObject
+    that contains images.
+
+    Generates images with their DPI at time of drawing.
+
+    """
+
+    for pdfimage, xobj in _image_xobjects(container):
+
+        # For each image that is drawn on this, check if we drawing the
+        # current image - yes this is O(n^2), but n == 1 almost always
+        for draw in contentsinfo.xobject_settings:
+            if draw.name != xobj:
+                continue
+
+            if draw.stack_depth == 0 and _is_unit_square(draw.shorthand):
+                # At least one PDF in the wild (and test suite) draws an image
+                # when the graphics stack depth is 0, meaning that the image
+                # gets drawn into a square of 1x1 PDF units (or 1/72",
+                # or 0.35 mm).  The equivalent DPI will be >100,000.  Exclude
+                # these from our DPI calculation for the page.
+                continue
+
+            yield ImageInfo(name=draw.name, pdfimage=pdfimage, shorthand=
+                            draw.shorthand)
+
+
+def _find_form_xobject_images(pdf, container, contentsinfo):
+    """Find any images that are in Form XObjects in the container
+
+    The container may be a page, or a parent Form XObject.
+
+    """
+    if '/Resources' not in container:
+        return
+    resources = container['/Resources']
+    if '/XObject' not in resources:
+        return
+    for xobj in resources['/XObject']:
+        candidate = resources['/XObject'][xobj]
+        if candidate['/Subtype'] != '/Form':
+            continue
+
+        form_xobject = candidate
+        for settings in contentsinfo.xobject_settings:
+            if settings.name != xobj:
+                continue
+
+            # Find images once for each time this Form XObject is drawn.
+            # This could be optimized to cache the multiple drawing events
+            # but in practice both Form XObjects and multiple drawing of the
+            # same object are both very rare.
+            ctm_shorthand = settings.shorthand
+            yield from _find_images(
+                pdf=pdf, container=form_xobject, shorthand=ctm_shorthand)
+
+
+def _find_images(*, pdf, container, shorthand=None):
+    """Find all individual instances of images drawn in the container
+
+    Usually the container is a page, but it may also be a Form XObject.
+
+    On a typical page images are stored inline or as regular images
+    in an XObject.
+
+    Form XObjects may include inline images, XObject images,
+    and recursively, other Form XObjects; and also vector drawing commands.
+
+    Every instance of an image being drawn somewhere is flattened and
+    treated as a unique image, since if the same image is drawn multiple times
+    on one page it may be drawn at differing resolutions, and our objective
+    is to find the resolution at which the page can be rastered without
+    downsampling.
+
+    """
+
+    if container.get('/Type') == '/Page' and '/Contents' in container:
+        # For a /Page the content stream is attached to the page's /Contents
+        page = container
+        contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
+        initial_shorthand = shorthand or UNIT_SQUARE
+    elif container.get('/Type') == '/XObject' and \
+            container['/Subtype'] == '/Form':
+        # For a Form XObject that content stream is attached to the XObject
+        contentstream = pypdf.pdf.ContentStream(container, pdf)
+
+        # Set the CTM to the state it was when the "Do" operator was
+        # encountered that is drawing this instance of the Form XObject
+        ctm = _matrix_from_shorthand(shorthand or UNIT_SQUARE)
+
+        # A Form XObject may provide its own matrix to map form space into
+        # user space. Get this if one exists
+        form_matrix = _matrix_from_shorthand(
+                container.get('/Matrix', UNIT_SQUARE))
+
+        # Concatenate form matrix with CTM to ensure CTM is correct for
+        # drawing this instance of the XObject
+        ctm = matrix_mult(form_matrix, ctm)
+        initial_shorthand = _shorthand_from_matrix(ctm)
+    else:
+        return
+
+    contentsinfo = _interpret_contents(contentstream, initial_shorthand)
+
+    yield from _find_inline_images(contentsinfo)
+    yield from _find_regular_images(container, contentsinfo)
+    yield from _find_form_xobject_images(pdf, container, contentsinfo)
+
+
+def _naive_find_text(*, pdf, page):
+    if not(page.get('/Type') == '/Page' and '/Contents' in page):
+        # Not a page, or has no /Contents => no text
+        return False
+
+    # First we check the main content stream    
+    contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
+    contentsinfo = _interpret_contents(contentstream, UNIT_SQUARE)
+    if contentsinfo.found_text:
+        return True
+
+    # Then see if there is a Form XObject with with a content stream
+    # that might have text.  For full completeness we should recursively
+    # search nested Form XObjects, as we do with images.  But that is
+    # rare.
+    if '/Resources' in page:
+        resources = page['/Resources']
+        if '/XObject' in resources:    
+            for xobj in resources['/XObject']:
+                candidate = resources['/XObject'][xobj]
+                if candidate['/Subtype'] != '/Form':
+                    continue
+                form_xobject = candidate                
+                # Content stream is attached to Form XObject dictionary
+                contentstream = pypdf.pdf.ContentStream(form_xobject, pdf)
+                sub_contentsinfo = _interpret_contents(
+                    contentstream, UNIT_SQUARE)
