v5.6.3 notes

Did not account for API difference in pathlib

.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 +1,54 @@
+# OCRmyPDF
+#
+FROM      ubuntu:17.10
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  software-properties-common python-software-properties \
+  python3-wheel \
+  python3-reportlab \
+  python3-venv \
+  ghostscript \
+  qpdf \
+  poppler-utils \
+  unpaper \
+  libffi-dev \ 
+  tesseract-ocr \
+  tesseract-ocr-eng \
+  tesseract-ocr-fra \
+  tesseract-ocr-spa \
+  tesseract-ocr-deu
+
+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
+
+# 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

@@ -1,16 +1,17 @@
 # OCRmyPDF polyglot
 #
-# VERSION               3.2
 FROM      jbarlow83/ocrmypdf:latest
-MAINTAINER James R. Barlow <jim@purplerock.ca>
+
+USER root
 
 # Update system and install our dependencies
-USER root
 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-wrapper.sh"]
+ENTRYPOINT ["/application/.docker/docker-wrapper.sh"]

.docker/Dockerfile.tess4

@@ -0,0 +1,64 @@
+# OCRmyPDF
+#
+FROM 	ubuntu:17.10
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  software-properties-common python-software-properties \
+  python3-wheel \
+  python3-reportlab \
+  python3-venv \
+  ghostscript \
+  qpdf \
+  poppler-utils \
+  unpaper \
+  libffi-dev
+
+RUN add-apt-repository ppa:alex-p/tesseract-ocr
+
+RUN apt-get update \
+	&& apt-get autoremove -y \
+	&& apt-get install -y --no-install-recommends \
+		tesseract-ocr \
+		tesseract-ocr-eng \
+		tesseract-ocr-fra \
+		tesseract-ocr-deu \
+		tesseract-ocr-spa \
+		tesseract-ocr-por \
+		tesseract-ocr-ara \
+		tesseract-ocr-rus \
+		tesseract-ocr-chi-sim 
+
+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
+
+# 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"]

.dockerignore

@@ -3,19 +3,13 @@
 *.pyc
 *.rst
 *.sublime*
-*/*.pyc
-*/*/*.pyc
-*/*/*/*.pyc
-*/*/*/*/*.pyc
-*/*/*/*/*/*.pyc
-*/*/*/*/*/*/*.pyc
-*/*/*/*/*/*/*/*.pyc
-.cache/
-.git/
-.ipynb_checkpoints/
+**/*.pyc
+.*/
+!.docker/
 .ruffus_history.sqlite
 bin/
 build/
+docs/
 dist/
 htmlcov/
 include/
@@ -27,5 +21,5 @@ tests/cache/
 tests/output/
 tests/resources/private/
 tmp/
-venv-3.4/
-venv-3.5/
+venv*/
+wheelhouse/

.gitattributes

@@ -6,5 +6,8 @@
 *.jar	binary
 *.pdf	binary
 *.PDF	binary
+*.png	binary
+*.jpg	binary
+
 
 .git_archival.txt  export-subst

.github/ISSUE_TEMPLATE

@@ -0,0 +1,3 @@
+Please include the command line and (if needed) 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).

.gitignore

@@ -1,8 +1,12 @@
 # Development environment
 *.pyc
 *.sublime-*
-venv-*/
+venv*/
 pyvenv.cfg
+tasks.py
+.bash_history
+.ruffus_history.sqlite
+.idea/
 
 # Package building
 *.egg-info/
@@ -15,6 +19,10 @@ wheelhouse/
 # Automatically generated files
 ocrmypdf/lib/_*.py
 ocrmypdf/version.py
+docs/_build/
+docs/_static/
+docs/_templates/
+docs/Makefile
 
 # Code coverage
 .coverage
@@ -23,8 +31,14 @@ htmlcov/
 # Testing
 log/
 /*.pdf
+/*.qdf
+*.ipynb
 .ipynb_checkpoints/
 tests/cache/
 tests/output/
-tests/resources/private
+tests/resources/private/
 tmp/
+pdfbox-app*.jar
+.vscode/
+IDEAS
+_Dockerfile.local

.travis.yml

@@ -1,44 +1,100 @@
+dist: trusty
 language: python
-
-cache: pip
-
 cache:
-    directories:
-        - $HOME/.ccache
-        - packages
-        - tests/cache
+  pip: true
+  directories:
+  - packages
+  - tests/cache
+  - $HOME/Library/Caches/Homebrew
 
-python:
-    - 3.4
+env:
+  global:
+    - secure: "oyX5xesoHD7qcDXKIxMyGZPi+H/WxcvfFkaviEmq84K1DDyHk48+9e92IKgrw8/lcTADnEo/AgVKfnhCPflFimk1xTkgaK4sUg1WLI2YjmaHcwl5SlBHa2rN3uGBwy1hyP92qyv/mMc9R59NtRJ8u76lbn6eN9wi7lkFWdE6BTw="  # DOCKERHUB_OCRMYPDF_TOKEN
+    - secure: "WlyII8YLsiUUyLtEA563GvEZmbneDb/T8q/P1uNbyQ2ps1U82tH0nSUV2CspSMxOFtZzPHCrRvnAmuTYKshBj+GNnBb1J9FKQmFwF+4NPeqsFdUkQ1NeeCmfIRShuNC3Otg2GGwj4Zssdg+QnVy43t2L11qizzfY+lY+MVzAYcM="  # DOCKERHUB_OCRMYPDF_TESS4_TOKEN
+    - secure: "hsf6MT+n2x3OiDM2fQyJZdV0/PWYmv81LdVqC6cfnHBE/8N3DloJRqQ7WfO14TxhiK9PEC7MpyCj0lSabUHEO7gSH6Vks6I1asoSkt8S9/bSMlhT4hei+pwVpeGEiU5xHVATNjY+D919VC3IFvc3XmjT74h/2SLhaZ+jhEmDggM="  # HOMEBREW_OCRMYPDF_TOKEN
+
+matrix:
+  include:
+    - os: linux
+      sudo: required
+      language: python
+      python: 3.5
+    - os: linux
+      sudo: required
+      language: python
+      python: 3.6
+    - os: osx
+      osx_image: xcode8
+      language: generic
 
 before_cache:
-    - rm -f $HOME/.cache/pip/log/debug.log
+- rm -f $HOME/.cache/pip/log/debug.log
 
-before_install:
-    # Ubuntu packages
-    - sudo add-apt-repository ppa:evl.ms/precise -y  # for Ghostscript 9.15
-    - sudo add-apt-repository ppa:lyrasis/precise-backports -y  # for Tesseract 3.03
-    - sudo add-apt-repository ppa:b-eltzner/qpdfview-exp -y  # for QPDF 5
-    - sudo add-apt-repository ppa:itachi-san/ffmpeg -y  # for libav 11.2 (for unpaper)
-    - sudo apt-get update -qq  # must go after all add-apt-repo
-    - sudo apt-get install -y ghostscript tesseract-ocr tesseract-ocr-deu tesseract-ocr-eng tesseract-ocr-fra qpdf poppler-utils libavformat56 libavcodec56 libavutil54 libffi-dev
-
-    # pip
-    - pip install --upgrade pip
-
-    # Download, make and install unpaper (using ccache)
-    - mkdir -p packages
-    - "[ -f packages/unpaper_6.1-1.deb ] || wget -q https://dl.dropboxusercontent.com/u/28971240/unpaper_6.1-1.deb -O packages/unpaper_6.1-1.deb"
-    - sudo dpkg -i packages/unpaper_6.1-1.deb
+before_install: |
+  if [[ "$TRAVIS_OS_NAME" == "linux" ]]; then
+    bash .travis/linux_before_install.sh
+  elif [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
+    brew update && brew bundle --file=.travis/Brewfile
+    pip3 install --upgrade pip
+    pip3 install wheel
+  fi
 
 install:
-    - pip install -r requirements.txt
-    - pip install -r test_requirements.txt
+- pip3 install .
+- pip3 install -r requirements.txt -r test_requirements.txt
 
 script:
-    - python setup.py clean
-    - python setup.py install
-    - python setup.py test
+- mv ocrmypdf dont_import_this_ocrmypdf
+- tesseract --version
+- qpdf --version
+- pytest
+- mv dont_import_this_ocrmypdf ocrmypdf
 
-os:
-    - linux
+deploy:
+  # release for main pypi
+  # 3.6 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
+  server: https://upload.pypi.org/legacy/
+  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"
+  skip_upload_docs: true
+
+  # test pypi
+- provider: pypi
+  server: https://testpypi.pypi.org/legacy/
+  user: ocrmypdf-travis
+  password:
+    secure: "DTFOmmNL6olA0+yXvp4u9jXZlZeqrJsJ0526jzqf4a3gZ6jnGTq5UI6WzRsslSyoMMfXKtHQebqHM6ogSgCZinyZ3ufHJo8fn9brxbEc2gsiWkbj5o3bGwdWMT1vNNE7XW0VCpw87rZ1EEwjl4FJHFudMlPR1yfU5+uq0k0PACo="
+  distributions: "sdist"
+  on:
+    branch: develop
+    tags: false
+    condition: $TRAVIS_OS_NAME == "osx"
+  skip_upload_docs: true
+
+  # null deploy for osx
+  # we really just want to run after_deploy *after* pypi upload is done, but
+  # after_deploy on runs if a given box deployed
+- provider: script
+  script: /usr/bin/true
+  on:
+    branch: master
+    tags: true
+    condition: $TRAVIS_OS_NAME == "osx"
+
+after_deploy: |
+  if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
+    bash .travis/osx_brew.sh
+  elif [[ "$TRAVIS_PYTHON_VERSION" == "3.6" && "$TRAVIS_OS_NAME" == "linux" ]]; then
+    curl -H "Content-Type: application/json" --data '{"build": true}' -X POST https://registry.hub.docker.com/u/jbarlow83/ocrmypdf/trigger/$DOCKERHUB_OCRMYPDF_TOKEN/
+    curl -H "Content-Type: application/json" --data '{"build": true}' -X POST https://registry.hub.docker.com/u/jbarlow83/ocrmypdf-tess4/trigger/$DOCKERHUB_OCRMYPDF_TESS4_TOKEN/
+  fi

.travis/Brewfile

@@ -0,0 +1,12 @@
+brew 'python'
+
+brew 'ghostscript'
+brew 'jbig2dec'
+brew 'leptonica'
+brew 'libffi'
+brew 'libtiff'
+brew 'libxml2'
+brew 'openjpeg'
+brew 'qpdf'
+brew 'tesseract'
+brew 'unpaper'

.travis/autobrew.py

@@ -0,0 +1,92 @@
+#!/usr/bin/env python3
+# © 2017-18 James R. Barlow: github.com/jbarlow83
+
+from string import Template
+from subprocess import run, PIPE
+import re
+
+recipe_template = Template("""
+class Ocrmypdf < Formula
+  include Language::Python::Virtualenv
+
+  desc "Adds an OCR text layer to scanned PDF files"
+  homepage "https://github.com/jbarlow83/OCRmyPDF"
+  ${ocrmypdf_url}
+  ${ocrmypdf_sha256}
+
+  depends_on "pkg-config" => :build
+  depends_on "freetype"
+  depends_on "ghostscript"
+  depends_on "jpeg"
+  depends_on "libpng"
+  depends_on "python"
+  depends_on "qpdf"
+  depends_on "tesseract"
+  depends_on "unpaper"
+
+${resources}
+  def install
+    venv = virtualenv_create(libexec, "python3")
+
+    resource("Pillow").stage do
+      inreplace "setup.py" do |s|
+        sdkprefix = MacOS::CLT.installed? ? "" : MacOS.sdk_path
+        s.gsub! "openjpeg.h", "probably_not_a_header_called_this_eh.h"
+        s.gsub! "ZLIB_ROOT = None", "ZLIB_ROOT = ('#{sdkprefix}/usr/lib', '#{sdkprefix}/usr/include')"
+        s.gsub! "JPEG_ROOT = None", "JPEG_ROOT = ('#{Formula["jpeg"].opt_prefix}/lib', '#{Formula["jpeg"].opt_prefix}/include')"
+        s.gsub! "FREETYPE_ROOT = None", "FREETYPE_ROOT = ('#{Formula["freetype"].opt_prefix}/lib', '#{Formula["freetype"].opt_prefix}/include')"
+      end
+
+      # avoid triggering "helpful" distutils code that doesn't recognize Xcode 7 .tbd stubs
+      ENV.append "CFLAGS", "-I#{MacOS.sdk_path}/System/Library/Frameworks/Tk.framework/Versions/8.5/Headers" unless MacOS::CLT.installed?
+      venv.pip_install Pathname.pwd
+    end
+
+    res = resources.map(&:name).to_set - ["Pillow"]
+
+    res.each do |r|
+      venv.pip_install resource(r)
+    end
+
+    venv.pip_install_and_link buildpath
+  end
+
+  test do
+    # Since we use Python 3, we require a UTF-8 locale
+    ENV["LC_ALL"] = "en_US.UTF-8"
+
+    system "#{bin}/ocrmypdf", "-f", "-q", "--deskew",
+                              test_fixtures("test.pdf"), "ocr.pdf"
+    assert_predicate testpath/"ocr.pdf", :exist?
+  end
+end
+""")
+
+def main():
+    p = run(['poet', '--single', 'ocrmypdf'],
+            encoding='utf-8', stdout=PIPE, check=True)
+
+    ocrmypdf_lines = p.stdout.splitlines()
+    ocrmypdf_url = ocrmypdf_lines[1].strip()
+    ocrmypdf_sha256 = ocrmypdf_lines[2].strip()
+
+    ocrmypdf_version = re.search(
+        r'ocrmypdf-(.+)\.tar.*', ocrmypdf_url).group(1)
+    print(f"Autobrewing {ocrmypdf_version}")
+
+    p = run(['poet', '--resources', 'ocrmypdf'],
+            encoding='utf-8', stdout=PIPE, check=True)
+
+    poet_resources = p.stdout
+
+    # Remove the duplicate "ocrmypdf" resource block
+    all_resources = poet_resources.split('resource')
+    kept_resources = [block for block in all_resources if 'ocrmypdf' not in block]
+    resources = 'resource'.join(kept_resources)
+
+    with open('ocrmypdf.rb', 'w') as out:
+        out.write(recipe_template.substitute(**locals()))
+
+
+if __name__ == '__main__':
+    main()

.travis/linux_before_install.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+# © 2017 James R. Barlow: github.com/jbarlow83
+set -euo pipefail
+set -x
+
+sudo add-apt-repository ppa:vshn/ghostscript -y
+sudo add-apt-repository ppa:heyarje/libav-11 -y
+sudo apt-get update -qq
+sudo apt-get install -y \
+	ghostscript \
+	qpdf \
+	poppler-utils \
+	libavformat56 \
+	libavcodec56 \
+	libavutil54 \
+	libffi-dev
+
+sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
+
+sudo apt-get update
+sudo apt-get autoremove -y
+sudo apt-get install -y --no-install-recommends \
+		tesseract-ocr \
+		tesseract-ocr-eng \
+		tesseract-ocr-fra \
+		tesseract-ocr-deu
+
+pip install --upgrade pip
+mkdir -p packages
+[ -f packages/unpaper_6.1-1.deb ] || 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

.travis/osx_brew.sh

@@ -0,0 +1,23 @@
+#!/bin/bash
+# © 2017 James R. Barlow: github.com/jbarlow83
+set -euo pipefail
+set -x
+
+pip3 install homebrew-pypi-poet
+python3 .travis/autobrew.py
+cat ocrmypdf.rb
+
+# brew audit crashes Travis
+#brew audit ocrmypdf.rb
+
+# Important: disable debug output so token is hidden
+set +x
+git clone https://$HOMEBREW_OCRMYPDF_TOKEN@github.com/jbarlow83/homebrew-ocrmypdf.git
+set -x
+
+pushd homebrew-ocrmypdf
+cp ../ocrmypdf.rb Formula/ocrmypdf.rb
+git add Formula/ocrmypdf.rb
+git commit -m "homebrew-ocrmypdf: automatic release $TRAVIS_BUILD_NUMBER $TRAVIS_TAG"
+git push origin master
+popd

Dockerfile

@@ -1,91 +0,0 @@
-# OCRmyPDF
-#
-# VERSION               3.2
-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
-# If this command takes too Docker hub's automated build will timeout,
-# so try it in portions
-RUN apt-get update && apt-get install -y --no-install-recommends \
-  locales \
-  python3 \
-  python3-pip \
-  python3-venv \
-  python3-reportlab \
-  python3-pil \
-  python3-wheel
-
-RUN apt-get install -y --no-install-recommends \
-  unpaper \
-  qpdf \
-  poppler-utils \
-  tesseract-ocr \
-  tesseract-ocr-deu tesseract-ocr-spa tesseract-ocr-eng tesseract-ocr-fra
-
-RUN apt-get install -qy --no-install-recommends \
-  libffi-dev \
-  libpython3-dev \
-  gcc
-
-# Install Ghostscript from Debian sid to work around JPEG 2000 issue in
-# Debian stretch libgs9 or gs 9.16~dfsg-2.1
-
-COPY ./share/etc-apt-sources.list /etc/apt/sources.list
-
-RUN apt-get update && apt-get install -y ghostscript/sid
-
-
-# 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
-
-
-# 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/
-
-# Replace stock Tesseract 3.04.00 font with improved sharp2.ttf that resolves
-# issues in many PDF viewers.
-# Discussion is in https://github.com/tesseract-ocr/tesseract/issues/182
-COPY ./share/sharp2.ttf /usr/share/tesseract-ocr/tessdata/pdf.ttf
-RUN chmod 644 /usr/share/tesseract-ocr/tessdata/pdf.ttf
-
-# Set this here to force a docker version, allowing non-tagged versions to
-# be built
-# ENV SETUPTOOLS_SCM_PRETEND_VERSION=v3.3.0
-
-# Install application and dependencies
-# In this arrangement Pillow and reportlab will be provided by the system
-# Even though ocrmypdf is locally present, pull from PyPI because
-# Dockerhub and setuptools_scm clash
-RUN . /appenv/bin/activate; \
-  pip install --upgrade pip \
-  && pip install ocrmypdf \
-  && pip install --no-cache-dir -r /application/test_requirements.txt
-
-# Remove the junk
-RUN apt-get remove -qy gcc
-RUN apt-get autoremove -y && apt-get clean -y
-RUN rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* /root/*
-
-USER docker
-WORKDIR /home/docker
-
-ENV OCRMYPDF_TEST_OUTPUT=/tmp/test-output
-ENV OCRMYPDF_SHARP_TTF=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.rst

@@ -1,4 +1,7 @@
-Copyright (c) 2013-2015, The OCRmyPDF Authors
+Copyright (c) 2014-2017, James R. Barlow
+
+Copyright (c) 2013-2014, Julien Pfefferkorn
+Copyright (c) 2013-2017, The OCRmyPDF Authors
 
 Permission is hereby granted, free of charge, to any person obtaining a
 copy of this software and associated documentation files (the
@@ -18,9 +21,3 @@ 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.
-
-----
-
-sRGB_IEC61966-2-1_black_scaled.icc 
-
-To anyone who acknowledges that the file "sRGB_IEC61966-2-1_black scaled.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY, permission to use, copy and distribute these file for any purpose is hereby granted without fee, provided that the file is not changed including the ICC copyright notice tag, and that the name of ICC shall not be used in advertising or publicity pertaining to distribution of the software without specific, written prior permission. ICC makes no representations about the suitability of this software for any purpose.

