Tweak release notes

Should extend test for other Asian languages

.travis.yml

@@ -1,3 +1,5 @@
+sudo: required
+dist: trusty
 language: python
 
 cache: pip
@@ -10,16 +12,15 @@ cache:
 
 python:
     - 3.4
+    - 3.5
 
 before_cache:
     - 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 add-apt-repository ppa:vshn/ghostscript -y  # for ghostscript 9.16 (trusty has 9.10)
+    - sudo add-apt-repository ppa:heyarje/libav-11 -y  # for libav11, which is 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
 

LICENSE.rst

@@ -18,9 +18,3 @@ 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.

README.rst

@@ -8,7 +8,7 @@ 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 +18,11 @@ 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 the `39 languages <https://code.google.com/p/tesseract-ocr/downloads/list>`_ 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 `release notes <RELEASE_NOTES.rst>`_.
 
 Motivation
 ----------
@@ -31,9 +31,9 @@ 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)
@@ -46,28 +46,45 @@ Installation
 
 Download OCRmyPDF here: https://github.com/jbarlow83/OCRmyPDF/releases
 
-You can install it to a Python virtual environment or system-wide. 
+These steps describe how to install OCRmyPDF on your system.
 
-Debian and Ubuntu
-~~~~~~~~~~~~~~~~~
+- `Installing on Debian and Ubuntu`_ (Debian stretch and Ubuntu 16.10 or later)
+- `Installing the Docker image`_
+- `Installing on Mac OS X`_
+- `Installing on Ubuntu 14.04 LTS`_
+- Installing and running on `Windows`_ using the Docker image
+
+If you prefer to install from source or install OCRmyPDF to a Python virtual environment, see steps for `Installing HEAD revision from sources`_.
+
+.. _Windows: `Installing on Windows`_
+
+
+Installing on Debian and Ubuntu
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Users of Debian 9 or later or Ubuntu 16.10 or later may simply
 ``apt-get install ocrmypdf``.
 
+.. _Docker:
+
 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
+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::
+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::
+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:
+
+.. code-block:: bash
 
    # Optional step for Mac OS X users
    docker-machine stop "yourVM"
@@ -76,29 +93,41 @@ OCRmyPDF will use all available CPU cores.  By default, the VirtualBox machine i
    eval $(docker-machine env "yourVM")
 
 Assuming you have a Docker engine running somewhere, you can run these commands to download
-the image::
+the image:
+
+.. code-block:: bash
 
    docker pull jbarlow83/ocrmypdf
 
-Then tag it to give a more convenient name, just ocrmypdf::
+Then tag it to give a more convenient name, just ocrmypdf:
+
+.. code-block:: bash
 
    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>`__::
+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>`_:
+
+.. code-block:: bash
 
    # 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::
+You can then run ocrmypdf using the command:
+
+.. code-block:: bash
 
    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::
+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:
+
+.. code-block:: bash
 
    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``:: 
+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 -v "$(pwd):/home/docker"   ocrmypdf --skip-text test.pdf output.pdf
 
@@ -110,85 +139,135 @@ 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/>`__.
+If it's not already present, `install Homebrew <http://brew.sh/>`_.
 
-Update Homebrew::
+Update Homebrew:
+
+.. code-block:: bash
 
    brew update
    
-Install or upgrade the required Homebrew packages, if any are missing::
+Install or upgrade the required Homebrew packages, if any are missing:
 
-   brew install libpng openjpeg jbig2dec     # image libraries
+.. 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
-   brew install tesseract
+   brew install unpaper   # optional
    
-Update the homebrew pip and install Pillow::
+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
+   
+.. 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::
+You can then install OCRmyPDF from PyPI:
+
+.. code-block:: bash
 
    pip3 install ocrmypdf
 
-The command line program should now be available::
+The command line program should now be available:
+
+.. code-block:: bash
 
    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.
+Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other options, because of bugs in Python package installation.
 
-Update apt-get::
+Add new "apt" repositories needed for backports of Ghostscript 9.16 and libav-11, which supports unpaper 6.1. This will replace Ghostscript on your system.
+
+.. code-block:: bash
+
+   sudo add-apt-repository ppa:vshn/ghostscript -y 
+   sudo add-apt-repository ppa:heyarje/libav-11 -y
+
+Update apt-get:
+
+.. code-block:: bash
 
    sudo apt-get update
    sudo apt-get upgrade
    
-Install system dependencies::
+Install system dependencies:
+
+.. code-block:: bash
 
    sudo apt-get install \
       zlib1g-dev \
       libjpeg-dev \
       libffi-dev \
+      libavformat56 libavcodec56 libavutil54 \
       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)::
+into your system Python, which could interfere with other programs):
+
+.. code-block:: bash
 
    sudo pip3 install ocrmypdf
    
-If you wish to install OCRmyPDF to a virtual environment to isolate system Python from modified, you can
+If you wish to install OCRmyPDF to a virtual environment to isolate the system Python, 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>`__::
+package <http://www.thefourtheye.in/2014/12/Python-venv-problem-with-ensurepip-in-Ubuntu.html>`_:
+
+.. code-block:: bash
 