+                if sub_contentsinfo.found_text:
+                    return True
+    return False
+
+
+def _page_has_text(infile, pageno):
+    doc = fitz.Document(infile)
+    text = doc.getPageText(pageno)
+    if text.strip() != '':
+        return True
+    return False
+
+
+def _pdf_get_pageinfo(pdf, pageno: int, infile):
+    pageinfo = {}
+    pageinfo['pageno'] = pageno
+    pageinfo['images'] = []
+
+    if isinstance(pdf, Path):
+        pdf = pypdf.PdfFileReader(str(pdf))
+    elif isinstance(pdf, str):
+        pdf = pypdf.PdfFileReader(pdf)
+
+    page = pdf.pages[pageno]
+
+    if fitz:
+        pageinfo['has_text'] = _page_has_text(str(infile), pageno)
+    else:
+        pageinfo['has_text'] = _naive_find_text(pdf=pdf, page=page)
+
+    width_pt = page.mediaBox.getWidth()
+    height_pt = page.mediaBox.getHeight()
+
+    userunit = page.get('/UserUnit', Decimal(1.0))
+    pageinfo['userunit'] = userunit
+    pageinfo['width_inches'] = width_pt * userunit / Decimal(72.0)
+    pageinfo['height_inches'] = height_pt * userunit / Decimal(72.0)
+
+    try:
+        pageinfo['rotate'] = int(page['/Rotate'])
+    except KeyError:
+        pageinfo['rotate'] = 0
+
+    userunit_shorthand = (userunit, 0, 0, userunit, 0, 0)
+    pageinfo['images'] = [im for im in
+                          _find_images(pdf=pdf, container=page,
+                                       shorthand=userunit_shorthand)]
+    if pageinfo['images']:
+        xres = max(image['dpi_w'] for image in pageinfo['images'])
+        yres = max(image['dpi_h'] for image in pageinfo['images'])
+        pageinfo['xres'], pageinfo['yres'] = xres, yres
+        pageinfo['width_pixels'] = \
+            int(round(xres * pageinfo['width_inches']))
+        pageinfo['height_pixels'] = \
+            int(round(yres * pageinfo['height_inches']))
+
+    return pageinfo
+
+
+def _pdf_get_all_pageinfo(infile):
+    with universal_open(infile, 'rb') as f:
+        pdf = pypdf.PdfFileReader(f)
+        return [PageInfo(pdf, n, infile) for n in range(pdf.numPages)]
+
+
+class PageInfo:
+    def __init__(self, pdf, pageno, infile):
+        self._pageno = pageno
+        self._infile = infile
+        self._pageinfo = _pdf_get_pageinfo(pdf, pageno, infile)
+
+    @property
+    def pageno(self):
+        return self._pageno
+
+    @property
+    def has_text(self):
+        return self._pageinfo['has_text']
+
+    @property
+    def width_inches(self):
+        return self._pageinfo['width_inches']
+
+    @property
+    def height_inches(self):
+        return self._pageinfo['height_inches']
+
+    @property
+    def width_pixels(self):
+        return int(round(self.width_inches * self.xres))
+
+    @property
+    def height_pixels(self):
+        return int(round(self.height_inches * self.yres))
+
+    @property
+    def rotation(self):
+        return self._pageinfo.get('rotate', None)
+
+    @rotation.setter
+    def rotation(self, value):
+        if value in (0, 90, 180, 270, 360, -90, -180, -270):
+            self._pageinfo['rotate'] = value
+        else:
+            raise ValueError("rotation must be a cardinal angle")
+
+    @property
+    def images(self):
+        return self._pageinfo['images']
+
+    @property
+    def xres(self):
+        return self._pageinfo.get('xres', None)
+
+    @property
+    def yres(self):
+        return self._pageinfo.get('yres', None)
+
+    @property
+    def userunit(self):
+        return self._pageinfo.get('userunit', None)
+
+    @property
+    def min_version(self):
+        if self.userunit is not None:
+            return '1.6'
+        else:
+            return '1.5'
+
+    def __repr__(self):
+        return (
+            '<PageInfo '
+            'pageno={} {}"x{}" rotation={} res={}x{} has_text={}>').format(
+            self.pageno, self.width_inches, self.height_inches,
+            self.rotation,
+            self.xres, self.yres, self.has_text
+        )
+
+
+class PdfInfo:
+    """Get summary information about a PDF
+    
+    """
+    def __init__(self, infile):
+        self._infile = infile
+        self._pages = _pdf_get_all_pageinfo(infile)
+        if fitz:
+            self._toc = fitz.Document(fspath(infile)).getToC()
+        else:
+            self._toc = []
+
+    @property
+    def pages(self):
+        return self._pages
+
+    @property
+    def min_version(self):
+        # The minimum PDF is the maximum version that any particular page needs
+        return max(page.min_version for page in self.pages)
+
+    @property
+    def has_userunit(self):
+        return any(page.userunit != 1.0 for page in self.pages)
+
+    @property
+    def filename(self):
+        if not isinstance(self._infile, (str, Path)):
+            raise NotImplementedError("can't get filename from stream")
+        return self._infile
+
+    @property
+    def table_of_contents(self):
+        return self._toc
+
+    def __getitem__(self, item):
+        return self._pages[item]
+
+    def __len__(self):
+        return len(self._pages)
+
+    def __repr__(self):
+        return "<PdfInfo('...'), page count={}>".format(len(self))
+
+    # def __getstate__(self):
+    #     state = {'_infile': self._infile}
+    #     return state
+    #
+    # def __setstate__(self, state):
+    #     self._infile = state['_infile']
+    #     self._pages = _pdf_get_all_pageinfo(self._infile)
+
+
+def main():
+    import argparse
+    parser = argparse.ArgumentParser()
+    parser.add_argument('infile')
+    args = parser.parse_args()
+    info = _pdf_get_all_pageinfo(args.infile)
+    from pprint import pprint
+    pprint(info)
+
+
+if __name__ == '__main__':
+    main()