MANIFEST.in

@@ -1,4 +1,51 @@
-recursive-exclude tests/output *
+# requirements
 include requirements.txt
 include test_requirements.txt
-include share/*
+include dev_requirements.txt
+
+# git
+include .git_archival.txt
+
+# docker
+include Dockerfile*
+include .dockerignore
+
+# tests
+include pytest.ini
+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-exclude tests/output *
+recursive-exclude tests/output_pageinfo *
+recursive-exclude tests/resources/private *
+recursive-exclude tests/cache *
+
+# documentation
+include *.rst
+include pipeline.svg
+recursive-exclude .github * 
+recursive-include docs *.py
+recursive-include docs *.rst
+recursive-include docs *.svg
+recursive-exclude docs/_build *
+
+
+# support files
+recursive-include ocrmypdf/data *
+include *.py
+exclude tasks.py
+exclude debug-runner.py
+recursive-exclude .travis *
+exclude .travis*
+
+
+# code
+recursive-include ocrmypdf *.py
+exclude ocrmypdf/lib/_leptonica.py
+include docker-wrapper.sh
+
+

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

@@ -1,14 +1,36 @@
 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/docker/build/jbarlow83/ocrmypdf.svg
+    :target: https://hub.docker.com/r/jbarlow83/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
+   `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
 -  When possible, inserts OCR information as a "lossless" operation without rendering vector information
@@ -18,11 +40,11 @@ Main features
 -  Provides debug mode to enable easy verification of the OCR results
 -  Processes pages in parallel when more than one CPU core is
    available
--  Uses `Tesseract OCR <https://github.com/tesseract-ocr/tesseract>`__ engine
--  Supports the `39 languages <https://code.google.com/p/tesseract-ocr/downloads/list>`__ recognized by Tesseract
+-  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
 ----------
@@ -31,223 +53,66 @@ 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 
+-  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
-
-You can install it to a Python virtual environment or system-wide. 
-
-Debian and Ubuntu
-~~~~~~~~~~~~~~~~~
+Linux, UNIX, and macOS are supported. Windows is not directly supported but there is a Docker image available that runs on Windows.
 
 Users of Debian 9 or later or Ubuntu 16.10 or later may simply
-``apt-get install ocrmypdf``.
 
-Installing the Docker image
-~~~~~~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: bash
 
-For many users, installing the Docker image 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 image 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 engine host, and then follow these optional steps to enable multiple CPUs::
+   brew install ocrmypdf
 
-   # 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 $(docker-machine env "yourVM")
-
-Assuming you have a Docker engine running somewhere, you can run these commands to download
-the image::
-
-   docker pull jbarlow83/ocrmypdf
-
-Then tag it to give a more convenient name, just ocrmypdf::
-
-   docker tag jbarlow83/ocrmypdf ocrmypdf
-
-This image contains language packs for English, French, Spanish and German. The alternative "polyglot" image provides `all available language packs <https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#languages>`__::
-
-   # Alternative step: If you need all language packs
-   docker pull jbarlow83/ocrmypdf-polyglot
-   docker tag jbarlow83/ocrmypdf-polyglot ocrmypdf
-
-You can then run ocrmypdf using the command::
-
-   docker run ocrmypdf --help
-  
-To execute the OCRmyPDF on a local file, you must `provide a writable volume to the Docker image <https://docs.docker.com/userguide/dockervolumes/>`__, such as this in this template::
-
-   docker run -v "$(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``:: 
-
-   docker run -v "$(pwd):/home/docker"   ocrmypdf --skip-text test.pdf output.pdf
-
-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.
-
-
-Installing on Mac OS X
-~~~~~~~~~~~~~~~~~~~~~~
-
-These instructions probably work on all Mac OS X versions later than 10.7 (Lion). OCRmyPDF is known to work on Yosemite and El Capitan, and regularly tested on El Capitan.
-
-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 libffi leptonica
-   brew install unpaper    # optional
-   brew install tesseract
-   
-Update the homebrew pip and install Pillow::
-
-   pip3 install --upgrade pip
-   pip3 install --upgrade pillow
-
-You can then install OCRmyPDF from PyPI::
-
-   pip3 install ocrmypdf
-
-The command line program should now be available::
-
-   ocrmypdf --help
-
-Installing on Ubuntu 14.04 LTS
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-Installing on Ubuntu 14.04 LTS (trusty) is more difficult than other options, because of certain bugs in Python package installation.
-
-Update apt-get::
-
-   sudo apt-get update
-   sudo apt-get upgrade
-   
-Install system dependencies::
-
-   sudo apt-get install \
-      zlib1g-dev \
-      libjpeg-dev \
-      libffi-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 on Windows
-~~~~~~~~~~~~~~~~~~~~~
-
-Direct installation on Windows is not possible.  Install the Docker container as described above.
-
-      
-Installing HEAD revision from sources
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-If you have ``git`` and ``python3.4`` or ``python3.5`` 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::
-
-   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::
-
-   pip3 install -e git+https://github.com/jbarlow83/OCRmyPDF.git
-   
-On certain Linux distributions such as Ubuntu, you may need to use 
-run the install command as superuser::
-
-   sudo pip3 install [-e] git+https://github.com/jbarlow83/OCRmyPDF.git
-   
-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 .
-
-However, ``ocrmypdf`` will only be accessible on the system PATH after
-you activate the virtual environment.
-
-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.
+For everyone else, `see our documentation <https://ocrmypdf.readthedocs.io/en/latest/installation.html>`_ for installation steps.
 
 Languages
 ---------
 
 OCRmyPDF uses Tesseract for OCR, and relies on its language packs. For Linux users,
-you can often find packages that provide language packs::
+you can often find packages that provide language packs:
+
+.. code-block:: bash
+
+   # Display a list of all Tesseract language packs
+   apt-cache search tesseract-ocr
 
    # Debian/Ubuntu users
-   sudo apt-get install tesseract-ocr-chi-sim
+   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.
 
-Support
--------
+Documentation and support
+-------------------------
 
-In case you detect an issue, please:
+Once ocrmypdf is installed, the built-in help which explains the command syntax and options can be accessed via:
 
--  Check if your issue is already known
+.. code-block:: bash
+
+   ocrmypdf --help
+
+Our `documentation is served on Read the Docs <https://ocrmypdf.readthedocs.io/en/latest/index.html>`_.
+
+If you detect an issue, please:
+
+-  Check whether your issue is already known
 -  If no problem report exists on github, please create one here:
    https://github.com/jbarlow83/OCRmyPDF/issues
 -  Describe your problem thoroughly
@@ -256,14 +121,21 @@ In case you detect an issue, please:
 -  If possible provide your input PDF file as well as the content of the
    temporary folder (using a file sharing service like Dropbox)
 
+Requirements
+------------
+
+Runs on CPython 3.6, and requires external program installations of Ghostscript, Tesseract OCR, QPDF, and Leptonica. ocrmypdf is pure Python, but uses CFFI to portably generate library bindings.
+
+Python 3.5 is fully supported, but support will end at the end of 2017, to allow f-strings in the code base.
+
 Press & Media
 -------------
 
--  `c't 1-2014, page 59 <http://heise.de/-2279695>`__:
+-  `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://heise.de/-2356670>`__
+   OCRmyPDF <http://heise.de/-2356670>`_
 
 Disclaimer
 ----------

RELEASE_NOTES.rst

@@ -1,691 +0,0 @@
-RELEASE NOTES
-=============
-
-Please always read this file before installing the package
-
-Download software here: https://github.com/jbarlow83/OCRmyPDF/tags
-
-
-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:
-=======
-
-Fixes
------
-
--  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 unexpect 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) 
--  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)
-
-
-Changes
--------
-
--  Logging output is now much easier to read
--  ``--deskew`` is now performed by Leptonica instead of unpaper (#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 "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 OS X 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 - 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: 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 +1,10 @@
+check-manifest >= 0.35
+twine >= 1.8.1
+coverage >= 4.4
+pytest-xdist >= 1.16.0
+GitPython == 2.1.3
+
+# Known good versions: 1.11.1
+# Known broken versions: 1.15.0
+setuptools-scm == 1.11.1
+setuptools-scm-git-archive == 1.0

docs/advanced.rst

@@ -0,0 +1,143 @@
+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 only 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 environment variable ``OCRMYPDF_TESSERACT`` for the full path *to the tesseract binary* first. 
+
+.. envvar:: OCRMYPDF_TESSERACT
+
+	Specify the location of the Tesseract binary.
+
+.. envvar:: TESSDATA_PREFIX
+
+	A Tesseract environment variable that overrides the path to Tesseract's data files.
+
+For example, if you are testing tesseract 4.00 and don't wish to disturb use an existing tesseract 3.04 installation, you can launch OCRmyPDF as follows:
+
+.. code-block:: bash
+
+	env \
+		OCRMYPDF_TESSERACT=/home/user/src/tesseract4/api/tesseract \
+		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 check the environment variable ``OCRMYPDF_{program}`` before asking the system to find ``{program}`` on the PATH. For example, you could redirect OCRmyPDF to ``OCRMYPDF_GS`` to override Ghostscript. The full list is below:
+
+.. envvar:: OCRMYPDF_GS
+
+	Specify the location of the Ghostscript binary.
+
+.. envvar:: OCRMYPDF_UNPAPER
+
+	Specify the location of the unpaper binary.
+
+.. envvar:: OCRMYPDF_QPDF
+
+	Specify the location of the qpdf binary.
+
+
+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, and newer, ``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.
+
+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.
+
+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 and 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 +1,212 @@
+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
+
+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 = '2017, James R. Barlow'
+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 +1,191 @@
+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.
+
+
+OCR languages other than English
+""""""""""""""""""""""""""""""""
+
+By default OCRmyPDF assumes the document is English. 
+
+.. code-block:: bash
+
+    ocrmypdf -l fre LeParisien.pdf LeParisien.pdf
+    ocrmypdf -l eng+fre 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
+--------------------
+
+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:
+
+.. code-block:: bash
+
+    img2pdf my-images*.jpg | ocrmypdf - myfile.pdf
+
+``img2pdf`` also has features to control the position of images on a page, if desired.
+
+For convenience, OCRmyPDF can 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.
+
+You can also use Tesseract 3.04+ directly to convert single page images or multi-page TIFFs to PDF:
+
+.. code-block:: bash
+
+    tesseract my-image.jpg output-prefix pdf
+
+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 +1,376 @@
+Installation
+============
+
+OCRmyPDF requires Python 3.5 (or newer) and Tesseract 3.04 (or newer).
+
+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
+
+Installing on macOS
+-------------------
+
+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:
+
++-----------------------------+---------------------------------------------+---------------------------------------------------------------------------------+
+| Image name                  | Download command                            | Notes                                                                           |
++-----------------------------+---------------------------------------------+---------------------------------------------------------------------------------+
+| ocrmypdf                    | ``docker pull jbarlow83/ocrmypdf``          | Latest ocrmypdf with Tesseract 3.04. Includes English, French, German, Spanish. |
++-----------------------------+---------------------------------------------+---------------------------------------------------------------------------------+
+| ocrmypdf-polyglot           | ``docker pull jbarlow83/ocrmypdf-polyglot`` | As above, with all available language packs.                                    |
++-----------------------------+---------------------------------------------+---------------------------------------------------------------------------------+
+| ocrmypdf-tess4              | ``docker pull jbarlow83/ocrmypdf-tess4``    | Latest ocrmypdf with Tesseract 4.00.00alpha and English, French, German,        |
+|                             |                                             | Spanish, Portuguese, Chinese Simplified, Arabic and Russian (the top 8).        |
++-----------------------------+---------------------------------------------+---------------------------------------------------------------------------------+
+
+For example:
+
+.. code-block:: bash
+
+    docker pull jbarlow83/ocrmypdf-tess4
+
+Then tag it to give a more convenient name, just ocrmypdf:
+
+.. code-block:: bash
+
+    docker tag jbarlow83/ocrmypdf-tess4 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 and 3.6 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 and install Pillow:
+
+.. code-block:: bash
+
+    pip3 install --upgrade pip
+    pip3 install --upgrade pillow
+
+You can then install OCRmyPDF from PyPI, for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf
+
+or system-wide:
+
+.. code-block:: bash
+
+    pip3 install ocrmypdf
+
+The command line program should now be available:
+
+.. code-block:: bash
+
+    ocrmypdf --help
+
+
+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
+
+Alternately, system-wide. Note that this may modify the system Python environment:
+
+.. code-block:: bash
+
+    sudo pip3 install ocrmypdf
+
+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
+
+
+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 \
+        ghostscript \
+        tesseract-ocr \
+        tesseract-ocr-eng \
+        libavformat56 libavcodec56 libavutil54 \
+        wget
+
+Now we need to install ``pip`` and let it install ocrmypdf:
+
+.. code-block:: bash
+
+    wget -O - -o /dev/null https://bootstrap.pypa.io/get-pip.py | python3.6
+    pip3.6 install ocrmypdf
+
+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.
+
+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 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
+
+On certain Linux distributions such as Ubuntu, you may need to use 
+run the install command as superuser:
+
+.. code-block:: bash
+
+    sudo pip3 install [-e] git+https://github.com/jbarlow83/OCRmyPDF.git
+
+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:
+
+.. 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.

docs/introduction.rst

@@ -0,0 +1,107 @@
+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.
+  
+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.
+
+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
+
+If you are looking for a micro web-frontend for OCRmyPDF, consider the third-party `OCRmyPDF-web <https://github.com/sseemayer/OCRmyPDF-web>`_. 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 +1,392 @@
+<?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="1544pt" height="1089pt"
+ viewBox="0.00 0.00 1544.00 1089.14" 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 1085.14)">
+<title>Pipeline:</title>
+<polygon fill="white" stroke="none" points="-4,4 -4,-1085.14 1540,-1085.14 1540,4 -4,4"/>
+<g id="clust1" class="cluster"><title>clustertasks</title>
+<polygon fill="none" stroke="black" points="8,-8 8,-1073.14 1528,-1073.14 1528,-8 8,-8"/>
+<text text-anchor="middle" x="768" y="-1045.14" 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="black" points="750.705,-1027.14 541.295,-1027.14 537.295,-1023.14 537.295,-991.141 746.705,-991.141 750.705,-995.141 750.705,-1027.14"/>
+<polyline fill="none" stroke="black" points="746.705,-1023.14 537.295,-1023.14 "/>
+<polyline fill="none" stroke="black" points="746.705,-1023.14 746.705,-991.141 "/>
+<polyline fill="none" stroke="black" points="746.705,-1023.14 750.705,-1027.14 "/>
+<text text-anchor="middle" x="644" y="-1003.14" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.triage</text>
+</g>
+<!-- t1 -->
+<g id="node2" class="node"><title>t1</title>
+<polygon fill="#efa03b" stroke="black" points="769.969,-969.141 522.031,-969.141 518.031,-965.141 518.031,-933.141 765.969,-933.141 769.969,-937.141 769.969,-969.141"/>
+<polyline fill="none" stroke="black" points="765.969,-965.141 518.031,-965.141 "/>
+<polyline fill="none" stroke="black" points="765.969,-965.141 765.969,-933.141 "/>
+<polyline fill="none" stroke="black" points="765.969,-965.141 769.969,-969.141 "/>
+<text text-anchor="middle" x="644" y="-945.141" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.repair_pdf</text>
+</g>
+<!-- t0&#45;&gt;t1 -->
+<g id="edge1" class="edge"><title>t0&#45;&gt;t1</title>
+<path fill="none" stroke="#0044a0" d="M644,-991.035C644,-987.344 644,-983.354 644,-979.384"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="647.5,-979.142 644,-969.142 640.5,-979.142 647.5,-979.142"/>
+</g>
+<!-- t2 -->
+<g id="node3" class="node"><title>t2</title>
+<polygon fill="#efa03b" stroke="black" points="869.82,-890.215 644,-907.116 418.18,-890.215 418.391,-862.869 869.609,-862.869 869.82,-890.215"/>
+<polygon fill="none" stroke="black" points="873.852,-893.922 644,-911.125 414.148,-893.922 414.419,-858.871 873.581,-858.871 873.852,-893.922"/>
+<text text-anchor="middle" x="644" y="-876.657" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.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="M644,-933.019C644,-929.4 644,-925.445 644,-921.399"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="647.5,-921.295 644,-911.295 640.5,-921.295 647.5,-921.295"/>
+</g>
+<!-- t18 -->
+<g id="node20" class="node"><title>t18</title>
+<polygon fill="#efa03b" stroke="black" points="383.516,-832.173 20.4844,-832.173 16.4844,-828.173 16.4844,-796.173 379.516,-796.173 383.516,-800.173 383.516,-832.173"/>
+<polyline fill="none" stroke="black" points="379.516,-828.173 16.4844,-828.173 "/>
+<polyline fill="none" stroke="black" points="379.516,-828.173 379.516,-796.173 "/>
+<polyline fill="none" stroke="black" points="379.516,-828.173 383.516,-832.173 "/>
+<text text-anchor="middle" x="200" y="-808.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.generate_postscript_stub</text>
+</g>
+<!-- t1&#45;&gt;t18 -->
+<g id="edge29" class="edge"><title>t1&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M518.14,-936.259C481.198,-930.408 441.05,-922.284 405,-911.141 344.394,-892.407 278.709,-859.14 238.448,-837.118"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="239.963,-833.956 229.517,-832.188 236.581,-840.085 239.963,-833.956"/>
+</g>
+<!-- t21 -->
+<g id="node22" class="node"><title>t21</title>
+<polygon fill="#efa03b" stroke="black" points="751.49,-156.424 963,-129.109 1174.51,-156.424 1174.31,-200.62 751.688,-200.62 751.49,-156.424"/>
+<polygon fill="none" stroke="black" points="747.474,-152.906 963,-125.073 1178.53,-152.906 1178.29,-204.623 747.705,-204.623 747.474,-152.906"/>
+<text text-anchor="middle" x="963" y="-162.639" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.merge_pages_qpdf</text>
+</g>
+<!-- t1&#45;&gt;t21 -->
+<g id="edge39" class="edge"><title>t1&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M769.832,-943.539C1011.96,-929.143 1519,-890.056 1519,-815.173 1519,-815.173 1519,-815.173 1519,-309.173 1519,-237.474 1342.87,-202.283 1188.87,-185.216"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1189.08,-181.719 1178.76,-184.118 1188.32,-188.678 1189.08,-181.719"/>
+</g>
+<!-- t3 -->
+<g id="node4" class="node"><title>t3</title>
+<polygon fill="#efa03b" stroke="black" points="712.317,-832.173 405.683,-832.173 401.683,-828.173 401.683,-796.173 708.317,-796.173 712.317,-800.173 712.317,-832.173"/>
+<polyline fill="none" stroke="black" points="708.317,-828.173 401.683,-828.173 "/>
+<polyline fill="none" stroke="black" points="708.317,-828.173 708.317,-796.173 "/>
+<polyline fill="none" stroke="black" points="708.317,-828.173 712.317,-832.173 "/>
+<text text-anchor="middle" x="557" y="-808.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.rasterize_preview</text>
+</g>
+<!-- t2&#45;&gt;t3 -->
+<g id="edge3" class="edge"><title>t2&#45;&gt;t3</title>
+<path fill="none" stroke="#0044a0" d="M614.06,-858.777C605.418,-852.173 595.983,-844.963 587.319,-838.342"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="589.37,-835.504 579.299,-832.213 585.119,-841.066 589.37,-835.504"/>
+</g>
+<!-- t4 -->
+<g id="node5" class="node"><title>t4</title>
+<polygon fill="#efa03b" stroke="black" points="688.082,-774.173 429.918,-774.173 425.918,-770.173 425.918,-738.173 684.082,-738.173 688.082,-742.173 688.082,-774.173"/>
+<polyline fill="none" stroke="black" points="684.082,-770.173 425.918,-770.173 "/>
+<polyline fill="none" stroke="black" points="684.082,-770.173 684.082,-738.173 "/>
+<polyline fill="none" stroke="black" points="684.082,-770.173 688.082,-774.173 "/>
+<text text-anchor="middle" x="557" y="-750.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.orient_page</text>
+</g>
+<!-- t2&#45;&gt;t4 -->
+<g id="edge5" class="edge"><title>t2&#45;&gt;t4</title>
+<path fill="none" stroke="#0044a0" d="M695.691,-858.661C705.833,-851.667 715.136,-842.912 721,-832.173 728.668,-818.13 730.919,-808.727 721,-796.173 714.915,-788.471 700.759,-782.127 683.18,-776.959"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="683.659,-773.462 673.091,-774.208 681.817,-780.216 683.659,-773.462"/>
+</g>
+<!-- t3&#45;&gt;t4 -->
+<g id="edge4" class="edge"><title>t3&#45;&gt;t4</title>
+<path fill="none" stroke="#0044a0" d="M557,-796.066C557,-792.375 557,-788.385 557,-784.415"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="560.5,-784.174 557,-774.174 553.5,-784.174 560.5,-784.174"/>
+</g>
+<!-- t5 -->
+<g id="node6" class="node"><title>t5</title>
+<polygon fill="#efa03b" stroke="black" points="1064.73,-716.173 689.268,-716.173 685.268,-712.173 685.268,-680.173 1060.73,-680.173 1064.73,-684.173 1064.73,-716.173"/>
+<polyline fill="none" stroke="black" points="1060.73,-712.173 685.268,-712.173 "/>
+<polyline fill="none" stroke="black" points="1060.73,-712.173 1060.73,-680.173 "/>
+<polyline fill="none" stroke="black" points="1060.73,-712.173 1064.73,-716.173 "/>
+<text text-anchor="middle" x="875" y="-692.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
+</g>
+<!-- t4&#45;&gt;t5 -->
+<g id="edge6" class="edge"><title>t4&#45;&gt;t5</title>
+<path fill="none" stroke="#0044a0" d="M653.706,-738.143C689.783,-731.789 730.943,-724.541 767.902,-718.033"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="768.965,-721.399 778.206,-716.218 767.751,-714.505 768.965,-721.399"/>
+</g>
+<!-- t15 -->
+<g id="node13" class="node"><title>t15</title>
+<polygon fill="#ff69b4" stroke="black" points="440.27,-396.173 59.7304,-396.173 55.7304,-392.173 55.7304,-360.173 436.27,-360.173 440.27,-364.173 440.27,-396.173"/>
+<polyline fill="none" stroke="black" points="436.27,-392.173 55.7304,-392.173 "/>
+<polyline fill="none" stroke="black" points="436.27,-392.173 436.27,-360.173 "/>
+<polyline fill="none" stroke="black" points="436.27,-392.173 440.27,-396.173 "/>
+<text text-anchor="middle" x="248" y="-372.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
+</g>
+<!-- t4&#45;&gt;t15 -->
+<g id="edge14" class="edge"><title>t4&#45;&gt;t15</title>
+<path fill="none" stroke="#0044a0" d="M425.857,-745.263C361.241,-732.425 297,-703.891 297,-641.173 297,-641.173 297,-641.173 297,-523.173 297,-480.18 277.182,-433.63 262.754,-405.436"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="265.677,-403.477 257.916,-396.265 259.486,-406.743 265.677,-403.477"/>
+</g>
+<!-- t12 -->
+<g id="node15" class="node"><title>t12</title>
+<polygon fill="#00cc66" stroke="black" points="781,-406.173 458.414,-378.173 781,-350.173 1103.59,-378.173 781,-406.173"/>
+<text text-anchor="middle" x="781" y="-372.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.select_image_layer</text>
+</g>
+<!-- t4&#45;&gt;t12 -->
+<g id="edge20" class="edge"><title>t4&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M471.449,-738.167C421.018,-722.605 367,-693.681 367,-641.173 367,-641.173 367,-641.173 367,-523.173 367,-480.717 347.218,-458.43 377,-428.173 392.917,-412.001 494.592,-399.721 591.948,-391.499"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="592.284,-394.983 601.96,-390.666 591.704,-388.007 592.284,-394.983"/>
+</g>
+<!-- t19 -->
+<g id="node18" class="node"><title>t19</title>
+<polygon fill="#efa03b" stroke="black" points="323.772,-270.173 78.2284,-270.173 74.2284,-266.173 74.2284,-234.173 319.772,-234.173 323.772,-238.173 323.772,-270.173"/>
+<polyline fill="none" stroke="black" points="319.772,-266.173 74.2284,-266.173 "/>
+<polyline fill="none" stroke="black" points="319.772,-266.173 319.772,-234.173 "/>
+<polyline fill="none" stroke="black" points="319.772,-266.173 323.772,-270.173 "/>
+<text text-anchor="middle" x="199" y="-246.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.skip_page</text>
+</g>
+<!-- t4&#45;&gt;t19 -->
+<g id="edge26" class="edge"><title>t4&#45;&gt;t19</title>
+<path fill="none" stroke="#0044a0" d="M425.782,-752.078C271.736,-744.795 37,-720.655 37,-641.173 37,-641.173 37,-641.173 37,-377.173 37,-325.92 88.7624,-292.989 133.769,-274.043"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="135.289,-277.203 143.249,-270.211 132.666,-270.713 135.289,-277.203"/>
+</g>
+<!-- t17 -->
+<g id="node19" class="node"><title>t17</title>
+<polygon fill="#66ccff" stroke="black" points="1499.67,-328.173 1094.33,-328.173 1090.33,-324.173 1090.33,-292.173 1495.67,-292.173 1499.67,-296.173 1499.67,-328.173"/>
+<polyline fill="none" stroke="black" points="1495.67,-324.173 1090.33,-324.173 "/>
+<polyline fill="none" stroke="black" points="1495.67,-324.173 1495.67,-292.173 "/>
+<polyline fill="none" stroke="black" points="1495.67,-324.173 1499.67,-328.173 "/>
+<text text-anchor="middle" x="1295" y="-304.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.ocr_tesseract_and_render_pdf</text>
+</g>
+<!-- t4&#45;&gt;t17 -->
+<g id="edge28" class="edge"><title>t4&#45;&gt;t17</title>
+<path fill="none" stroke="#0044a0" d="M688.477,-752.611C947.521,-745.804 1499,-721.841 1499,-641.173 1499,-641.173 1499,-641.173 1499,-455.173 1499,-407.201 1508.25,-383.766 1474,-350.173 1465.98,-342.302 1448.74,-335.804 1427.99,-330.522"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1428.77,-327.11 1418.23,-328.18 1427.14,-333.917 1428.77,-327.11"/>
+</g>
+<!-- t6 -->
+<g id="node7" class="node"><title>t6</title>
+<polygon fill="#efa03b" stroke="black" points="1089.08,-658.173 664.92,-658.173 660.92,-654.173 660.92,-622.173 1085.08,-622.173 1089.08,-626.173 1089.08,-658.173"/>
+<polyline fill="none" stroke="black" points="1085.08,-654.173 660.92,-654.173 "/>
+<polyline fill="none" stroke="black" points="1085.08,-654.173 1085.08,-622.173 "/>
+<polyline fill="none" stroke="black" points="1085.08,-654.173 1089.08,-658.173 "/>
+<text text-anchor="middle" x="875" y="-634.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.preprocess_remove_background</text>
+</g>
+<!-- t5&#45;&gt;t6 -->
+<g id="edge7" class="edge"><title>t5&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M875,-680.066C875,-676.375 875,-672.385 875,-668.415"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="878.5,-668.174 875,-658.174 871.5,-668.174 878.5,-668.174"/>
+</g>
+<!-- t11 -->
+<g id="node14" class="node"><title>t11</title>
+<polygon fill="#efa03b" stroke="black" points="1096,-484.173 712.187,-456.173 1096,-428.173 1479.81,-456.173 1096,-484.173"/>
+<text text-anchor="middle" x="1096" y="-450.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.select_visible_page_image</text>
+</g>
+<!-- t5&#45;&gt;t11 -->
+<g id="edge18" class="edge"><title>t5&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M1051.07,-680.0px8C1072.76,-674.712 1090.1,-667.576 1098,-658.173 1136.53,-612.317 1122.5,-537.602 1109.06,-493.335"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1112.33,-492.043 1105.96,-483.573 1105.65,-494.164 1112.33,-492.043"/>
+</g>
+<!-- t7 -->
+<g id="node8" class="node"><title>t7</title>
+<polygon fill="#efa03b" stroke="black" points="1039.68,-600.173 720.324,-600.173 716.324,-596.173 716.324,-564.173 1035.68,-564.173 1039.68,-568.173 1039.68,-600.173"/>
+<polyline fill="none" stroke="black" points="1035.68,-596.173 716.324,-596.173 "/>
+<polyline fill="none" stroke="black" points="1035.68,-596.173 1035.68,-564.173 "/>
+<polyline fill="none" stroke="black" points="1035.68,-596.173 1039.68,-600.173 "/>
+<text text-anchor="middle" x="878" y="-576.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.preprocess_deskew</text>
+</g>
+<!-- t6&#45;&gt;t7 -->
+<g id="edge8" class="edge"><title>t6&#45;&gt;t7</title>
+<path fill="none" stroke="#0044a0" d="M875.916,-622.066C876.114,-618.375 876.328,-614.385 876.541,-610.415"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="880.049,-610.347 877.089,-600.174 873.059,-609.972 880.049,-610.347"/>
+</g>
+<!-- t6&#45;&gt;t11 -->
+<g id="edge17" class="edge"><title>t6&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M990.695,-622.152C1016.54,-616.369 1039.2,-609.066 1049,-600.173 1078.78,-573.156 1089.74,-526.964 1093.74,-494.403"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1097.24,-494.666 1094.82,-484.351 1090.28,-493.924 1097.24,-494.666"/>
+</g>
+<!-- t8 -->
+<g id="node9" class="node"><title>t8</title>
+<polygon fill="#efa03b" stroke="black" points="1030.88,-542.173 729.117,-542.173 725.117,-538.173 725.117,-506.173 1026.88,-506.173 1030.88,-510.173 1030.88,-542.173"/>
+<polyline fill="none" stroke="black" points="1026.88,-538.173 725.117,-538.173 "/>
+<polyline fill="none" stroke="black" points="1026.88,-538.173 1026.88,-506.173 "/>
+<polyline fill="none" stroke="black" points="1026.88,-538.173 1030.88,-542.173 "/>
+<text text-anchor="middle" x="878" y="-518.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.preprocess_clean</text>
+</g>
+<!-- t7&#45;&gt;t8 -->
+<g id="edge9" class="edge"><title>t7&#45;&gt;t8</title>
+<path fill="none" stroke="#0044a0" d="M878,-564.066C878,-560.375 878,-556.385 878,-552.415"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="881.5,-552.174 878,-542.174 874.5,-552.174 881.5,-552.174"/>
+</g>
+<!-- t7&#45;&gt;t11 -->
+<g id="edge16" class="edge"><title>t7&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M995.441,-564.147C1011.26,-558.713 1026.62,-551.566 1040,-542.173 1057.45,-529.919 1070.89,-510.334 1080.22,-493.051"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1083.55,-494.243 1084.99,-483.747 1077.31,-491.053 1083.55,-494.243"/>
+</g>
+<!-- t9 -->
+<g id="node10" class="node"><title>t9</title>
+<polygon fill="#efa03b" stroke="black" points="694.109,-474.173 389.891,-474.173 385.891,-470.173 385.891,-438.173 690.109,-438.173 694.109,-442.173 694.109,-474.173"/>
+<polyline fill="none" stroke="black" points="690.109,-470.173 385.891,-470.173 "/>
+<polyline fill="none" stroke="black" points="690.109,-470.173 690.109,-438.173 "/>
+<polyline fill="none" stroke="black" points="690.109,-470.173 694.109,-474.173 "/>
+<text text-anchor="middle" x="540" y="-450.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.select_ocr_image</text>
+</g>
+<!-- t8&#45;&gt;t9 -->
+<g id="edge10" class="edge"><title>t8&#45;&gt;t9</title>
+<path fill="none" stroke="#0044a0" d="M790.517,-506.09C743.487,-496.907 685.383,-485.561 636.976,-476.109"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="637.602,-472.665 627.116,-474.183 636.26,-479.535 637.602,-472.665"/>
+</g>
+<!-- t8&#45;&gt;t11 -->
+<g id="edge15" class="edge"><title>t8&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M934.424,-506.09C958.693,-498.742 987.536,-490.01 1014.09,-481.971"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1015.2,-485.293 1023.75,-479.045 1013.17,-478.593 1015.2,-485.293"/>
+</g>
+<!-- t10 -->
+<g id="node11" class="node"><title>t10</title>
+<polygon fill="#00cc66" stroke="black" points="1440.1,-396.173 1125.9,-396.173 1121.9,-392.173 1121.9,-360.173 1436.1,-360.173 1440.1,-364.173 1440.1,-396.173"/>
+<polyline fill="none" stroke="black" points="1436.1,-392.173 1121.9,-392.173 "/>
+<polyline fill="none" stroke="black" points="1436.1,-392.173 1436.1,-360.173 "/>
+<polyline fill="none" stroke="black" points="1436.1,-392.173 1440.1,-396.173 "/>
+<text text-anchor="middle" x="1281" y="-372.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
+</g>
+<!-- t9&#45;&gt;t10 -->
+<g id="edge11" class="edge"><title>t9&#45;&gt;t10</title>
+<path fill="none" stroke="#0044a0" d="M631.992,-438.111C655.104,-434.327 679.903,-430.697 703,-428.173 884.404,-408.343 931.549,-425.567 1113,-406.173 1133.45,-403.987 1155.2,-400.994 1176.03,-397.794"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1176.83,-401.211 1186.17,-396.209 1175.75,-394.295 1176.83,-401.211"/>
+</g>
+<!-- t9&#45;&gt;t15 -->
+<g id="edge13" class="edge"><title>t9&#45;&gt;t15</title>
+<path fill="none" stroke="#0044a0" d="M474.47,-438.117C429.651,-426.452 370.0px1,-410.957 323.342,-398.782"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="324.123,-395.369 313.564,-396.237 322.36,-402.143 324.123,-395.369"/>
+</g>
+<!-- t13 -->
+<g id="node12" class="node"><title>t13</title>
+<polygon fill="#00cc66" stroke="black" points="683.043,-328.173 374.957,-328.173 370.957,-324.173 370.957,-292.173 679.043,-292.173 683.043,-296.173 683.043,-328.173"/>
+<polyline fill="none" stroke="black" points="679.043,-324.173 370.957,-324.173 "/>
+<polyline fill="none" stroke="black" points="679.043,-324.173 679.043,-292.173 "/>
+<polyline fill="none" stroke="black" points="679.043,-324.173 683.043,-328.173 "/>
+<text text-anchor="middle" x="527" y="-304.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.render_hocr_page</text>
+</g>
+<!-- t10&#45;&gt;t13 -->
+<g id="edge12" class="edge"><title>t10&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M1186.17,-360.113C1162.35,-356.33 1136.79,-352.699 1113,-350.173 929.956,-330.738 880.531,-342.246 693.485,-328.255"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="693.548,-324.75 683.311,-327.48 693.016,-331.73 693.548,-324.75"/>
+</g>
+<!-- t14 -->
+<g id="node17" class="node"><title>t14</title>
+<polygon fill="#00cc66" stroke="black" points="1072.92,-328.173 705.08,-328.173 701.08,-324.173 701.08,-292.173 1068.92,-292.173 1072.92,-296.173 1072.92,-328.173"/>
+<polyline fill="none" stroke="black" points="1068.92,-324.173 701.08,-324.173 "/>
+<polyline fill="none" stroke="black" points="1068.92,-324.173 1068.92,-292.173 "/>
+<polyline fill="none" stroke="black" points="1068.92,-324.173 1072.92,-328.173 "/>
+<text text-anchor="middle" x="887" y="-304.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.render_hocr_debug_page</text>
+</g>
+<!-- t10&#45;&gt;t14 -->
+<g id="edge25" class="edge"><title>t10&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M1179.28,-360.133C1124,-350.873 1055.54,-339.405 998.77,-329.896"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="999.118,-326.405 988.677,-328.205 997.962,-333.309 999.118,-326.405"/>
+</g>
+<!-- t16 -->
+<g id="node16" class="node"><title>t16</title>
+<polygon fill="#00cc66" stroke="black" points="636.299,-270.173 345.701,-270.173 341.701,-266.173 341.701,-234.173 632.299,-234.173 636.299,-238.173 636.299,-270.173"/>
+<polyline fill="none" stroke="black" points="632.299,-266.173 341.701,-266.173 "/>
+<polyline fill="none" stroke="black" points="632.299,-266.173 632.299,-234.173 "/>
+<polyline fill="none" stroke="black" points="632.299,-266.173 636.299,-270.173 "/>
+<text text-anchor="middle" x="489" y="-246.173" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.combine_layers</text>
+</g>
+<!-- t13&#45;&gt;t16 -->
+<g id="edge22" class="edge"><title>t13&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M515.392,-292.066C512.519,-287.833 509.379,-283.205 506.299,-278.666"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="509.048,-276.483 500.537,-270.174 503.256,-280.414 509.048,-276.483"/>
+</g>
+<!-- t15&#45;&gt;t16 -->
+<g id="edge21" class="edge"><title>t15&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M263.466,-359.936C281.645,-340.695 313.598,-309.935 347,-292.173 361.3,-284.568 377.219,-278.339 393.079,-273.27"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="394.389,-276.53 402.932,-270.262 392.345,-269.835 394.389,-276.53"/>
+</g>
+<!-- t11&#45;&gt;t12 -->
+<g id="edge19" class="edge"><title>t11&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M1009.92,-434.405C967.591,-424.191 916.524,-411.87 873.373,-401.459"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="873.983,-398.006 863.441,-399.063 872.341,-404.811 873.983,-398.006"/>
+</g>
+<!-- t11&#45;&gt;t14 -->
+<g id="edge24" class="edge"><title>t11&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M1334.06,-445.464C1386.49,-438.209 1431.63,-426.12 1449,-406.173 1465.34,-387.403 1466.07,-368.289 1449,-350.173 1422.45,-321.987 1153.75,-330.0px 1083.05,-328.037"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1083,-324.531 1072.85,-327.564 1082.68,-331.524 1083,-324.531"/>
+</g>
+<!-- t11&#45;&gt;t17 -->
+<g id="edge27" class="edge"><title>t11&#45;&gt;t17</title>
+<path fill="none" stroke="#0044a0" d="M1278.64,-441.463C1364.44,-433.122 1450.86,-421.174 1464,-406.173 1480.4,-387.448 1479.49,-369.65 1464,-350.173 1457.83,-342.418 1443.67,-336.074 1426.02,-330.926"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1426.46,-327.418 1415.89,-328.188 1424.63,-334.176 1426.46,-327.418"/>
+</g>
+<!-- t12&#45;&gt;t16 -->
+<g id="edge23" class="edge"><title>t12&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M624.426,-363.718C509.174,-353.04 370.992,-338.277 362,-328.173 351.364,-316.22 352.491,-305.04 362,-292.173 366.819,-285.651 372.71,-280.186 379.287,-275.609"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="381.274,-278.496 387.956,-270.274 377.605,-272.534 381.274,-278.496"/>
+</g>
+<!-- t20 -->
+<g id="node21" class="node"><title>t20</title>
+<polygon fill="#efa03b" stroke="black" points="234.918,-156.426 480,-129.117 725.082,-156.426 724.853,-200.614 235.147,-200.614 234.918,-156.426"/>
+<polygon fill="none" stroke="black" points="230.903,-152.847 480,-125.09 729.097,-152.847 728.829,-204.616 231.171,-204.616 230.903,-152.847"/>
+<text text-anchor="middle" x="480" y="-162.639" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.merge_pages_ghostscript</text>
+</g>
+<!-- t16&#45;&gt;t20 -->
+<g id="edge33" class="edge"><title>t16&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M487.091,-233.881C486.46,-228.165 485.726,-221.515 484.974,-214.703"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="488.446,-214.253 483.87,-204.697 481.488,-215.021 488.446,-214.253"/>
+</g>
+<!-- t16&#45;&gt;t21 -->
+<g id="edge37" class="edge"><title>t16&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M602.543,-234.159C645.082,-227.582 693.805,-219.809 738,-212.173 748.572,-210.346 759.416,-208.431 770.357,-206.468"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="771.206,-209.872 780.427,-204.653 769.964,-202.983 771.206,-209.872"/>
+</g>
+<!-- t14&#45;&gt;t20 -->
+<g id="edge34" class="edge"><title>t14&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M837.094,-292.063C776.411,-271.259 672.165,-235.52 591.922,-208.01"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="592.855,-204.63 582.26,-204.697 590.584,-211.251 592.855,-204.63"/>
+</g>
+<!-- t14&#45;&gt;t21 -->
+<g id="edge38" class="edge"><title>t14&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M896.319,-292.063C906.85,-272.729 924.406,-240.497 938.861,-213.958"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="942.118,-215.294 943.828,-204.838 935.971,-211.946 942.118,-215.294"/>
+</g>
+<!-- t19&#45;&gt;t20 -->
+<g id="edge30" class="edge"><title>t19&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M257.907,-234.081C284.597,-226.336 317.298,-216.848 349.316,-207.558"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="350.595,-210.831 359.224,-204.683 348.645,-204.108 350.595,-210.831"/>
+</g>
+<!-- t19&#45;&gt;t21 -->
+<g id="edge35" class="edge"><title>t19&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M323.792,-235.042C326.891,-234.738 329.964,-234.447 333,-234.173 512.532,-217.929 559.043,-233.851 738,-212.173 751.869,-210.493 766.169,-208.472 780.522,-206.237"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="781.136,-209.683 790.462,-204.657 780.037,-202.77 781.136,-209.683"/>
+</g>
+<!-- t17&#45;&gt;t20 -->
+<g id="edge32" class="edge"><title>t17&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M1195.47,-292.133C1071.63,-270.93 856.728,-234.138 694.688,-206.395"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="695.028,-202.903 684.581,-204.665 693.846,-209.802 695.028,-202.903"/>
+</g>
+<!-- t17&#45;&gt;t21 -->
+<g id="edge36" class="edge"><title>t17&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M1254.29,-292.063C1205.19,-271.428 1121.13,-236.1 1055.9,-208.683"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1056.99,-205.345 1046.42,-204.697 1054.28,-211.799 1056.99,-205.345"/>
+</g>
+<!-- t18&#45;&gt;t20 -->
+<g id="edge31" class="edge"><title>t18&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M120.266,-796.105C70.9512,-780.34 17,-751.202 17,-699.173 17,-699.173 17,-699.173 17,-309.173 17,-269.597 31.8505,-255.79 65,-234.173 92.9782,-215.927 154.085,-202.475 221.024,-192.724"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="221.533,-196.187 230.939,-191.311 220.546,-189.256 221.533,-196.187"/>
+</g>
+<!-- t22 -->
+<g id="node23" class="node"><title>t22</title>
+<polygon fill="#efa03b" stroke="black" points="554.065,-47.332 721,-20.0049 887.935,-47.332 887.779,-91.5483 554.221,-91.5483 554.065,-47.332"/>
+<polygon fill="none" stroke="black" points="550.054,-43.9306 721,-15.9468 891.946,-43.9306 891.764,-95.5532 550.236,-95.5532 550.054,-43.9306"/>
+<text text-anchor="middle" x="721" y="-53.5531" font-family="Times,serif" font-size="20.00">ocrmypdf.pipeline.copy_final</text>
+</g>
+<!-- t20&#45;&gt;t22 -->
+<g id="edge41" class="edge"><title>t20&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M556.585,-133.609C580.784,-122.857 607.72,-110.888 632.67,-99.8018"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="634.408,-102.859 642.125,-95.6004 631.566,-96.4625 634.408,-102.859"/>
+</g>
+<!-- t21&#45;&gt;t22 -->
+<g id="edge40" class="edge"><title>t21&#45;&gt;t22</title>
+<path fill="none" stroke="#0044a0" d="M888.417,-134.636C863.556,-123.635 835.613,-111.27 809.783,-99.8401"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="810.933,-96.5216 800.372,-95.6756 808.1,-102.923 810.933,-96.5216"/>
+</g>
+</g>
+</svg>