test_requirements.txt

@@ -1,2 +1,5 @@
-img2pdf>=0.1.5
-pytest>=2.7.2
+pytest >= 3.2
+pytest-helpers-namespace
+pytest-xdist
+pytest-cov
+pytest-timeout == 1.2.1

tests/cache/2400dpi/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/2400dpi/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -0,0 +1,25 @@
+a la Waterman
+
+4 ons linzen
+
+3 liter water
+
+3 uien
+
+bloem, boter
+
+2 kopjes melk
+
+laurier, kruidnagel, kerrie, zout
+
+De linzgen wassen en in -l liter kokend wa-
+ter 1 dag laten weken, 2 liter water bij
+de linzen voegen, zonder het water waarin
+ze geweekt zijn af te gieten, De helft van
+de uien bakken met laurier en Kruicdnagel.
+Alle uien, kerrie en gout bij de linzen
+voegen, Alles aan de kook brengen,. Van de
+bloem met boter en melk een papje maken en
+verder afmaken met de soep, Als de linzen
+gaar Zijn is de soep klaar.
+

tests/cache/aspect/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -0,0 +1,3 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
+Too few characters. Skipping this page
+OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -0,0 +1,4 @@
+ 
+
+This should be a perfect circle:
+

tests/cache/aspect/__-l__eng__000001.image__000001.rendered__pdf__txt/stderr.bin

@@ -0,0 +1,3 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
+Too few characters. Skipping this page
+OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__000001.image__000001.rendered__pdf__txt/txt.bin