docs/release_notes.rst

@@ -0,0 +1,709 @@
+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.
+
+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: 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: 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: fix final merge failure for PDFs with more pages than the system file handle limit (``ulimit -n``)
+-   Fix issue #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 cause qpdf to infinite-loop
+-   Address issue #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)
+-   Using a non Latin-1 language with the "hocr" renderer now warns about possible OCR quality and recommends workarounds (#176)
+-   Output file path added to error message when that location is not writable (#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, 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.
+-   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.
+-   Fixed issue #143, added ``--quiet`` to suppress "INFO" messages
+-   Fixed issue #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, '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, ocrmypdf now stops and terminates if Ghostscript fails on an intermediate step, as it is not possible to proceed.
+-   Fixed issue #160, exception thrown on certain invalid arguments instead of error message
+
+
+v4.5.5
+------
+
+-   Automated update of macOS homebrew tap
+-   Fixed issue #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) (thanks to @TomRaz)
+-   Fix an issue where pages with no images might trigger "cannot write mode P as JPEG" (#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. ``--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, 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; 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) with PDFs that omit the optional /BitsPerComponent parameter on images
+-   Removed non-free file milk.pdf
+
+
+v4.2.4
+------
+
+-   Fixed an error (#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)
+
+    +  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
+
+-   Fixes an issue where, with certain settings, monochrome images in PDFs would be converted to 8-bit grayscale, increasing file size (#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) 
+-   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)
+
+
+Changes
+^^^^^^^
+
+-   Logging output is now much easier to read
+-   ``--deskew`` is now performed by Leptonica instead of unpaper (#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 "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 - 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: 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)
+------------------------
+
+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

@@ -1,23 +1,10 @@
-from enum import IntEnum
-import os
+# © 2017 James R. Barlow: github.com/jbarlow83
 
+import pkg_resources
 
-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
-    child_process_error = 7
-    other_error = 15
+PROGRAM_NAME = 'ocrmypdf'
 
+# Official PEP 396
+__version__ = pkg_resources.get_distribution('ocrmypdf').version
 
-def get_program(name):
-    envvar = 'OCRMYPDF_' + name.upper()
-    return os.environ.get(envvar, name)
-
-
-def page_number(input_file):
-    return int(os.path.basename(input_file)[0:6])
+VERSION = __version__

ocrmypdf/__main__.py

@@ -0,0 +1,839 @@
+#!/usr/bin/env python3
+# © 2015-17 James R. Barlow: github.com/jbarlow83
+
+from tempfile import mkdtemp
+from collections.abc import Sequence
+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
+from .exec import tesseract, qpdf, ghostscript
+from . import PROGRAM_NAME, VERSION
+
+from .exceptions import ExitCode, ExitCodeException, MissingDependencyError
+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
+
+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'], 
+    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'."
+         )
+
+# 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=int, 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=float, 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=float, 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', 'tess4', '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"
+         "tess4 - deprecated alias for 'sandwich'"
+    )
+advanced.add_argument(
+    '--tesseract-timeout', default=180.0, type=float, 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=float, 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.")
+
+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 == 'tess4':
+        log.warning("The 'tess4' PDF renderer has been renamed to 'sandwich'. "
+                    "Please use --pdf-renderer=sandwich.")
+        options.pdf_renderer = 'sandwich'
+
+    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")
+
+    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 tesseract.v4():
+        log.info(
+            "Tesseract v4.x.alpha found.")
+    if options.tesseract_oem and not tesseract.v4():
+        log.warning(
+            "--tesseract-oem requires Tesseract 4.x -- argument ignored")
+    if options.pdf_renderer == 'tess4' and not tesseract.has_textonly_pdf():
+        raise MissingDependencyError(
+            "--pdf-renderer tess4 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'"
+        )
+
+
+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(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)
+    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 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 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(e_args, options, log):
+    """Walk through a RethrownJobError and find the first exception.
+
+    The exit code will be based on this, even if multiple exceptions occurred
+    at the same time."""
+
+    if isinstance(e_args, Sequence) and isinstance(e_args[0], str) and \
+            len(e_args) == 5:
+        return do_ruffus_exception(e_args, options, log)
+    elif is_iterable_notstr(e_args):
+        for exc in e_args:
+            return traverse_ruffus_exception(exc, options, log)
+
+
+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.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 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)
+    _log.debug('ocrmypdf ' + VERSION)
+    _log.debug('tesseract ' + tesseract.version())
+    _log.debug('qpdf ' + qpdf.version())
+
+    check_options(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
+
+    # 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()))
+
+    # 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()
+    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')
+
+        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)
+                return ExitCode.input_file
+
+        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."""))
+                return ExitCode.bad_args
+        elif not is_file_writable(options.output_file):
+            _log.error(
+                "Output file location (" + options.output_file + ") " +
+                "is not a writable file.")
+            return ExitCode.file_access_error
+
+        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
+
+        # Ruffus flattens exception to 5 element tuples. Because of a bug
+        # in <= 2.6.3 it may present either the single:
+        #   (task, job, exc, value, stack)
+        # or something like:
+        #   [[(task, job, exc, value, stack)]]
+        #
+        # Generally cross-process exception marshalling doesn't work well
+        # and ruffus doesn't support because BaseException has its own
+        # implementation of __reduce__ that attempts to reconstruct the
+        # exception based on e.__init__(e.args).
+        #
+        # Attempting to log the exception directly marshalls it to the logger
+        # which is probably in another process, so it's better to log only
+        # data from the exception at this point.
+
+        exitcode = traverse_ruffus_exception(e.args, 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))
+    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
+
+    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())

ocrmypdf/_unicodefun.py

@@ -0,0 +1,108 @@
+# 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."""
+    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)

ocrmypdf/exceptions.py

@@ -0,0 +1,58 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+
+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 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 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

ocrmypdf/exec/__init__.py

@@ -0,0 +1,48 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+"""Wrappers to manage subprocess calls"""
+
+import os
+import re
+import sys
+from subprocess import run, STDOUT, PIPE, CalledProcessError
+from ..exceptions import MissingDependencyError
+
+
+def get_program(name):
+    "Check environment variables for overrides to this program"
+    envvar = 'OCRMYPDF_' + name.upper()
+    return os.environ.get(envvar, name)
+
+
+def get_version(program, *, 
+        version_arg='--version', regex=r'(\d+(\.\d+)*)'):
+    "Get the version of the specified program, "
+    args_prog = [
+        get_program(program),
+        version_arg
+    ]
+    try:
+        proc = run(
+            args_prog, close_fds=True, universal_newlines=True,
+            stdout=PIPE, stderr=STDOUT, check=True)
+        output = proc.stdout
+    except CalledProcessError as e:
+        if get_program(program) == program:
+            raise MissingDependencyError(
+                "Could not find program '{}' on the PATH".format(
+                    program)) from e
+        else:
+            raise MissingDependencyError(
+                "Could not find program '{}'".format(
+                    get_program(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

ocrmypdf/exec/ghostscript.py

@@ -0,0 +1,156 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+
+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_program, get_version
+from ..exceptions import SubprocessOutputError, MissingDependencyError
+from ..helpers import fspath
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('gs')
+
+
+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 = [
+            get_program('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",
+        ]
+
+    with NamedTemporaryFile(delete=True) as gs_pdf:
+        args_gs = [
+            get_program("gs"),
+            "-dQUIET",
+            "-dBATCH",
+            "-dNOPAUSE",
+            "-dCompatibilityLevel=" + str(pdf_version),
+            "-dNumRenderingThreads=" + str(threads),
+            "-sDEVICE=pdfwrite",
+            "-dAutoRotatePages=/None",
+            "-sColorConversionStrategy=/RGB",
+            "-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()

ocrmypdf/exec/qpdf.py

@@ -0,0 +1,203 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+
+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_program, 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 = [
+        get_program('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 = [
+        get_program('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(
+            [get_program('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 = [
+            get_program('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 _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 = [
+        get_program('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)
+
+

ocrmypdf/exec/tesseract.py

@@ -0,0 +1,353 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+import re
+import shutil
+from functools import lru_cache
+from collections import namedtuple
+from textwrap import dedent
+import PyPDF2 as pypdf
+from subprocess import PIPE, CalledProcessError, \
+    TimeoutExpired, check_output, STDOUT
+
+from ..exceptions import MissingDependencyError, TesseractConfigError
+from ..helpers import page_number
+from . import get_program, 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 = [
+        get_program('tesseract'),
+        '--print-parameters'
+    ]
+    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 = [
+        get_program('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 = [
+        get_program('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, hopefully also 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:
+        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)

ocrmypdf/exec/unpaper.py

@@ -1,26 +1,14 @@
-#!/usr/bin/env python3
 # © 2015 James R. Barlow: github.com/jbarlow83
 # unpaper documentation:
 # https://github.com/Flameeyes/unpaper/blob/master/doc/basic-concepts.md
 
-from subprocess import CalledProcessError, STDOUT, check_output, check_call
+from subprocess import CalledProcessError, STDOUT, check_output
 from tempfile import NamedTemporaryFile
 import sys
 import os
 from functools import lru_cache
-from . import ExitCode, get_program
-
-
-@lru_cache(maxsize=1)
-def version():
-    args_unpaper = [
-        get_program('unpaper'),
-        '--version'
-    ]
-    version = check_output(
-        args_unpaper, close_fds=True, universal_newlines=True,
-        stderr=STDOUT, timeout=5)
-    return version.strip()
+from ..exceptions import MissingDependencyError
+from . import get_program, get_version
 
 
 try:
@@ -30,6 +18,11 @@ 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 = [
         get_program('unpaper'),
@@ -47,17 +40,17 @@ def run(input_file, output_file, dpi, log, mode_args):
                 im = im.convert(mode='1')
             else:
                 im = im.convert(mode='RGB')
-        except IOError:
+        except IOError as e:
             log.error(
                     "Could not convert image with type " + im.mode)
-            sys.exit(ExitCode.missing_dependency)
+            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)
+        raise MissingDependencyError() from e
 
     with NamedTemporaryFile(suffix=suffix) as input_pnm, \
             NamedTemporaryFile(suffix=suffix, mode="r+b") as output_pnm:
@@ -77,7 +70,8 @@ def run(input_file, output_file, dpi, log, mode_args):
             raise e from e
         else:
             log.debug(stdout)
-            Image.open(output_pnm.name).save(output_file)
+            # unpaper sets dpi to 72
+            Image.open(output_pnm.name).save(output_file, dpi=(dpi, dpi))
 
 
 def deskew(input_file, output_file, dpi, log):

ocrmypdf/ghostscript.py

@@ -1,60 +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
-from . import get_program
-from .pdfa import SRGB_ICC_PROFILE
-
-
-def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
-                  pageno=1):
-    with NamedTemporaryFile(delete=True) as tmp:
-        args_gs = [
-            get_program('gs'),
-            '-dQUIET',
-            '-dBATCH',
-            '-dNOPAUSE',
-            '-sDEVICE=%s' % raster_device,
-            '-dFirstPage=%i' % pageno,
-            '-dLastPage=%i' % pageno,
-            '-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 = [
-            get_program("gs"),
-            "-dQUIET",
-            "-dBATCH",
-            "-dNOPAUSE",
-            '-dNumRenderingThreads=' + str(threads),
-            "-sDEVICE=pdfwrite",
-            "-dAutoRotatePages=/None",
-            "-sColorConversionStrategy=/RGB",
-            "-sProcessColorModel=DeviceRGB",
-            "-dJPEGQ=95",
-            "-dPDFA=2",
-            "-sPDFACompatibilityPolicy=2",
-            "-sOutputFile=" + gs_pdf.name,
-        ]
-        args_gs.extend(pdf_pages)
-        check_call(args_gs)
-        copy(gs_pdf.name, output_file)

ocrmypdf/helpers.py

@@ -0,0 +1,135 @@
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+from functools import partial
+from collections.abc import Iterable
+from contextlib import suppress, contextmanager
+from pathlib import Path
+import sys
+import os
+
+
+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):
+    return int(os.path.basename(input_file)[0:6])
+
+
+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):
+            resolve = lambda: Path(os.path.realpath(str(p)))
+        else:
+            resolve = lambda: p.resolve(strict=False)
+        p = resolve()
+
+    # 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

ocrmypdf/hocrtransform.py

@@ -1,12 +1,33 @@
 #!/usr/bin/env 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
-##############################################################################
+# 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

ocrmypdf/leptonica.py

@@ -1,14 +1,10 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 #
-# © 2013-15: jbarlow83 from Github (https://github.com/jbarlow83)
+# © 2013-16: 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.
+# Python FFI wrapper for Leptonica library
 
-from __future__ import print_function, absolute_import, division
 import argparse
 import sys
 import os
@@ -17,6 +13,8 @@ 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'))
 
@@ -43,22 +41,34 @@ class LeptonicaErrorTrap(object):
 
     """
     def __enter__(self):
+        from io import UnsupportedOperation
         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())
+        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
-        os.dup2(self.old_stderr_fileno, sys.stderr.fileno())
+        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)
@@ -83,6 +93,13 @@ class LeptonicaIOError(LeptonicaError):
     pass
 
 
+class RemoveColormap(Enum):
+    to_binary = 0
+    to_grayscale = 1
+    to_full_color = 2
+    based_on_src = 3
+
+
 class Pix:
     """Wrapper around leptonica's PIX object.
 
@@ -96,29 +113,55 @@ class Pix:
 
     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, cpix):
-        self.cpix = ffi.gc(cpix, Pix._pix_destroy)
+    def __init__(self, pix):
+        self._pix = ffi.gc(pix, Pix._pix_destroy)
 
     def __repr__(self):
-        if self.cpix:
+        if self._pix:
             s = "<leptonica.Pix image size={0}x{1} depth={2} at 0x{3:x}>"
-            return s.format(self.cpix.w, self.cpix.h, self.cpix.d,
-                            int(ffi.cast("intptr_t", self.cpix)))
+            return s.format(self._pix.w, self._pix.h, self._pix.d,
+                            int(ffi.cast("intptr_t", self._pix)))
         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.cpix, data, size)
+        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)
 
@@ -126,39 +169,101 @@ class Pix:
         cdata_bytes = ffi.new('char[]', state['data'])
         cdata_uint32 = ffi.cast('l_uint32 *', cdata_bytes)
 
-        self.cpix = lept.pixDeserializeFromMemory(
+        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.cpix.w
+        return self._pix.w
 
     @property
     def height(self):
-        return self.cpix.h
+        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, filename):
+    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(
-                filename.encode(sys.getfilesystemencoding())))
+            return cls(lept.pixRead(os.fsencode(filename)))
 
     def write_implied_format(
-            self, filename, jpeg_quality=0, jpeg_progressive=0):
+            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(
-                filename.encode(sys.getfilesystemencoding()),
-                self.cpix, jpeg_quality, jpeg_progressive)
+                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.
@@ -170,16 +275,21 @@ class Pix:
             for skew angle
         """
         with LeptonicaErrorTrap():
-            return Pix(lept.pixDeskew(self.cpix, reduction_factor))
+            return Pix(lept.pixDeskew(self._pix, reduction_factor))
 
-    def scale(self, scalex, scaley):
+    def scale(self, scale_xy):
         "Returns the pix object rescaled according to the proportions given."
         with LeptonicaErrorTrap():
-            return Pix(lept.pixScale(self.cpix, scalex, scaley))
+            return Pix(lept.pixScale(self._pix, scale_xy[0], scale_xy[1]))
 
     def rotate180(self):
         with LeptonicaErrorTrap():
-            return Pix(lept.pixRotate180(ffi.NULL, self.cpix))
+            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).
@@ -189,31 +299,132 @@ class Pix:
         with LeptonicaErrorTrap():
             angle = ffi.new('float *', 0.0)
             confidence = ffi.new('float *', 0.0)
-            result = lept.pixFindSkew(self.cpix, angle, confidence)
+            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
+
+            removal_type - RemovalColormap()
+        """
+
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixRemoveColormap(self._pix, removal_type))
+
     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_cpix = ffi.new('PIX **')
+            p_pix = ffi.new('PIX **')
 
             result = lept.pixOtsuAdaptiveThreshold(
-                self.cpix,
+                self._pix,
                 sx, sy,
                 smoothx, smoothy,
                 scorefract,
                 ffi.NULL,
-                p_cpix)
+                p_pix)
             if result == 0:
-                return Pix(p_cpix[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)):
+        with LeptonicaErrorTrap():
+            return Pix(lept.pixBackgroundNorm(
+                self._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():
@@ -224,16 +435,17 @@ class Pix:
         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.
+            # images.  Ubuntu trusty has 1.70. Ubuntu PPA
+            # ppa:rebuntu16/avidemux+unofficial has "leptonlib" 1.73.
             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.cpix, pix1_count, tab8)
-            lept.pixCountPixels(pix2.cpix, pix2_count, tab8)
-            pixn = Pix(lept.pixAnd(ffi.NULL, pix1.cpix, pix2.cpix))
-            lept.pixCountPixels(pixn.cpix, pixn_count, 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.
@@ -244,7 +456,7 @@ class Pix:
             return correlation
         else:
             correlation = ffi.new('float *', 0.0)
-            result = lept.pixCorrelationBinary(pix1.cpix, pix2.cpix,
+            result = lept.pixCorrelationBinary(pix1._pix, pix2._pix,
                                                correlation)
             if result != 0:
                 raise LeptonicaError("Correlation failed")
@@ -252,11 +464,48 @@ class Pix:
 
     @staticmethod
     def _pix_destroy(pix):
-        ptr_to_pix = ffi.new('PIX **', pix)
-        lept.pixDestroy(ptr_to_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.
@@ -286,6 +535,22 @@ def deskew(infile, outfile, dpi):
         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")
@@ -301,35 +566,4 @@ if __name__ == '__main__':
     parser_deskew.set_defaults(func=deskew)
 
     args = parser.parse_args()
-
-    if get_leptonica_version() != u'leptonica-1.69':
-        print("Unexpected leptonica version: %s" % getLeptonicaVersion())
-
-    args.func(args)
-
-
-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)
-            print('{0} {1}  {2}'.format(rotate_angle, angle, confidence), file=sys.stderr)
-
-
+    args.func(args)

ocrmypdf/lib/__init__.py

@@ -0,0 +1,2 @@
+"""Bindings to external libraries"""
+# © 2017 James R. Barlow: github.com/jbarlow83

ocrmypdf/lib/compile_leptonica.py

@@ -1,8 +1,11 @@
+#!/usr/bin/env python3
+# © 2017 James R. Barlow: github.com/jbarlow83
+
 from cffi import FFI
 
-ffi = FFI()
-ffi.set_source("ocrmypdf.lib._leptonica", None)
-ffi.cdef("""
+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;
@@ -42,18 +45,43 @@ struct PixColormap
     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;
+
 """)
 
-ffi.cdef("""
+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 );
@@ -62,6 +90,10 @@ 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,
@@ -72,9 +104,121 @@ pixOtsuAdaptiveThreshold(PIX       *pixs,
                          PIX      **ppixth,
                          PIX      **ppixd);
 
-void lept_free(void *ptr);
+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);
+
+void                 
+boxDestroy(BOX  **pbox);
+
+void
+lept_free(void *ptr);  
 """)
 
 
 if __name__ == '__main__':
-    ffi.compile()
+    ffibuilder.compile(verbose=True)

ocrmypdf/pageinfo.py

@@ -2,12 +2,17 @@
 # © 2015 James R. Barlow: github.com/jbarlow83
 
 from subprocess import Popen, PIPE
-from decimal import Decimal, getcontext
-from math import hypot
+from decimal import Decimal
+from math import hypot, isclose
 import re
 import sys
 import PyPDF2 as pypdf
 from collections import namedtuple
+import warnings
+
+
+warnings.warn("ocrmypdf.pageinfo is deprecated'; use ocrmypdf.pdfinfo",
+              DeprecationWarning)
 
 matrix_mult = pypdf.pdf.utils.matrixMultiply
 
@@ -52,6 +57,9 @@ FRIENDLY_COMP = {
 }
 
 
+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
 
@@ -73,38 +81,96 @@ def _shorthand_from_matrix(matrix):
     return tuple(map(float, (a, b, c, d, e, f)))
 
 
-ContentsInfo = namedtuple('ContentsInfo', ['raster_settings', 'inline_images'])
+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'])
 
 
-def _interpret_contents(contentstream):
+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((1, 0, 0, 1, 0, 0))
-    image_raster_settings = []
+    ctm = _matrix_from_shorthand(initial_shorthand)
+    xobject_settings = []
     inline_images = []
 
-    for op in operations:
+    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")
+                raise RuntimeError(
+                    "PDF graphics stack overflow, command %i" % n)
         elif command == b'Q':
-            ctm = stack.pop()
+            try:
+                ctm = stack.pop()
+            except IndexError:
+                raise RuntimeError(
+                    "PDF graphics stack underflow, command %i" % n)
         elif command == b'cm':
             ctm = matrix_mult(
-                ctm, _matrix_from_shorthand(operands))
+                _matrix_from_shorthand(operands), ctm)
         elif command == b'Do':
             image_name = operands[0]
-            image_raster_settings.append(
-                (image_name, _shorthand_from_matrix(ctm)))
+            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_images.append(
-                (settings, _shorthand_from_matrix(ctm)))
+            inline = InlineSettings(
+                settings=settings, shorthand=_shorthand_from_matrix(ctm),
+                stack_depth=len(stack))
+            inline_images.append(inline)
 
     return ContentsInfo(
-        raster_settings=image_raster_settings,
+        xobject_settings=xobject_settings,
         inline_images=inline_images)
 
 
@@ -172,48 +238,92 @@ def _get_dpi(ctm_shorthand, image_size):
     return (dpi_w, dpi_h)
 
 
-def _find_page_inline_images(page, pageinfo, contentsinfo):
-    "Find inline images on the page"
+def _find_inline_images(contentsinfo):
+    "Find inline images in the contentstream"
 
-    for n, im in enumerate(contentsinfo.inline_images):
-        settings, shorthand = im
+    for n, inline in enumerate(contentsinfo.inline_images):
         image = {}
         image['name'] = str('inline-%02d' % n)
-        image['width'] = settings['/W']
-        image['height'] = settings['/H']
-        image['bpc'] = settings['/BPC']
-        image['color'] = FRIENDLY_COLORSPACE.get(settings['/CS'], '-')
+        image['width'] = inline.settings['/W']
+        image['height'] = inline.settings['/H']
+        if '/BPC' in inline.settings:
+            image['bpc'] = inline.settings['/BPC']
+        else:
+            image['bpc'] = 8
+        if '/CS' in inline.settings:
+            image['color'] = FRIENDLY_COLORSPACE.get(inline.settings['/CS'], '-')
+        else:
+            image['color'] = '-'
         image['comp'] = FRIENDLY_COMP.get(image['color'], '?')
-        if '/F' in settings:
-            filter_ = settings['/F']
+        if '/F' in inline.settings:
+            filter_ = inline.settings['/F']
             if isinstance(filter_, pypdf.generic.ArrayObject):
                 filter_ = filter_[0]
             image['enc'] = FRIENDLY_ENCODING.get(filter_, 'image')
         else:
             image['enc'] = 'image'
 
-        dpi_w, dpi_h = _get_dpi(shorthand, (image['width'], image['height']))
+        dpi_w, dpi_h = _get_dpi(
+            inline.shorthand, (image['width'], image['height']))
         image['dpi_w'], image['dpi_h'] = Decimal(dpi_w), Decimal(dpi_h)
         yield image
 
 
-def _find_page_regular_images(page, pageinfo, contentsinfo):
-    "Find images stored in XObject resources"
+def _image_xobjects(container):
+    """Search for all XObject-based images in the container
 
-    try:
-        page['/Resources']['/XObject']
-    except KeyError:
+    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
-    for xobj in page['/Resources']['/XObject']:
-        # PyPDF2 returns the keys as an iterator
-        pdfimage = page['/Resources']['/XObject'][xobj]
-        if pdfimage['/Subtype'] != '/Image':
-            continue
+    resources = container['/Resources']
+    if '/XObject' not in resources:
+        return
+    for xobj in resources['/XObject']:
+        candidate = resources['/XObject'][xobj]
+        if candidate['/Subtype'] == '/Image':
+            image = candidate
+            yield (image, 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):
         image = {}
-        image['name'] = str(xobj)
+        image['name'] = xobj
         image['width'] = pdfimage['/Width']
         image['height'] = pdfimage['/Height']
-        image['bpc'] = pdfimage['/BitsPerComponent']
+        if '/BitsPerComponent' in pdfimage:
+            image['bpc'] = pdfimage['/BitsPerComponent']
+        else:
+            image['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:
+            image['type'] = 'stencil' if pdfimage['/ImageMask'].value \
+                            else 'image'
+        else:
+            image['type'] = 'image'
         if '/Filter' in pdfimage:
             filter_ = pdfimage['/Filter']
             if isinstance(filter_, pypdf.generic.ArrayObject):
@@ -230,34 +340,130 @@ def _find_page_regular_images(page, pageinfo, contentsinfo):
             image['color'] = 'jpx' if image['enc'] == 'jpx' else '?'
 
         image['comp'] = FRIENDLY_COMP.get(image['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 image['comp'] == '?' and image['enc'] in ('ccitt', 'jbig2'):
+            image['comp'] = FRIENDLY_COMP['gray']
+
         image['dpi_w'] = image['dpi_h'] = 0
 
-        for raster in contentsinfo.raster_settings:
+        for xobj in contentsinfo.xobject_settings:
             # Loop in case the same image is display multiple times on a page
-            if raster[0] != image['name']:
+            if xobj.name != image['name']:
+                continue
+
+            if xobj.stack_depth == 0 and _is_unit_square(xobj.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
-            shorthand = raster[1]
 
             dpi_w, dpi_h = _get_dpi(
-                shorthand, (image['width'], image['height']))
+                xobj.shorthand, (image['width'], image['height']))
 
             # When image is used multiple times take the highest DPI it is
             # rendered at
             image['dpi_w'] = max(dpi_w, image.get('dpi_w', 0))
             image['dpi_h'] = max(dpi_h, image.get('dpi_h', 0))
 
-        image['dpi_w'] = Decimal(image['dpi_w'])
-        image['dpi_h'] = Decimal(image['dpi_h'])
-        image['dpi'] = (image['dpi_w'] * image['dpi_h']) ** Decimal(0.5)
+        DPI_PREC = Decimal('1.000')
+        dpi = Decimal(image['dpi_w'] * image['dpi_h']).sqrt()
+        image['dpi_w'] = Decimal(image['dpi_w']).quantize(DPI_PREC)
+        image['dpi_h'] = Decimal(image['dpi_h']).quantize(DPI_PREC)
+        image['dpi'] = dpi.quantize(DPI_PREC)
         yield image
 
 
-def _find_page_images(page, pageinfo, contentsinfo):
-    yield from _find_page_inline_images(page, pageinfo, contentsinfo)
-    yield from _find_page_regular_images(page, pageinfo, contentsinfo)
+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, form_xobject, 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 _page_has_text(pdf, page):
+    if not '/Contents' in page:
+        return False
+
     # Simple test
     text = page.extractText()
     if text.strip() != '':
@@ -295,14 +501,12 @@ def _pdf_get_pageinfo(infile, pageno: int):
     pageinfo['height_inches'] = height_pt / Decimal(72.0)
 
     try:
-        contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
-    except AttributeError as e:
-        return pageinfo
-
-    contentsinfo = _interpret_contents(contentstream)
-    pageinfo['images'] = [im for im in _find_page_images(
-                                page, pageinfo, contentsinfo)]
+        pageinfo['rotate'] = int(page['/Rotate'])
+    except KeyError:
+        pageinfo['rotate'] = 0
 
+    pageinfo['images'] = [im for im in
+                          _find_images(pdf, page)]
     if pageinfo['images']:
         xres = max(image['dpi_w'] for image in pageinfo['images'])
         yres = max(image['dpi_h'] for image in pageinfo['images'])
@@ -317,7 +521,6 @@ def _pdf_get_pageinfo(infile, pageno: int):
 
 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

@@ -1,14 +1,13 @@
-#!/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
 import codecs
 import pkg_resources
+import PyPDF2 as pypdf
 
-ICC_PROFILE_RELPATH = 'data/sRGB_IEC61966-2-1_black_scaled.icc'
+ICC_PROFILE_RELPATH = 'data/sRGB.icc'
 
 SRGB_ICC_PROFILE = pkg_resources.resource_filename(
     'ocrmypdf', ICC_PROFILE_RELPATH)
@@ -18,20 +17,15 @@ SRGB_ICC_PROFILE = pkg_resources.resource_filename(
 # 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.
+% 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
 
-[ /Title <$title>
-  /Author <$author>
-  /Subject <$subject>
-  /Keywords <$keywords>
-  /Creator <$creator>
+[$docinfo
   /DOCINFO pdfmark
 
 % Define an ICC profile :
@@ -69,13 +63,20 @@ def encode_text_string(s: str) -> str:
 
     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.
+    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 = codecs.encode(b'\xfe\xff' + utf16_bytes, 'hex')
     ascii_hex_str = ascii_hex_bytes.decode('ascii').lower()
@@ -83,20 +84,30 @@ def encode_text_string(s: str) -> str:
 
 
 def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
-    pdfmark_utf16 = {k: encode_text_string(v) for k, v in pdfmark.items()}
+    # 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')
+    docinfo_line_template = '  {key} <{value}>'
+
+    def docinfo_gen():
+        for key in docinfo_keys:
+            if key in pdfmark and pdfmark[key].strip() != '':
+                line = docinfo_line_template.format(
+                    key=key, value=encode_text_string(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,
-                          title=pdfmark_utf16.get('title', ''),
-                          author=pdfmark_utf16.get('author', ''),
-                          subject=pdfmark_utf16.get('subject', ''),
-                          creator=pdfmark_utf16.get('creator', ''),
-                          keywords=pdfmark_utf16.get('keywords', ''))
+                          docinfo=docinfo)
     return result
 
 
-def generate_pdfa_def(target_filename, pdfmark, icc='sRGB'):
+def generate_pdfa_ps(target_filename, pdfmark, icc='sRGB'):
     if icc == 'sRGB':
         icc_profile = SRGB_ICC_PROFILE
     else:
@@ -104,7 +115,48 @@ def generate_pdfa_def(target_filename, pdfmark, icc='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
+    # 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)
+    xmp = pdf.getXmpMetadata()
+
+    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
+

ocrmypdf/pdfinfo.py

@@ -0,0 +1,725 @@
+#!/usr/bin/env python3
+# © 2015 James R. Barlow: github.com/jbarlow83
+
+from subprocess import Popen, PIPE
+from decimal import Decimal
+from math import hypot, isclose
+import re
+import sys
+import PyPDF2 as pypdf
+from collections import namedtuple
+from collections.abc import MutableMapping, Mapping
+import warnings
+from pathlib import Path
+from enum import Enum
+from .helpers import universal_open
+
+
+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'])
+
+
+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 = []
+
+    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)
+
+    return ContentsInfo(
+        xobject_settings=xobject_settings,
+        inline_images=inline_images)
+
+
+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 _page_has_text(pdf, page):
+    if not '/Contents' in page:
+        return False
+
+    # Simple test
+    text = page.extractText()
+    if text.strip() != '':
+        return True
+
+    # More nuanced test to deal with quirks of Tesseract PDF generation
+    # Check if there's a Glyphless font
+    try:
+        font = page['/Resources']['/Font']
+    except KeyError:
+        pass
+    else:
+        font_objects = list(font.keys())
+        for font_object in font_objects:
+            basefont = font[font_object]['/BaseFont']
+            if basefont.endswith('GlyphLessFont'):
+                return True
+
+    return False
+
+
+def _pdf_get_pageinfo(pdf, pageno: int):
+    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]
+
+    pageinfo['has_text'] = _page_has_text(pdf, 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) for n in range(pdf.numPages)]
+
+
+class PageInfo:
+    def __init__(self, pdf, pageno):
+        self._pageno = pageno
+        self._pageinfo = _pdf_get_pageinfo(pdf, pageno)
+
+    @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)
+
+    @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)
+
+    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()

ocrmypdf/qpdf.py

@@ -1,87 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-from subprocess import CalledProcessError, check_output, STDOUT, check_call
-import sys
-import os
-
-from . import ExitCode, get_program
-
-
-def check(input_file, log):
-    args_qpdf = [
-        get_program('qpdf'),
-        '--check',
-        input_file
-    ]
-
-    try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=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 repair(input_file, output_file, log):
-    args_qpdf = [
-        get_program('qpdf'), input_file, output_file
-    ]
-    try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 3 and e.output.find("operation succeeded"):
-            log.debug('qpdf found and fixed errors: ' + e.output)
-            log.debug(e.output)
-            return
-
-        if e.returncode == 2 and e.output.find("invalid password"):
-            log.error("{0}: this PDF is password-protected - password must "
-                      "be removed for OCR".format(input_file))
-            sys.exit(ExitCode.input_file)
-        elif e.returncode == 2:
-            log.error("{0}: not a valid PDF, and could not repair it.".format(
-                      input_file))
-            log.error("Details: " + e.output)
-            sys.exit(ExitCode.input_file)
-        else:
-            log.error("{0}: unknown error".format(
-                      input_file))
-            log.error(e.output)
-            sys.exit(ExitCode.unknown)
-
-
-def get_npages(input_file, log):
-    try:
-        pages = check_output(
-            [get_program('qpdf'), '--show-npages', input_file],
-             universal_newlines=True, close_fds=True)
-    except CalledProcessError as e:
-        if e.returncode == 2 and e.output.find('No such file'):
-            log.error(e.output)
-            sys.exit(ExitCode.input_file)
-    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 = [
-            get_program('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)

ocrmypdf/tesseract.py

@@ -1,260 +0,0 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
-
-import sys
-import os
-import re
-import shutil
-from functools import lru_cache
-from . import ExitCode, get_program, page_number
-from collections import namedtuple
-
-from subprocess import Popen, PIPE, CalledProcessError, \
-    TimeoutExpired, check_output, STDOUT
-try:
-    from subprocess import DEVNULL
-except ImportError:
-    DEVNULL = open(os.devnull, 'wb')
-
-
-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 3.02.02' />
-  <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>'''
-
-
-@lru_cache(maxsize=1)
-def version():
-    args_tess = [
-        get_program('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 = [
-        get_program('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:])
-
-
-def get_orientation(input_file, language: list, timeout: float, log):
-    args_tesseract = [
-        get_program('tesseract'),
-        '-l', '+'.join(language),
-        '-psm', '0',
-        input_file,
-        'stdout'
-    ]
-
-    try:
-        stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, timeout=timeout)
-    except TimeoutExpired:
-        return OrientationConfidence(angle=0, confidence=0.0)
-    except CalledProcessError as e:
-        tesseract_log_output(log, e.output, input_file)
-        if ('Too few characters. Skipping this page' in e.output or
-                'Image too large' in e.output):
-            return OrientationConfidence(0, 0)
-        raise e from e
-    else:
-        osd = {}
-        for line in stdout.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, hopefully also 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):
-    lines = stdout.splitlines()
-    prefix = "{0:4d}: [tesseract] ".format(page_number(input_file))
-    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())
-        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, pageinfo):
-    with open(output_hocr, 'w', encoding="utf-8") as f:
-        f.write(HOCR_TEMPLATE.format(
-            pageinfo['width_pixels'],
-            pageinfo['height_pixels']))
-
-
-def generate_hocr(input_file, output_hocr, language: list, tessconfig: list,
-                  timeout: float, pageinfo_getter, pagesegmode: int, log):
-
-    badxml = os.path.splitext(output_hocr)[0] + '.badxml'
-
-    args_tesseract = [
-        get_program('tesseract'),
-        '-l', '+'.join(language)
-    ]
-
-    if pagesegmode is not None:
-        args_tesseract.extend(['-psm', str(pagesegmode)])
-
-    args_tesseract.extend([
-        input_file,
-        badxml,
-        'hocr'
-    ] + tessconfig)
-    try:
-        stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, 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, pageinfo_getter())
-    except CalledProcessError as e:
-        tesseract_log_output(log, e.output, input_file)
-        if 'Image too large' in e.output:
-            _generate_null_hocr(output_hocr, pageinfo_getter())
-            return
-
-        raise e from e
-    else:
-        tesseract_log_output(log, stdout, input_file)
-        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_hocr, 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)
-
-
-def generate_pdf(input_image, skip_pdf, output_pdf, language: list,
-                 tessconfig: list, timeout: float, pagesegmode: int, log):
-    '''Use Tesseract to render a PDF.
-
-    input_image -- image to analyze
-    skip_pdf -- if we time out, use this file as output
-    language -- list of languages to consider
-    tessconfig -- tesseract configuration
-    timeout -- timeout (seconds)
-    log -- logger object
-    '''
-
-    args_tesseract = [
-        get_program('tesseract'),
-        '-l', '+'.join(language)
-    ]
-
-    if pagesegmode is not None:
-        args_tesseract.extend(['-psm', str(pagesegmode)])
-
-    args_tesseract.extend([
-        input_image,
-        os.path.splitext(output_pdf)[0],  # Tesseract appends suffix
-        'pdf'
-    ] + tessconfig)
-
-    try:
-        stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, timeout=timeout)
-    except TimeoutExpired:
-        page_timedout(log, input_image)
-        shutil.copy(skip_pdf, output_pdf)
-    except CalledProcessError as e:
-        tesseract_log_output(log, e.output, input_image)
-        if 'Image too large' in e.output:
-            shutil.copy(skip_pdf, output_pdf)
-            return
-        raise e from e
-    else:
-        tesseract_log_output(log, stdout, input_image)