@@ -0,0 +1,4 @@
+ 
+
+This should be a perfect circle:
+

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/hocr.bin

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ <head>
+  <title></title>
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
+  <meta name='ocr-system' content='tesseract 4.0.0-beta.1' />
+  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
+</head>
+<body>
+  <div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.4nwu81i8/000001.ocr.png"; bbox 0 0 1000 800; ppageno 0'>
+   <div class='ocr_carea' id='block_1_1' title="bbox 296 96 704 504">
+    <p class='ocr_par' id='par_1_1' lang='eng' title="bbox 296 96 704 504">
+     <span class='ocr_line' id='line_1_1' title="bbox 296 96 704 504; baseline 0 296; x_size 169.33333; x_descenders 42.333332; x_ascenders 42.333336"><span class='ocrx_word' id='word_1_1' title='bbox 296 96 704 504; x_wconf 95'><strong> </strong></span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_2' title="bbox 150 592 841 622">
+    <p class='ocr_par' id='par_1_2' lang='eng' title="bbox 150 592 841 622">
+     <span class='ocr_line' id='line_1_2' title="bbox 150 592 841 622; baseline 0 -6; x_size 30; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_2' title='bbox 150 592 230 616; x_wconf 96'><strong>This</strong></span> <span class='ocrx_word' id='word_1_3' title='bbox 260 592 384 616; x_wconf 95'><strong>should</strong></span> <span class='ocrx_word' id='word_1_4' title='bbox 413 592 449 616; x_wconf 95'><strong>be</strong></span> <span class='ocrx_word' id='word_1_5' title='bbox 479 600 493 616; x_wconf 95'><strong>a</strong></span> <span class='ocrx_word' id='word_1_6' title='bbox 523 592 668 622; x_wconf 95'><strong>perfect</strong></span> <span class='ocrx_word' id='word_1_7' title='bbox 698 592 841 616; x_wconf 55'><strong>circle:</strong></span> 
+     </span>
+    </p>
+   </div>
+  </div>
+ </body>
+</html>

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/stderr.bin

@@ -0,0 +1,3 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
+Too few characters. Skipping this page
+OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/txt.bin

@@ -0,0 +1,4 @@
+ 
+
+This should be a perfect circle:
+

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers!
+
+¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+¢ One or all tracks may be TRANSPOSED at the touch of a key.
+e Exclusive real-time ERASE function makes editing FAST.
+* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+¢ Optional SMPTE time code synchronization.
+
+© Optional remote control.
+
+Recording a Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+— existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot-recording. To overdub a new part,
+select a different track and start recording—while you
+record, the first track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections,
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+* Non-destructive recording—existing notes are not erased while recording.
+¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+© Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+¢ Any TIME SIGNATURE may be used, and may be changed within a song.
+
+linn
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 708-8131 TELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000002.ocr.png__000002.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000002.ocr.png__000002.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers!
+
+¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+¢ One or all tracks may be TRANSPOSED at the touch of a key.
+e Exclusive real-time ERASE function makes editing FAST.
+* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+¢ Optional SMPTE time code synchronization.
+
+© Optional remote control.
+
+Recording a Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+— existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot-recording. To overdub a new part,
+select a different track and start recording—while you
+record, the first track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections,
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+* Non-destructive recording—existing notes are not erased while recording.
+¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+© Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+¢ Any TIME SIGNATURE may be used, and may be changed within a song.
+
+linn
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 708-8131 TELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000003.ocr.png__000003.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000003.ocr.png__000003.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers!
+
+¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+¢ One or all tracks may be TRANSPOSED at the touch of a key.
+e Exclusive real-time ERASE function makes editing FAST.
+* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+¢ Optional SMPTE time code synchronization.
+
+© Optional remote control.
+
+Recording a Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+— existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot-recording. To overdub a new part,
+select a different track and start recording—while you
+record, the first track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections,
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+* Non-destructive recording—existing notes are not erased while recording.
+¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+© Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+¢ Any TIME SIGNATURE may be used, and may be changed within a song.
+
+linn
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 708-8131 TELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000004.ocr.png__000004.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000004.ocr.png__000004.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers!
+
+¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+¢ One or all tracks may be TRANSPOSED at the touch of a key.
+e Exclusive real-time ERASE function makes editing FAST.
+* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+¢ Optional SMPTE time code synchronization.
+
+© Optional remote control.
+
+Recording a Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+— existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot-recording. To overdub a new part,
+select a different track and start recording—while you
+record, the first track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections,
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+* Non-destructive recording—existing notes are not erased while recording.
+¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+© Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+¢ Any TIME SIGNATURE may be used, and may be changed within a song.
+
+linn
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 708-8131 TELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__osd__--psm__0__000001.ocr.preview.jpg__stdout/stderr.bin

@@ -0,0 +1 @@
+Warning. Invalid resolution 0 dpi. Using 70 instead.

tests/cache/cardinal/__-l__osd__--psm__0__000001.ocr.preview.jpg__stdout/stdout.bin

@@ -0,0 +1,6 @@
+Page number: 0
+Orientation in degrees: 0
+Rotate: 0
+Orientation confidence: 18.91
+Script: Latin
+Script confidence: 0.97

tests/cache/cardinal/__-l__osd__--psm__0__000002.ocr.preview.jpg__stdout/stderr.bin

@@ -0,0 +1 @@
+Warning. Invalid resolution 0 dpi. Using 70 instead.