pipeline.svg

@@ -1,297 +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="1452pt" height="824pt"
- viewBox="0.00 0.00 1452.00 824.08" 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(4 820.083)">
-<title>Pipeline:</title>
-<polygon fill="white" stroke="none" points="-4,4 -4,-820.083 1448,-820.083 1448,4 -4,4"/>
-<g id="clust1" class="cluster"><title>clustertasks</title>
-<polygon fill="none" stroke="black" points="8,-8 8,-808.083 1436,-808.083 1436,-8 8,-8"/>
-<text text-anchor="middle" x="722" y="-780.083" 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="black" points="1256.54,-762.083 1033.46,-762.083 1029.46,-758.083 1029.46,-726.083 1252.54,-726.083 1256.54,-730.083 1256.54,-762.083"/>
-<polyline fill="none" stroke="black" points="1252.54,-758.083 1029.46,-758.083 "/>
-<polyline fill="none" stroke="black" points="1252.54,-758.083 1252.54,-726.083 "/>
-<polyline fill="none" stroke="black" points="1252.54,-758.083 1256.54,-762.083 "/>
-<text text-anchor="middle" x="1143" y="-738.083" font-family="Times,serif" font-size="20.00">repair_pdf</text>
-</g>
-<!-- t1 -->
-<g id="node2" class="node"><title>t1</title>
-<polygon fill="#efa03b" stroke="black" points="1234.11,-683.155 1030,-700.057 825.888,-683.155 826.078,-655.806 1233.92,-655.806 1234.11,-683.155"/>
-<polygon fill="none" stroke="black" points="1238.13,-686.834 1030,-704.069 821.866,-686.834 822.11,-651.808 1237.89,-651.808 1238.13,-686.834"/>
-<text text-anchor="middle" x="1030" y="-669.596" 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="#0044a0" d="M1113.9,-725.961C1103.58,-719.89 1091.66,-712.873 1080.09,-706.069"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1081.75,-702.982 1071.35,-700.927 1078.2,-709.015 1081.75,-702.982"/>
-</g>
-<!-- t14 -->
-<g id="node16" class="node"><title>t14</title>
-<polygon fill="#efa03b" stroke="black" points="1428.08,-625.109 1089.92,-625.109 1085.92,-621.109 1085.92,-589.109 1424.08,-589.109 1428.08,-593.109 1428.08,-625.109"/>
-<polyline fill="none" stroke="black" points="1424.08,-621.109 1085.92,-621.109 "/>
-<polyline fill="none" stroke="black" points="1424.08,-621.109 1424.08,-589.109 "/>
-<polyline fill="none" stroke="black" points="1424.08,-621.109 1428.08,-625.109 "/>
-<text text-anchor="middle" x="1257" y="-601.109" font-family="Times,serif" font-size="20.00">generate_postscript_stub</text>
-</g>
-<!-- t0&#45;&gt;t14 -->
-<g id="edge22" class="edge"><title>t0&#45;&gt;t14</title>
-<path fill="none" stroke="#0044a0" d="M1219.32,-726.004C1229.98,-720.592 1239.75,-713.466 1247,-704.083 1261.92,-684.78 1263.1,-656.385 1261.35,-635.486"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1264.8,-634.881 1260.22,-625.328 1257.84,-635.653 1264.8,-634.881"/>
-</g>
-<!-- t2 -->
-<g id="node3" class="node"><title>t2</title>
-<polygon fill="#efa03b" stroke="black" points="1029.88,-625.109 748.117,-625.109 744.117,-621.109 744.117,-589.109 1025.88,-589.109 1029.88,-593.109 1029.88,-625.109"/>
-<polyline fill="none" stroke="black" points="1025.88,-621.109 744.117,-621.109 "/>
-<polyline fill="none" stroke="black" points="1025.88,-621.109 1025.88,-589.109 "/>
-<polyline fill="none" stroke="black" points="1025.88,-621.109 1029.88,-625.109 "/>
-<text text-anchor="middle" x="887" y="-601.109" font-family="Times,serif" font-size="20.00">rasterize_preview</text>
-</g>
-<!-- t1&#45;&gt;t2 -->
-<g id="edge2" class="edge"><title>t1&#45;&gt;t2</title>
-<path fill="none" stroke="#0044a0" d="M980.788,-651.715C965.266,-644.498 948.187,-636.558 932.905,-629.452"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="934.195,-626.193 923.652,-625.15 931.244,-632.54 934.195,-626.193"/>
-</g>
-<!-- t3 -->
-<g id="node4" class="node"><title>t3</title>
-<polygon fill="#efa03b" stroke="black" points="1128.15,-567.109 893.852,-567.109 889.852,-563.109 889.852,-531.109 1124.15,-531.109 1128.15,-535.109 1128.15,-567.109"/>
-<polyline fill="none" stroke="black" points="1124.15,-563.109 889.852,-563.109 "/>
-<polyline fill="none" stroke="black" points="1124.15,-563.109 1124.15,-531.109 "/>
-<polyline fill="none" stroke="black" points="1124.15,-563.109 1128.15,-567.109 "/>
-<text text-anchor="middle" x="1009" y="-543.109" font-family="Times,serif" font-size="20.00">orient_page</text>
-</g>
-<!-- t1&#45;&gt;t3 -->
-<g id="edge4" class="edge"><title>t1&#45;&gt;t3</title>
-<path fill="none" stroke="#0044a0" d="M1037.63,-651.508C1042.21,-633.96 1045.92,-609.455 1039,-589.109 1037.4,-584.404 1034.98,-579.822 1032.19,-575.552"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1034.85,-573.258 1026.1,-567.282 1029.21,-577.409 1034.85,-573.258"/>
-</g>
-<!-- t2&#45;&gt;t3 -->
-<g id="edge3" class="edge"><title>t2&#45;&gt;t3</title>
-<path fill="none" stroke="#0044a0" d="M924.267,-589.003C936.328,-583.467 949.857,-577.257 962.502,-571.453"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="964.333,-574.463 971.961,-567.111 961.413,-568.102 964.333,-574.463"/>
-</g>
-<!-- t4 -->
-<g id="node5" class="node"><title>t4</title>
-<polygon fill="#efa03b" stroke="black" points="1027.3,-509.109 676.701,-509.109 672.701,-505.109 672.701,-473.109 1023.3,-473.109 1027.3,-477.109 1027.3,-509.109"/>
-<polyline fill="none" stroke="black" points="1023.3,-505.109 672.701,-505.109 "/>
-<polyline fill="none" stroke="black" points="1023.3,-505.109 1023.3,-473.109 "/>
-<polyline fill="none" stroke="black" points="1023.3,-505.109 1027.3,-509.109 "/>
-<text text-anchor="middle" x="850" y="-485.109" font-family="Times,serif" font-size="20.00">rasterize_with_ghostscript</text>
-</g>
-<!-- t3&#45;&gt;t4 -->
-<g id="edge5" class="edge"><title>t3&#45;&gt;t4</title>
-<path fill="none" stroke="#0044a0" d="M960.431,-531.003C943.787,-525.141 924.999,-518.524 907.704,-512.433"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="908.867,-509.132 898.272,-509.111 906.541,-515.734 908.867,-509.132"/>
-</g>
-<!-- t9 -->
-<g id="node11" class="node"><title>t9</title>
-<polygon fill="#efa03b" stroke="black" points="314.109,-277.109 19.8906,-277.109 15.8906,-273.109 15.8906,-241.109 310.109,-241.109 314.109,-245.109 314.109,-277.109"/>
-<polyline fill="none" stroke="black" points="310.109,-273.109 15.8906,-273.109 "/>
-<polyline fill="none" stroke="black" points="310.109,-273.109 310.109,-241.109 "/>
-<polyline fill="none" stroke="black" points="310.109,-273.109 314.109,-277.109 "/>
-<text text-anchor="middle" x="165" y="-253.109" font-family="Times,serif" font-size="20.00">select_image_layer</text>
-</g>
-<!-- t3&#45;&gt;t9 -->
-<g id="edge14" class="edge"><title>t3&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M889.675,-542.297C675.505,-530.255 248,-498.192 248,-434.109 248,-434.109 248,-434.109 248,-374.109 248,-337.597 219.83,-304.936 196.312,-284.031"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="198.437,-281.243 188.562,-277.404 193.888,-286.563 198.437,-281.243"/>
-</g>
-<!-- t15 -->
-<g id="node14" class="node"><title>t15</title>
-<polygon fill="#efa03b" stroke="black" points="1391.34,-451.109 1170.66,-451.109 1166.66,-447.109 1166.66,-415.109 1387.34,-415.109 1391.34,-419.109 1391.34,-451.109"/>
-<polyline fill="none" stroke="black" points="1387.34,-447.109 1166.66,-447.109 "/>
-<polyline fill="none" stroke="black" points="1387.34,-447.109 1387.34,-415.109 "/>
-<polyline fill="none" stroke="black" points="1387.34,-447.109 1391.34,-451.109 "/>
-<text text-anchor="middle" x="1279" y="-427.109" font-family="Times,serif" font-size="20.00">skip_page</text>
-</g>
-<!-- t3&#45;&gt;t15 -->
-<g id="edge19" class="edge"><title>t3&#45;&gt;t15</title>
-<path fill="none" stroke="#0044a0" d="M1060.32,-531.078C1078.95,-524.591 1100.07,-516.881 1119,-509.109 1159.3,-492.568 1204.11,-471.304 1236.04,-455.645"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1237.67,-458.74 1245.1,-451.181 1234.58,-452.461 1237.67,-458.74"/>
-</g>
-<!-- t13 -->
-<g id="node15" class="node"><title>t13</title>
-<polygon fill="#efa03b" stroke="black" points="1388.24,-277.109 1007.76,-277.109 1003.76,-273.109 1003.76,-241.109 1384.24,-241.109 1388.24,-245.109 1388.24,-277.109"/>
-<polyline fill="none" stroke="black" points="1384.24,-273.109 1003.76,-273.109 "/>
-<polyline fill="none" stroke="black" points="1384.24,-273.109 1384.24,-241.109 "/>
-<polyline fill="none" stroke="black" points="1384.24,-273.109 1388.24,-277.109 "/>
-<text text-anchor="middle" x="1196" y="-253.109" font-family="Times,serif" font-size="20.00">tesseract_ocr_and_render_pdf</text>
-</g>
-<!-- t3&#45;&gt;t13 -->
-<g id="edge21" class="edge"><title>t3&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M1040.79,-531.094C1071.69,-511.983 1114,-477.834 1114,-434.109 1114,-434.109 1114,-434.109 1114,-374.109 1114,-337.749 1141.83,-305.063 1165.07,-284.11"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1167.46,-286.664 1172.72,-277.466 1162.87,-281.377 1167.46,-286.664"/>
-</g>
-<!-- t5 -->
-<g id="node6" class="node"><title>t5</title>
-<polygon fill="#efa03b" stroke="black" points="830.742,-451.109 535.258,-451.109 531.258,-447.109 531.258,-415.109 826.742,-415.109 830.742,-419.109 830.742,-451.109"/>
-<polyline fill="none" stroke="black" points="826.742,-447.109 531.258,-447.109 "/>
-<polyline fill="none" stroke="black" points="826.742,-447.109 826.742,-415.109 "/>
-<polyline fill="none" stroke="black" points="826.742,-447.109 830.742,-451.109 "/>
-<text text-anchor="middle" x="681" y="-427.109" font-family="Times,serif" font-size="20.00">preprocess_deskew</text>
-</g>
-<!-- t4&#45;&gt;t5 -->
-<g id="edge6" class="edge"><title>t4&#45;&gt;t5</title>
-<path fill="none" stroke="#0044a0" d="M798.376,-473.003C780.521,-467.087 760.346,-460.402 741.823,-454.264"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="742.901,-450.934 732.308,-451.111 740.699,-457.579 742.901,-450.934"/>
-</g>
-<!-- t8 -->
-<g id="node10" class="node"><title>t8</title>
-<polygon fill="#efa03b" stroke="black" points="980.119,-335.109 665.881,-335.109 661.881,-331.109 661.881,-299.109 976.119,-299.109 980.119,-303.109 980.119,-335.109"/>
-<polyline fill="none" stroke="black" points="976.119,-331.109 661.881,-331.109 "/>
-<polyline fill="none" stroke="black" points="976.119,-331.109 976.119,-299.109 "/>
-<polyline fill="none" stroke="black" points="976.119,-331.109 980.119,-335.109 "/>
-<text text-anchor="middle" x="821" y="-311.109" font-family="Times,serif" font-size="20.00">select_image_for_pdf</text>
-</g>
-<!-- t4&#45;&gt;t8 -->
-<g id="edge12" class="edge"><title>t4&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M848.03,-473.087C846.196,-457.857 843.261,-434.941 840,-415.109 836.1,-391.393 830.78,-364.592 826.761,-345.181"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="830.152,-344.296 824.68,-335.224 823.3,-345.728 830.152,-344.296"/>
-</g>
-<!-- t6 -->
-<g id="node7" class="node"><title>t6</title>
-<polygon fill="#efa03b" stroke="black" points="801.95,-393.109 524.05,-393.109 520.05,-389.109 520.05,-357.109 797.95,-357.109 801.95,-361.109 801.95,-393.109"/>
-<polyline fill="none" stroke="black" points="797.95,-389.109 520.05,-389.109 "/>
-<polyline fill="none" stroke="black" points="797.95,-389.109 797.95,-357.109 "/>
-<polyline fill="none" stroke="black" points="797.95,-389.109 801.95,-393.109 "/>
-<text text-anchor="middle" x="661" y="-369.109" font-family="Times,serif" font-size="20.00">preprocess_clean</text>
-</g>
-<!-- t5&#45;&gt;t6 -->
-<g id="edge7" class="edge"><title>t5&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M674.891,-415.003C673.495,-411.095 671.979,-406.852 670.48,-402.652"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="673.731,-401.351 667.072,-393.111 667.139,-403.705 673.731,-401.351"/>
-</g>
-<!-- t5&#45;&gt;t8 -->
-<g id="edge11" class="edge"><title>t5&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M781.944,-415.046C792.967,-409.579 803.051,-402.434 811,-393.109 822.009,-380.195 824.454,-361.215 824.174,-345.581"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="827.659,-345.2 823.617,-335.407 820.669,-345.583 827.659,-345.2"/>
-</g>
-<!-- t7 -->
-<g id="node8" class="node"><title>t7</title>
-<polygon fill="#efa03b" stroke="black" points="643.666,-335.109 354.334,-335.109 350.334,-331.109 350.334,-299.109 639.666,-299.109 643.666,-303.109 643.666,-335.109"/>
-<polyline fill="none" stroke="black" points="639.666,-331.109 350.334,-331.109 "/>
-<polyline fill="none" stroke="black" points="639.666,-331.109 639.666,-299.109 "/>
-<polyline fill="none" stroke="black" points="639.666,-331.109 643.666,-335.109 "/>
-<text text-anchor="middle" x="497" y="-311.109" font-family="Times,serif" font-size="20.00">ocr_tesseract_hocr</text>
-</g>
-<!-- t6&#45;&gt;t7 -->
-<g id="edge8" class="edge"><title>t6&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M610.903,-357.003C593.736,-351.141 574.358,-344.524 556.518,-338.433"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="557.384,-335.03 546.79,-335.111 555.122,-341.654 557.384,-335.03"/>
-</g>
-<!-- t6&#45;&gt;t8 -->
-<g id="edge10" class="edge"><title>t6&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M709.875,-357.003C726.624,-351.141 745.529,-344.524 762.933,-338.433"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="764.142,-341.718 772.425,-335.111 761.83,-335.111 764.142,-341.718"/>
-</g>
-<!-- t10 -->
-<g id="node9" class="node"><title>t10</title>
-<polygon fill="#efa03b" stroke="black" points="620.109,-277.109 335.891,-277.109 331.891,-273.109 331.891,-241.109 616.109,-241.109 620.109,-245.109 620.109,-277.109"/>
-<polyline fill="none" stroke="black" points="616.109,-273.109 331.891,-273.109 "/>
-<polyline fill="none" stroke="black" points="616.109,-273.109 616.109,-241.109 "/>
-<polyline fill="none" stroke="black" points="616.109,-273.109 620.109,-277.109 "/>
-<text text-anchor="middle" x="476" y="-253.109" font-family="Times,serif" font-size="20.00">render_hocr_page</text>
-</g>
-<!-- t7&#45;&gt;t10 -->
-<g id="edge9" class="edge"><title>t7&#45;&gt;t10</title>
-<path fill="none" stroke="#0044a0" d="M490.585,-299.003C489.12,-295.095 487.528,-290.852 485.953,-286.652"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="489.164,-285.245 482.376,-277.111 482.61,-287.703 489.164,-285.245"/>
-</g>
-<!-- t11 -->
-<g id="node13" class="node"><title>t11</title>
-<polygon fill="#efa03b" stroke="black" points="985.486,-277.109 642.514,-277.109 638.514,-273.109 638.514,-241.109 981.486,-241.109 985.486,-245.109 985.486,-277.109"/>
-<polyline fill="none" stroke="black" points="981.486,-273.109 638.514,-273.109 "/>
-<polyline fill="none" stroke="black" points="981.486,-273.109 981.486,-241.109 "/>
-<polyline fill="none" stroke="black" points="981.486,-273.109 985.486,-277.109 "/>
-<text text-anchor="middle" x="812" y="-253.109" font-family="Times,serif" font-size="20.00">render_hocr_debug_page</text>
-</g>
-<!-- t7&#45;&gt;t11 -->
-<g id="edge18" class="edge"><title>t7&#45;&gt;t11</title>
-<path fill="none" stroke="#0044a0" d="M592.793,-299.079C628.531,-292.726 669.303,-285.478 705.912,-278.969"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="706.887,-282.351 716.12,-277.155 705.661,-275.459 706.887,-282.351"/>
-</g>
-<!-- t12 -->
-<g id="node12" class="node"><title>t12</title>
-<polygon fill="#efa03b" stroke="black" points="607.082,-219.109 348.918,-219.109 344.918,-215.109 344.918,-183.109 603.082,-183.109 607.082,-187.109 607.082,-219.109"/>
-<polyline fill="none" stroke="black" points="603.082,-215.109 344.918,-215.109 "/>
-<polyline fill="none" stroke="black" points="603.082,-215.109 603.082,-183.109 "/>
-<polyline fill="none" stroke="black" points="603.082,-215.109 607.082,-219.109 "/>
-<text text-anchor="middle" x="476" y="-195.109" font-family="Times,serif" font-size="20.00">add_text_layer</text>
-</g>
-<!-- t10&#45;&gt;t12 -->
-<g id="edge15" class="edge"><title>t10&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M476,-241.003C476,-237.312 476,-233.322 476,-229.352"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="479.5,-229.111 476,-219.111 472.5,-229.111 479.5,-229.111"/>
-</g>
-<!-- t8&#45;&gt;t9 -->
-<g id="edge13" class="edge"><title>t8&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M661.842,-299.864C658.871,-299.606 655.921,-299.354 653,-299.109 509.812,-287.095 471.159,-289.734 324.542,-277.219"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="324.526,-273.705 314.262,-276.332 323.924,-280.679 324.526,-273.705"/>
-</g>
-<!-- t8&#45;&gt;t11 -->
-<g id="edge17" class="edge"><title>t8&#45;&gt;t11</title>
-<path fill="none" stroke="#0044a0" d="M818.251,-299.003C817.64,-295.204 816.979,-291.087 816.322,-287.002"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="819.775,-286.429 814.732,-277.111 812.864,-287.539 819.775,-286.429"/>
-</g>
-<!-- t8&#45;&gt;t13 -->
-<g id="edge20" class="edge"><title>t8&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M935.04,-299.079C978.311,-292.618 1027.78,-285.23 1071.94,-278.636"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1072.48,-282.093 1081.86,-277.155 1071.45,-275.17 1072.48,-282.093"/>
-</g>
-<!-- t9&#45;&gt;t12 -->
-<g id="edge16" class="edge"><title>t9&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M259.577,-241.079C294.86,-234.726 335.115,-227.478 371.259,-220.969"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="372.116,-224.372 381.337,-219.155 370.875,-217.482 372.116,-224.372"/>
-</g>
-<!-- t16 -->
-<g id="node17" class="node"><title>t16</title>
-<polygon fill="#efa03b" stroke="black" points="1031.47,-105.333 1196,-78.005 1360.53,-105.333 1360.37,-149.551 1031.63,-149.551 1031.47,-105.333"/>
-<polygon fill="none" stroke="black" points="1027.46,-101.94 1196,-73.9453 1364.54,-101.94 1364.36,-153.556 1027.64,-153.556 1027.46,-101.94"/>
-<text text-anchor="middle" x="1196" y="-111.555" font-family="Times,serif" font-size="20.00">merge_pages</text>
-</g>
-<!-- t12&#45;&gt;t16 -->
-<g id="edge26" class="edge"><title>t12&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M607.222,-185.246C722.142,-172.229 890.172,-153.196 1017.34,-138.791"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1018.04,-142.235 1027.58,-137.632 1017.25,-135.279 1018.04,-142.235"/>
-</g>
-<!-- t11&#45;&gt;t16 -->
-<g id="edge27" class="edge"><title>t11&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M859.086,-240.997C916.223,-220.232 1014.29,-184.591 1089.94,-157.099"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1091.32,-160.323 1099.52,-153.618 1088.92,-153.744 1091.32,-160.323"/>
-</g>
-<!-- t15&#45;&gt;t16 -->
-<g id="edge23" class="edge"><title>t15&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M1321.98,-414.938C1359.16,-396.916 1407,-364.681 1407,-318.109 1407,-318.109 1407,-318.109 1407,-258.109 1407,-211.469 1370.69,-179.509 1328.2,-158.076"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1329.57,-154.854 1319.04,-153.66 1326.53,-161.158 1329.57,-154.854"/>
-</g>
-<!-- t13&#45;&gt;t16 -->
-<g id="edge25" class="edge"><title>t13&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M1196,-240.997C1196,-221.914 1196,-190.265 1196,-163.926"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1199.5,-163.759 1196,-153.759 1192.5,-163.759 1199.5,-163.759"/>
-</g>
-<!-- t14&#45;&gt;t16 -->
-<g id="edge24" class="edge"><title>t14&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M1326.64,-588.956C1373.58,-572.748 1427,-543.103 1427,-492.109 1427,-492.109 1427,-492.109 1427,-258.109 1427,-220.539 1416.55,-207.537 1388,-183.109 1376.82,-173.542 1364.16,-165.377 1350.82,-158.413"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1351.99,-155.087 1341.47,-153.781 1348.88,-161.358 1351.99,-155.087"/>
-</g>
-<!-- t17 -->
-<g id="node18" class="node"><title>t17</title>
-<polygon fill="#efa03b" stroke="black" points="1310.18,-52 1085.82,-52 1081.82,-48 1081.82,-16 1306.18,-16 1310.18,-20 1310.18,-52"/>
-<polyline fill="none" stroke="black" points="1306.18,-48 1081.82,-48 "/>
-<polyline fill="none" stroke="black" points="1306.18,-48 1306.18,-16 "/>
-<polyline fill="none" stroke="black" points="1306.18,-48 1310.18,-52 "/>
-<text text-anchor="middle" x="1196" y="-28" font-family="Times,serif" font-size="20.00">copy_final</text>
-</g>
-<!-- t16&#45;&gt;t17 -->
-<g id="edge28" class="edge"><title>t16&#45;&gt;t17</title>
-<path fill="none" stroke="#0044a0" d="M1196,-73.8665C1196,-69.8921 1196,-65.942 1196,-62.1676"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1199.5,-62.1213 1196,-52.1214 1192.5,-62.1214 1199.5,-62.1213"/>
-</g>
-</g>
-</svg>

pytest.ini

@@ -1,2 +0,0 @@
-[pytest]
-norecursedirs = lib 

requirements.txt

@@ -1,6 +1,9 @@
-ruffus==2.6.3
-Pillow==3.1.1
-reportlab==3.2.0
-PyPDF2==1.25.1
-img2pdf==0.2
-cffi==1.5.2
+# 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 == 2.6.3
+Pillow == 5.0.0
+reportlab == 3.4.0
+PyPDF2 == 1.26.0
+img2pdf == 0.2.4
+cffi == 1.11.2

setup.cfg

@@ -1,4 +1,14 @@
 [bdist_wheel]
-python-tag = py34
+python-tag = py35
+
 [aliases]
-test=pytest
+test=pytest
+
+[check-manifest]
+ignore =
+	.github
+
+[tool:pytest]
+norecursedirs = lib .pc .git output cache resources
+testpaths = tests
+addopts = -n auto

setup.py

@@ -5,11 +5,11 @@
 from __future__ import print_function, unicode_literals
 
 import sys
-if sys.version_info < (3, 4):
-    print("Python 3.4 or newer is required")
+if sys.version_info < (3, 5):
+    print("Python 3.5 or newer is required", file=sys.stderr)
     sys.exit(1)
 
-from setuptools import setup  # nopep8
+from setuptools import setup, find_packages  # nopep8
 from subprocess import STDOUT, check_output, CalledProcessError  # nopep8
 from collections.abc import Mapping  # nopep8
 import re  # nopep8
@@ -148,32 +148,35 @@ 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 \
+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='5.1.1',   # limited by Travis CI / Ubuntu 14.04 backports
         package='qpdf',
         version_check_args=['--version']
     )
+else:
+    print('Skipping external program tests because of --force')
 
 
 if 'upload' in sys.argv[1:]:
@@ -182,23 +185,30 @@ if 'upload' in sys.argv[1:]:
 
 tests_require = open('test_requirements.txt').read().splitlines()
 
+
+def readme():
+    with open('README.rst') as f:
+        return f.read()
+
 setup(
     name='ocrmypdf',
     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'],
+    license='MIT',
+    packages=find_packages(exclude=["tests", "tests.*"]),
     keywords=['PDF', 'OCR', 'optical character recognition', 'PDF/A', 'scanning'],
     classifiers=[
-        "Programming Language :: Python :: 3",
+        "Programming Language :: Python :: 3.5",
+        "Programming Language :: Python :: 3.6",
         "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 :: MIT License",
         "Operating System :: MacOS :: MacOS X",
         "Operating System :: POSIX",
         "Operating System :: POSIX :: BSD",
@@ -207,29 +217,30 @@ setup(
         "Topic :: Text Processing :: Indexing",
         "Topic :: Text Processing :: Linguistic",
         ],
+    python_requires='>=3.5',
     setup_requires=[
-        'setuptools_scm',
-        'cffi>=1.5.0',
-        'pytest-runner'
+        'setuptools_scm',       # so that version will work
+        'cffi>=1.9.1',          # to build the leptonica module
+        'pytest-runner'         # to enable python setup.py test
     ],
     use_scm_version={'version_scheme': 'post-release'},
     cffi_modules=[
-        'ocrmypdf/lib/compile_leptonica.py:ffi'
+        'ocrmypdf/lib/compile_leptonica.py:ffibuilder'
     ],
     install_requires=[
-        'ruffus==2.6.3',
-        'Pillow>=3.0.0',
-        'reportlab>=3.1.44',
-        'PyPDF2>=1.25.1',
-        'img2pdf>=0.2.0',
-        'cffi>=1.5.0'
+        'ruffus==2.6.3',        # pinned - ocrmypdf implements a 2.6.3 workaround
+        'Pillow>=4.0.0',        # Pillow < 4 has BytesIO/TIFF bug w/img2pdf 0.2.3
+        'reportlab>=3.3.0',     # oldest released version with sane image handling
+        'PyPDF2>=1.26',         # pure Python, so track HEAD closely
+        'img2pdf>=0.2.3',       # pure Python, so track HEAD closely
+        'cffi>=1.9.1'           # must be a setup and install requirement
     ],
     tests_require=tests_require,
     entry_points={
         'console_scripts': [
-            'ocrmypdf = ocrmypdf.main:run_pipeline'
+            'ocrmypdf = ocrmypdf.__main__:run_pipeline'
         ],
     },
-    package_data={'ocrmypdf': ['data/sRGB_IEC61966-2-1_black_scaled.icc']},
+    package_data={'ocrmypdf': ['data/sRGB.icc']},
     include_package_data=True,
     zip_safe=False)

share/etc-apt-sources.list

@@ -1,4 +0,0 @@
-deb http://httpredir.debian.org/debian stretch main
-deb http://httpredir.debian.org/debian stretch-updates main
-deb http://security.debian.org stretch/updates main
-deb http://ftp.de.debian.org/debian sid main contrib non-free

test_requirements.txt

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

tests/conftest.py

@@ -0,0 +1,147 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+import platform
+
+pytest_plugins = ['helpers_namespace']
+
+import pytest
+from pathlib import Path
+from subprocess import Popen, PIPE
+
+
+if sys.version_info.major < 3:
+    print("Requires Python 3.4+")
+    sys.exit(1)
+
+
+@pytest.helpers.register
+def is_linux():
+    return platform.system() == 'Linux'
+
+
+@pytest.helpers.register
+def is_macos():
+    return platform.system() == 'Darwin'
+
+
+@pytest.helpers.register
+def running_in_docker():
+    # Docker creates a file named /.dockerenv (newer versions) or
+    # /.dockerinit (older) -- this is undocumented, not an offical test
+    return os.path.exists('/.dockerenv') or os.path.exists('/.dockerinit')
+
+
+@pytest.helpers.register
+def running_in_travis():
+    return os.environ.get('TRAVIS') == 'true'
+
+
+TESTS_ROOT = os.path.abspath(os.path.dirname(__file__))
+SPOOF_PATH = os.path.join(TESTS_ROOT, 'spoof')
+PROJECT_ROOT = os.path.dirname(TESTS_ROOT)
+OCRMYPDF = [sys.executable, '-m', 'ocrmypdf']
+
+
+@pytest.helpers.register
+def spoof(**kwargs):
+    """Modify environment variables to override subprocess executables
+
+    spoof(program1='replacement', ...)
+
+    Before running any executable, ocrmypdf checks the environment variable
+    OCRMYPDF_PROGRAMNAME to override default program name/location, e.g.
+    OCRMYPDF_GS redirects from the system path Ghostscript ("gs") to elsewhere.
+
+    """
+    env = os.environ.copy()
+
+    for replace_program, with_spoof in kwargs.items():
+        spoofer = os.path.join(SPOOF_PATH, with_spoof)
+        if not os.access(spoofer, os.X_OK):
+            os.chmod(spoofer, 0o755)
+        env['OCRMYPDF_' + replace_program.upper()] = spoofer
+    return env
+
+
+@pytest.fixture
+def spoof_tesseract_noop():
+    return spoof(tesseract='tesseract_noop.py')
+
+
+@pytest.fixture
+def spoof_tesseract_cache():
+    if running_in_docker():
+        return os.environ.copy()
+    return spoof(tesseract="tesseract_cache.py")
+
+
+@pytest.fixture
+def resources():
+    return Path(TESTS_ROOT) / 'resources'
+
+
+@pytest.fixture
+def ocrmypdf_exec():
+    return OCRMYPDF
+
+
+@pytest.fixture(scope="function")
+def outdir(tmpdir):
+    return Path(str(tmpdir))
+
+
+@pytest.fixture(scope="function")
+def outpdf(tmpdir):
+    return str(Path(str(tmpdir)) / 'out.pdf')
+
+
+@pytest.fixture(scope="function")
+def no_outpdf(tmpdir):
+    """This just documents the fact that a test is not expected to produce
+    output. Unfortunately an assertion failure inside a test fixture produces
+    an error rather than a test failure, so no testing is done. It's up to
+    the test to confirm that no output file was created."""
+    return str(Path(str(tmpdir)) / 'no_output.pdf')
+
+
+@pytest.helpers.register
+def check_ocrmypdf(input_file, output_file, *args, env=None):
+    "Run ocrmypdf and confirmed that a valid file was created"
+
+    p, out, err = run_ocrmypdf(input_file, output_file, *args, env=env)
+    #print(err)  # ensure py.test collects the output, use -s to view
+    assert p.returncode == 0, "<stderr>\n" + err + "\n</stderr>"
+    assert os.path.exists(str(output_file)), "Output file not created"
+    assert os.stat(str(output_file)).st_size > 100, "PDF too small or empty"
+    assert out == "", \
+        "The following was written to stdout and should not have been: \n" + \
+        "<stdout>\n" + out + "\n</stdout>"
+    return output_file
+
+
+@pytest.helpers.register
+def run_ocrmypdf(input_file, output_file, *args, env=None):
+    "Run ocrmypdf and let caller deal with results"
+
+    if env is None:
+        env = os.environ
+
+    p_args = OCRMYPDF + [str(arg) for arg in args] + \
+             [str(input_file), str(output_file)]
+    p = Popen(
+        p_args, close_fds=True, stdout=PIPE, stderr=PIPE,
+        universal_newlines=True, env=env)
+    out, err = p.communicate()
+    #print(err)
+
+    return p, out, err
+
+
+@pytest.helpers.register
+def first_page_dimensions(pdf):
+    from ocrmypdf import pdfinfo
+    info = pdfinfo.PdfInfo(pdf)
+    page0 = info[0]
+    return (page0.width_inches, page0.height_inches)

tests/resources/README.rst

@@ -9,21 +9,35 @@ Files derived from free sources
 These test resources come from free sources, under either public domain or Creative Commons licenses.
 In some cases they were converted from one image format to another without other changes.
 
-+---------------------+--------------------------------------------------------------------------------+
-| File                | Source                                                                         |
-+=====================+================================================================================+
-| c02-22.pdf          | `Project Gutenberg`_, Adventures of Huckleberry Finn, page 22                  |
-+---------------------+--------------------------------------------------------------------------------+
-| congress.jpg        | `US Congressional Records`_                                                    |
-+---------------------+--------------------------------------------------------------------------------+
-| graph.pdf           | `Wikimedia: Pandas text analysis.png`_                                         |
-+---------------------+--------------------------------------------------------------------------------+
-| lichtenstein.pdf    | `Wikimedia: JPEG2000 Lichtenstein`_ (Creative Commons BY-SA 3.0)               |
-+---------------------+--------------------------------------------------------------------------------+
-| LinnSequencer.jpg,  | `Wikimedia: LinnSequencer`_ (Creative Commons Attribution-ShareAlike 3.0)      |
-| linn.pdf, linn.txt  |                                                                                |
-+---------------------+--------------------------------------------------------------------------------+
- 
+.. list-table:: 
+    :widths: 20 50 30
+    :header-rows: 1
+
+    *   - File
+        - Source
+        - License
+    *   - c02-22.pdf
+        - `Project Gutenberg`_, Adventures of Huckleberry Finn, page 22
+        - Public Domain
+    *   - congress.jpg
+        - `US Congressional Records`_
+        - Public Domain
+    *   - graph.pdf
+        - `Wikimedia: Pandas text analysis.png`_
+        - Public Domain
+    *   - lichtenstein.pdf
+        - `Wikimedia: JPEG2000 Lichtenstein`_
+        - Creative Commons BY-SA 3.0
+    *   - LinnSequencer.jpg, linn.pdf, linn.txt
+        - `Wikimedia: LinnSequencer`_
+        - Creative Commons BY-SA 3.0
+    *   - typewriter.png, 2400dpi.pdf
+        - `Wikimedia: Triumph typewrtier text Linzensoep`_
+        - Creative Commons BY-SA 2.5
+    *   - baiona.png
+        - `Wikimedia: Baionako udalerri mugakideak`_
+        - Creative Commons BY-SA 4.0
+
 
 Files generated for this project
 ================================
@@ -31,28 +45,84 @@ Files generated for this project
 The following test resources were crafted specifically for this project, and can be used
 under the terms of the license in LICENSE.rst.
 
-- blank.pdf (a blank PDF page)
-- cmyk.pdf (a CMYK image created in Photoshop)
-- enormous.pdf (a very lage page)
-- francais.pdf (a page containing French accented characters)
-- hugemono.pdf (large monochrome JBIG2 page with pixel dimensions of 35000x35000)
-- invalid.pdf (a PDF file header followed by EOF marker)
-- missing_docinfo.pdf (PDF file with no /DocumentInfo section)
+.. list-table:: 
+    :widths: 20 20 60
+    :header-rows: 1
+
+    *   - File
+        - Contributor
+        - Purpose
+    *   - aspect.pdf
+        - @jbarlow83
+        - test image with 200 x 100 DPI resolution
+    *   - blank.pdf
+        - @jbarlow83
+        - blank PDF generated by Adobe Illustrator CC 17, containing a lot of application-specific metadata/bloat
+    *   - cmyk.pdf
+        - @jbarlow83
+        - a CMYK image created in Photoshop
+    *   - crom.png
+        - @jbarlow83
+        - test for non-dictionary words        
+    *   - enormous.pdf
+        - @jbarlow83
+        - very large PDF page
+    *   - epson.pdf
+        - @lowesjam
+        - a linearized PDF containing some unusual indirect objects, created by an Epson printer; printout of a Wikipedia article (CC BY-SA)
+    *   - formxobject.pdf
+        - @jbarlow83
+        - hand-crafted PDF containing an image inside a Form XObject
+    *   - francais.pdf
+        - @jbarlow83
+        - a page containing French accents (diacritics)  
+    *   - hugemono.pdf
+        - @jbarlow83
+        - large monochrome 35000x35000 image in JBIG2 encoding 
+    *   - invalid.pdf
+        - @jbarlow83
+        - a PDF file header followed by EOF marker
+    *   - masks.pdf
+        - @supergrobi
+        - file containing explicit masks and a stencil mask drawn without a proper transformation matrix; printout of a German Wikipedia article (CC BY-SA)
+    *   - missing_docinfo.pdf
+        - @jbarlow83
+        - PDF file with no /DocumentInfo section 
+    *   - overlay.pdf
+        - @maxandersen
+        - PDF file generated by PDFPen pro that triggered content stream parse errors
+    *   - negzero.pdf
+        - @jbarlow83
+        - copy of formxobject.pdf with token that qpdf doesn't like
+    *   - no_contents.pdf
+        - @jbarlow83
+        - synthetic PDF with a blank page that has no /Contents entry
+    *   - trivial.pdf
+        - @jbarlow83
+        - smallest possible valid PDF-1.3 with all required fields
+    *   - vector.pdf
+        - @Catscratch
+        - a PDF with vector art and text rendered as curves with no fonts
 
 
 Assemblies
 ==========
 
-These test resources are assemblies from other previously mentioned files, released under the same license terms as their input files.
+These test resources are assemblies or derivatives from other previously mentioned files, released under the same license terms as their input files.
 
-- cardinal.pdf (four cardinal directions, rotated copies of LinnSequencer.jpg)
+- cardinal.pdf (four cardinal directions, baked-in rotated copies of LinnSequencer.jpg)
 - ccitt.pdf (LinnSequencer.jpg, converted to CCITT encoding)
+- encrypted_algo4.pdf (congress.jpg, encrypted with algorithm 4 - not supported by PyPDF2)
 - graph_ocred.pdf (from graph.pdf)
 - jbig2.pdf (congress.jpg, converted to JBIG2 encoding)
 - multipage.pdf (from several other files)
 - palette.pdf (congress.jpg, converted to a 256-color palette)
-- skew.pdf (from c02-22.pdf)
-- skew-encrypted.pdf (skew.pdf with encrypted applied)
+- rotated_skew.pdf (a /Rotate'd and skewed document from LinnSequencer.jpg)
+- skew.pdf (from LinnSequencer.jpg, skew simulated by adjusting the transformation matrix)
+- skew-encrypted.pdf (skew.pdf with encryption - access supported by PyPDF2, password is "password")
+- baiona_gray.png (from baiona.png)
+- poster.pdf (from LinnSequencer.jpg)
+
 
 
 .. _`Wikimedia: LinnSequencer`: https://upload.wikimedia.org/wikipedia/en/b/b7/LinnSequencer_hardware_MIDI_sequencer_brochure_page_2_300dpi.jpg
@@ -63,4 +133,10 @@ These test resources are assemblies from other previously mentioned files, relea
 
 .. _`Wikimedia: Pandas text analysis.png`: https://en.wikipedia.org/wiki/File:Pandas_text_analysis.png
 
-.. _`Wikimedia: JPEG2000 Lichtenstein`: https://en.wikipedia.org/wiki/JPEG_2000#/media/File:Jpeg2000_2-level_wavelet_transform-lichtenstein.png
+.. _`Wikimedia: JPEG2000 Lichtenstein`: https://en.wikipedia.org/wiki/JPEG_2000#/media/File:Jpeg2000_2-level_wavelet_transform-lichtenstein.png
+
+.. _`Linux (Wikipedia Article)`: https://de.wikipedia.org/wiki/Linux 
+
+.. _`Wikimedia: Triumph typewrtier text Linzensoep`: https://commons.wikimedia.org/wiki/File:Triumph.typewriter_text_Linzensoep.gif
+
+.. _`Wikimedia: Baionako udalerri mugakideak`: https://commons.wikimedia.org/wiki/File:Baionako_udalerri_mugakideak.png

tests/resources/no_contents.pdf

@@ -0,0 +1,21 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 2 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /Count 1 /Kids [ 3 0 R ] /Type /Pages >>
+endobj
+3 0 obj
+<< /MediaBox [ 0 0 720 720 ] /Parent 2 0 R /Resources << >> /Type /Page >>
+endobj
+xref
+0 4
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+trailer << /Root 1 0 R /Size 4 /ID [<52bba3c78160d0c6e851b59110e5d076><52bba3c78160d0c6e851b59110e5d076>] >>
+startxref
+213
+%%EOF

tests/resources/trivial.pdf

@@ -0,0 +1,27 @@
+%PDF-1.3
+%¿÷¢þ
+1 0 obj
+<< /Pages 2 0 R /Type /Catalog >>
+endobj
+2 0 obj
+<< /Count 1 /Kids [ 3 0 R ] /Type /Pages >>
+endobj
+3 0 obj
+<< /Contents 4 0 R /MediaBox [ 0 0 300 300 ] /Parent 2 0 R /Resources << >> /Type /Page >>
+endobj
+4 0 obj
+<< /Length 0 >>
+stream
+endstream
+endobj
+xref
+0 5
+0000000000 65535 f 
+0000000015 00000 n 
+0000000064 00000 n 
+0000000123 00000 n 
+0000000229 00000 n 
+trailer << /Root 1 0 R /Size 5 /ID [<8088c23e0edc07ef4fbed3daa55f52ed><8088c23e0edc07ef4fbed3daa55f52ed>] >>
+startxref
+277
+%%EOF

tests/spoof/gs_feature_elision.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+from subprocess import check_call
+
+"""Replicate one type of Ghostscript feature elision warning during
+PDF/A creation."""
+
+
+def real_ghostscript(argv):
+    gs_args = ['gs'] + argv[1:]
+    os.execvp("gs", gs_args)
+    return  # Not reachable
+
+
+elision_warning = """GPL Ghostscript 9.20: Setting Overprint Mode to 1
+not permitted in PDF/A-2, overprint mode not set"""
+
+
+def main():
+    if '--version' in sys.argv:
+        print('9.20')
+        print('SPOOFED: ' + os.path.basename(__file__))
+        sys.exit(0)
+
+    gs_args = ['gs'] + sys.argv[1:]
+    check_call(gs_args)
+
+    if '-sDEVICE=pdfwrite' in sys.argv[1:]:
+        print(elision_warning)
+
+    sys.exit(0)
+
+if __name__ == '__main__':
+    main()

tests/spoof/gs_pdfa_failure.py

@@ -0,0 +1,44 @@
+#!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+
+"""Replicate Ghostscript PDF/A conversion failure by suppressing some
+arguments"""
+
+
+def real_ghostscript(argv):
+    gs_args = ['gs'] + argv[1:]
+    os.execvp("gs", gs_args)
+    return  # Not reachable
+
+
+def main():
+    if '--version' in sys.argv:
+        print('9.20')
+        print('SPOOFED: ' + os.path.basename(__file__))
+        sys.exit(0)
+
+    # Unless some argument is calling for PDFA generation, forward to
+    # real ghostscript
+    if not any(arg.startswith('-dPDFA') for arg in sys.argv):
+        real_ghostscript(sys.argv)
+        return
+
+    # Remove the two arguments that tell ghostscript to create a PDF/A
+    # Does not remove the Postscript definition file - not necessary
+    # to cause PDF/A creation failure
+    argv = []
+    for arg in sys.argv:
+        if arg.startswith('-dPDFA'):
+            continue
+        elif arg.startswith('-dPDFACompatibilityPolicy'):
+            continue
+        argv.append(arg)
+
+    real_ghostscript(argv)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/gs_raster_failure.py

@@ -0,0 +1,33 @@
+#!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+
+"""Replicate Ghostscript raster failure while allowing rendering"""
+
+
+def real_ghostscript(argv):
+    gs_args = ['gs'] + argv[1:]
+    os.execvp("gs", gs_args)
+    return  # Not reachable
+
+
+def main():
+    if '--version' in sys.argv:
+        print('9.20')
+        print('SPOOFED: ' + os.path.basename(__file__))
+        sys.exit(0)
+
+    # For any rendering calls (device == pdfwrite) call real ghostscript
+    if '-sDEVICE=pdfwrite' in sys.argv:
+        real_ghostscript(sys.argv)
+        return
+
+    # Fail
+    print("ERROR: Ghost story archive not found")
+    sys.exit(1)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/gs_render_failure.py

@@ -0,0 +1,33 @@
+#!/usr/bin/env python3
+# © 2016-18 James R. Barlow: github.com/jbarlow83
+
+import sys
+import os
+
+"""Replicate Ghostscript render failure while allowing rasterizing"""
+
+
+def real_ghostscript(argv):
+    gs_args = ['gs'] + argv[1:]
+    os.execvp("gs", gs_args)
+    return  # Not reachable
+
+
+def main():
+    if '--version' in sys.argv:
+        print('9.20')
+        print('SPOOFED: ' + os.path.basename(__file__))
+        sys.exit(0)
+
+    # For any rasterize calls (device != pdfwrite) call real ghostscript
+    if '-sDEVICE=pdfwrite' not in sys.argv:
+        real_ghostscript(sys.argv)
+        return
+
+    # Fail
+    print("ERROR: Casper is not a friendly ghost")
+    sys.exit(1)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/qpdf_dummy_return2.py

@@ -1,8 +1,13 @@
 #!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
 import sys
 
 
 def main():
+    if sys.argv[-1] == '--version':
+        print('qpdf version 7.0.0')
+        sys.exit(0)
     print('qpdf dummy')
     sys.exit(2)
 

tests/spoof/tesseract_badutf8.py

@@ -0,0 +1,58 @@
+#!/usr/bin/env python3
+# © 2017 James R. Barlow: github.com/jbarlow83
+import sys
+import img2pdf
+import PyPDF2 as pypdf
+from PIL import Image
+
+
+"""Tesseract bad utf8 spoof
+
+In 'hocr' mode or 'pdf' mode, return error code 1 and some non-Unicode
+text because tesseract seems to do that in some cases related to 
+language pack version mismatches
+
+"""
+
+
+VERSION_STRING = '''tesseract 3.05.01
+ leptonica-1.72
+  libjpeg 8d : libpng 1.6.19 : libtiff 4.0.6 : zlib 1.2.5
+SPOOFED
+'''
+
+# Japanese "Invalid UTF-8" encoded in Shift JIS
+BAD_UTF8 = b'\x96\xb3\x8c\xf8\x82\xc8UTF-8\x0a'
+
+
+def main():
+    if sys.argv[1] == '--version':
+        print(VERSION_STRING, file=sys.stderr)
+        sys.exit(0)
+    elif sys.argv[1] == '--list-langs':
+        print('List of available languages (1):\neng', file=sys.stderr)
+        sys.exit(0)
+    elif sys.argv[1] == '--print-parameters':
+        print("Some parameters", file=sys.stderr)
+        print("textonly_pdf\t1\tSome help text")
+        sys.exit(0)
+    elif sys.argv[-2] in ('hocr', 'pdf'):
+        sys.stdout.buffer.write(BAD_UTF8)
+        sys.exit(1)
+    elif sys.argv[-1] == 'stdout':
+        # input file is at sys.argv[-2] but we don't look at it
+        print("""Orientation: 0
+Orientation in degrees: 0
+Orientation confidence: 100.00
+Script: 1
+Script confidence: 100.00""", file=sys.stderr)
+    else:
+        print("Spoof doesn't understand arguments", file=sys.stderr)
+        print(sys.argv, file=sys.stderr)
+        sys.exit(1)
+
+    sys.exit(0)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/tesseract_big_image_error.py

@@ -1,20 +1,16 @@
 #!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
 import sys
 
 
-VERSION_STRING = '''tesseract 3.04.00
- leptonica-1.72
-  libjpeg 8d : libpng 1.6.19 : libtiff 4.0.6 : zlib 1.2.5
+VERSION_STRING = '''tesseract 3.05.01
+ leptonica-1.74.4
+  libjpeg 9b : libpng 1.6.32 : libtiff 4.0.8 : zlib 1.2.8
 SPOOFED: return error claiming image too big
 '''
 
-"""Simulates a Tesseract crash
-
-It isn't strictly necessary to crash the process and that has unwanted
-side effects like triggering core dumps or error reporting, logging and such.
-It's enough to dump some text to stderr and return an error code.
-
-Follows the POSIX? convention of returning 128 + signal number.
+"""Simulates an error of Tesseract failing on attempts to process large images
 
 """
 
@@ -24,13 +20,16 @@ def main():
         print(VERSION_STRING, file=sys.stderr)
         sys.exit(0)
     elif sys.argv[1] == '--list-langs':
-        print('List of available languages (1):\neng', file=sys.stderr)
+        print('List of available languages (1):\neng\n', file=sys.stderr)
         sys.exit(0)
-    elif sys.argv[-1] == 'hocr':
+    elif sys.argv[1] == '--print-parameters':
+        print('A parameter list would go here\ntextonly_pdf 0\n', file=sys.stderr)
+        sys.exit(0)
+    elif sys.argv[-2] == 'hocr':
         print("Image too large: (33830, 14959)\n"
               "Error during processing.", file=sys.stderr)
         sys.exit(1)
-    elif sys.argv[-1] == 'pdf':
+    elif sys.argv[-2] == 'pdf':
         print("Image too large: (33830, 14959)\n"
               "Error during processing.", file=sys.stderr)
         sys.exit(1)

tests/spoof/tesseract_cache.py

@@ -1,4 +1,6 @@
 #!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
+
 import sys
 import os
 import hashlib
@@ -6,6 +8,34 @@ import shutil
 import subprocess
 
 
+"""Cache output of tesseract to speed up test suite
+
+The cache is keyed by a hash that includes the tesseract version, some of
+the command line, and the binary dump of the input file, and this file itself.
+ Therefore any updates to this file invalidate cache. Uses SHA-1 because it is
+ fast and defeating a hash collision here is not exactly a priority. :P
+
+The output files, stdout, and stderr are replicated on a cache hit. The output
+files are either a .pdf and .txt or .hocr and .txt.
+
+Page orientation checks are also cached (-psm 0 stdout)
+
+Errors and crashes are not cached. If the arguments don't match a known
+caching template then real tesseract is called with the same arguments.
+
+Things not checked:
+-changes to tesseract installation that don't affect --version
+
+Assumes Tesseract 3.04 or higher.
+
+Will fail on Tesseract 3.02.02 in "hocr" mode because it doesn't produce
+the incorrect file extension. Will fail on 3.03 because that has no sidecar
+text support. Will fail to replicate a 3.04 bug if wrong parameter order is
+given.
+
+"""
+
+
 CACHE_PATH = os.path.abspath(os.path.join(
         os.path.dirname(__file__), '..', 'cache'))
 
@@ -16,8 +46,15 @@ def real_tesseract():
     return  # Not reachable
 
 def main():
-    operation = sys.argv[-1]
+    operation = sys.argv[-2]
+    sidecar = False
+    if sys.argv[-1] == 'txt':
+        sidecar = True
+    elif sys.argv[-1] == 'stdout':
+        operation = 'stdout'
+
     # For anything unexpected operation, defer to real tesseract binary
+    # Currently this includes all use of "--tesseract-config"
     if operation != 'hocr' and operation != 'pdf' and operation != 'stdout':
         real_tesseract()
         return  # Not reachable
@@ -29,39 +66,66 @@ def main():
 
     m = hashlib.sha1()
 
-    version = subprocess.check_output(
+    tess_version = subprocess.check_output(
         ['tesseract', '--version'],
         stderr=subprocess.STDOUT)
 
-    m.update(version)
+    m.update(tess_version)
+
+    # Insert this source file into the hash function, to ensure that any
+    # changes to this file invalidate previous hashes
+    with open(__file__, 'rb') as f:
+        m.update(f.read())
+
     m.update(operation.encode())
 
     try:
         lang = sys.argv[sys.argv.index('-l') + 1]
         m.update(lang.encode())
     except ValueError:
-        pass
+        m.update(b'default-lang')
+
     try:
-        psm = sys.argv[sys.argv.index('-psm') + 1]
-        m.update(psm.encode())
+        textonly = sys.argv[sys.argv.index('-c') + 1]
+        m.update(textonly.encode())
     except ValueError:
-        pass
+        m.update(b'textonly_pdf=0')
+
+    psm_arg = ''
+    if '--psm' in sys.argv:
+        psm_arg = '--psm'
+    elif '-psm' in sys.argv:
+        psm_arg = '-psm'
+    if psm_arg:
+        try:
+            psm = sys.argv[sys.argv.index(psm_arg) + 1]
+            m.update(psm.encode())
+        except ValueError:
+            m.update(b'default-psm')
+    else:
+        m.update(b'default-psm')
 
     if operation == 'stdout' and psm != '0':
         real_tesseract()
         return
 
     if operation == 'stdout':
+        # tesseract [--options] ... input stdout
         input_file = sys.argv[-2]
         output_file = 'stdout'
+        sidecar_file = ''
     else:
-        input_file = sys.argv[-3]
-        output_file = sys.argv[-2]
+        # tesseract [--options] ... input output txt hocr|pdf
+        input_file = sys.argv[-4]
+        output_file = sys.argv[-3]
+        sidecar_file = sys.argv[-3]
 
     if operation == 'hocr':
         output_file += '.hocr'
+        sidecar_file += '.txt'
     elif operation == 'pdf':
         output_file += '.pdf'
+        sidecar_file += '.txt'
 
     with open(input_file, 'rb') as f:
         m.update(f.read())
@@ -72,12 +136,14 @@ def main():
         print("Tesseract cache hit", file=sys.stderr)
         if operation != 'stdout':
             shutil.copy(cache_name, output_file)
+            if sidecar:
+                shutil.copy(cache_name + '.sidecar', sidecar_file)
 
         # Replicate output
-        with open(cache_name + '.stdout', 'r') as f:
-            print(f.read(), end='')
-        with open(cache_name + '.stderr', 'r') as f:
-            print(f.read(), end='', file=sys.stderr)
+        with open(cache_name + '.stdout', 'rb') as f:
+            sys.stdout.buffer.write(f.read())
+        with open(cache_name + '.stderr', 'rb') as f:
+            sys.stderr.buffer.write(f.read())
         sys.exit(0)
 
     # Cache miss
@@ -86,23 +152,22 @@ def main():
     # Call tesseract
     p = subprocess.Popen(
             ['tesseract'] + sys.argv[1:],
-            stdout=subprocess.PIPE, stderr=subprocess.PIPE,
-            universal_newlines=True)
+            stdout=subprocess.PIPE, stderr=subprocess.PIPE)
     stdout, stderr = p.communicate()
 
     if p.returncode != 0:
         # Do not cache errors or crashes
         print("Tesseract error", file=sys.stderr)
-        print(stdout, end='')
-        print(stderr, end='', file=sys.stderr)
+        sys.stdout.buffer.write(stdout)
+        sys.stderr.buffer.write(stderr)
         return p.returncode
 
-    with open(cache_name + '.stdout', 'w') as f:
+    with open(cache_name + '.stdout', 'wb') as f:
         f.write(stdout)
-    with open(cache_name + '.stderr', 'w') as f:
+    with open(cache_name + '.stderr', 'wb') as f:
         f.write(stderr)
-    print(stdout, end='')
-    print(stderr, end='', file=sys.stderr)
+    sys.stdout.buffer.write(stdout)
+    sys.stderr.buffer.write(stderr)
 
     # Insert file into cache
     if output_file != 'stdout':
@@ -110,6 +175,8 @@ def main():
             shutil.copy(output_file, cache_name)
         else:
             print("Could not find output file", file=sys.stderr)
+        if sidecar and os.path.exists(sidecar_file):
+            shutil.copy(sidecar_file, cache_name + '.sidecar')
     else:
         open(cache_name, 'w').close()
 

tests/spoof/tesseract_crash.py

@@ -1,22 +1,23 @@
 #!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
 import sys
 import os
 import signal
 
 
-VERSION_STRING = '''tesseract 3.04.00
- leptonica-1.72
-  libjpeg 8d : libpng 1.6.19 : libtiff 4.0.6 : zlib 1.2.5
+VERSION_STRING = '''tesseract 3.05.01
+ leptonica-1.74.4
+  libjpeg 9b : libpng 1.6.32 : libtiff 4.0.8 : zlib 1.2.8
 SPOOFED: CRASH ON OCR or -psm 0
 '''
 
-"""Simulates a Tesseract crash
+"""Simulates a Tesseract crash when asked to run OCR
 
 It isn't strictly necessary to crash the process and that has unwanted
 side effects like triggering core dumps or error reporting, logging and such.
 It's enough to dump some text to stderr and return an error code.
 
-Follows the POSIX? convention of returning 128 + signal number.
+Follows the POSIX(?) convention of returning 128 + signal number.
 
 """
 
@@ -28,10 +29,14 @@ def main():
     elif sys.argv[1] == '--list-langs':
         print('List of available languages (1):\neng', file=sys.stderr)
         sys.exit(0)
-    elif sys.argv[-1] == 'hocr':
+    elif sys.argv[1] == '--print-parameters':
+        print('A parameter list would go here\ntextonly_pdf 0\n',
+              file=sys.stderr)
+        sys.exit(0)
+    elif sys.argv[-2] == 'hocr':
         print("KABOOM! Tesseract failed for some reason", file=sys.stderr)
         sys.exit(128 + signal.SIGSEGV)
-    elif sys.argv[-1] == 'pdf':
+    elif sys.argv[-2] == 'pdf':
         print("KABOOM! Tesseract failed for some reason", file=sys.stderr)
         sys.exit(128 + signal.SIGSEGV)
     elif sys.argv[-1] == 'stdout':

tests/spoof/tesseract_noop.py

@@ -1,10 +1,24 @@
 #!/usr/bin/env python3
+# © 2016 James R. Barlow: github.com/jbarlow83
 import sys
 import img2pdf
+import PyPDF2 as pypdf
 from PIL import Image
 
 
-VERSION_STRING = '''tesseract 3.04.00
+"""Tesseract no-op spoof
+
+To quickly run tests where getting OCR output is not necessary.
+
+In 'hocr' mode, create a .hocr file that specifies no text found.
+
+In 'pdf' mode, convert the image to PDF using another program.
+
+In orientation check mode, report the orientation is upright.
+"""
+
+
+VERSION_STRING = '''tesseract 3.05.01
  leptonica-1.72
   libjpeg 8d : libpng 1.6.19 : libtiff 4.0.6 : zlib 1.2.5
 SPOOFED
@@ -40,19 +54,42 @@ def main():
     elif sys.argv[1] == '--list-langs':
         print('List of available languages (1):\neng', file=sys.stderr)
         sys.exit(0)
-    elif sys.argv[-1] == 'hocr':
-        inputf = sys.argv[-3]
-        output = sys.argv[-2]
+    elif sys.argv[1] == '--print-parameters':
+        print("Some parameters", file=sys.stderr)
+        print("textonly_pdf\t1\tSome help text")
+        sys.exit(0)
+    elif sys.argv[-2] == 'hocr':
+        inputf = sys.argv[-4]
+        output = sys.argv[-3]
         with Image.open(inputf) as im, \
                 open(output + '.hocr', 'w', encoding='utf-8') as f:
             w, h = im.size
             f.write(HOCR_TEMPLATE.format(str(w), str(h)))
-    elif sys.argv[-1] == 'pdf':
-        inputf = sys.argv[-3]
-        output = sys.argv[-2]
-        pdf_bytes = img2pdf.convert([inputf], dpi=300)
-        with open(output + '.pdf', 'wb') as f:
-            f.write(pdf_bytes)
+        with open(output + '.txt', 'w') as f:
+            f.write('')
+    elif sys.argv[-2] == 'pdf':
+        if 'textonly_pdf=1' in sys.argv:
+            inputf = sys.argv[-4]
+            output = sys.argv[-3]
+            with Image.open(inputf) as im:
+                dpi = im.info['dpi']
+                pagesize = im.size[0] / dpi[0], im.size[1] / dpi[1]
+                ptsize = pagesize[0] * 72, pagesize[1] * 72
+
+            pdf_out = pypdf.PdfFileWriter()
+            pdf_out.addBlankPage(ptsize[0], ptsize[1])
+            with open(output + '.pdf', 'wb') as f:
+                pdf_out.write(f)
+            with open(output + '.txt', 'w') as f:
+                f.write('')
+        else:
+            inputf = sys.argv[-4]
+            output = sys.argv[-3]
+            pdf_bytes = img2pdf.convert([inputf], dpi=300)
+            with open(output + '.pdf', 'wb') as f:
+                f.write(pdf_bytes)
+            with open(output + '.txt', 'w') as f:
+                f.write('')
     elif sys.argv[-1] == 'stdout':
         inputf = sys.argv[-2]
         print("""Orientation: 0