    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
+   python3 -m venv --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.
+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 is not motivated to figure how to set up a Ubuntu PPA to distribute it. You can create a .deb package to do the job of installing unpaper 6.1 (for x86 64-bit only):
+
+.. code-block:: bash
+
+   wget -q https://dl.dropboxusercontent.com/u/28971240/unpaper_6.1-1.deb -O unpaper_6.1-1.deb
+   sudo dpkg -i unpaper_6.1-1.deb
+
 
 Installing on Windows
 ~~~~~~~~~~~~~~~~~~~~~
 
-Direct installation on Windows is not possible.  Install the Docker container as described above.
+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.
       
 Installing HEAD revision from sources
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -196,24 +275,32 @@ 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::
+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::
+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::
+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::
+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
-   pyvenv venv
+   python3 -m venv
    source venv/bin/activate
    cd OCRmyPDF
    pip3 install .
@@ -221,7 +308,9 @@ to not install as superuser, you can install the package in a Python virtual env
 However, ``ocrmypdf`` will only be accessible on the system PATH after
 you activate the virtual environment.
 
-To run the program::
+To run the program:
+
+.. code-block:: bash
    
    ocrmypdf --help
 
@@ -234,10 +323,15 @@ 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
+   sudo 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.
@@ -245,9 +339,17 @@ languages can be requested.
 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
+
+The `Wiki <https://github.com/jbarlow83/OCRmyPDF/wiki>`_ page also contains some tips and suggests.
+
+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
@@ -259,11 +361,11 @@ In case you detect an issue, please:
 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

@@ -1,9 +1,60 @@
 RELEASE NOTES
 =============
 
-Please always read this file before installing the package
 
-Download software here: https://github.com/jbarlow83/OCRmyPDF/tags
+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:
@@ -30,9 +81,6 @@ v4.0.6:
 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
@@ -56,7 +104,7 @@ New features
 Fixes
 -----
 
--  Show stack trace if unexpect errors occur
+-  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
@@ -329,363 +377,6 @@ Notes and known issues
 v2.2-stable (2014-09-29):
 =========================
 
-New features
-------------
+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.
 
-- 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"
+For details on older versions, see the `final version of its release notes <https://github.com/fritz-hh/OCRmyPDF/blob/7fd3dbdf42ca53a619412ce8add7532c5e81a9d1/RELEASE_NOTES.md>`_.

ocrmypdf/__init__.py

@@ -1,5 +1,6 @@
 from enum import IntEnum
 import os
+from collections.abc import Iterable
 
 
 class ExitCode(IntEnum):
@@ -7,11 +8,13 @@ class ExitCode(IntEnum):
     bad_args = 1
     input_file = 2
     missing_dependency = 3
-    invalid_output_pdfa = 4
+    invalid_output_pdf = 4
     file_access_error = 5
     already_done_ocr = 6
     child_process_error = 7
+    encrypted_pdf = 8
     other_error = 15
+    ctrl_c = 130
 
 
 def get_program(name):
@@ -21,3 +24,7 @@ def get_program(name):
 
 def page_number(input_file):
     return int(os.path.basename(input_file)[0:6])
+
+
+def is_iterable_notstr(thing):
+    return isinstance(thing, Iterable) and not isinstance(thing, str)

ocrmypdf/ghostscript.py

@@ -14,6 +14,7 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
         args_gs = [
             get_program('gs'),
             '-dQUIET',
+            '-dSAFER',
             '-dBATCH',
             '-dNOPAUSE',
             '-sDEVICE=%s' % raster_device,

ocrmypdf/leptonica.py

@@ -17,6 +17,7 @@ from tempfile import TemporaryFile
 from ctypes.util import find_library
 from .lib._leptonica import ffi
 from functools import lru_cache
+from enum import Enum
 
 lept = ffi.dlopen(find_library('lept'))
 
@@ -83,6 +84,13 @@ 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.
 
@@ -98,14 +106,14 @@ class Pix:
     in a threadsafe manner if a Python threading.Lock protects the data.
     """
 
-    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>"
 
@@ -113,12 +121,16 @@ class Pix:
         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)
 
@@ -126,16 +138,41 @@ 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)
 
     @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):
@@ -158,7 +195,29 @@ class Pix:
         with LeptonicaErrorTrap():
             lept.pixWriteImpliedFormat(
                 filename.encode(sys.getfilesystemencoding()),
-                self.cpix, jpeg_quality, jpeg_progressive)
+                self._pix, jpeg_quality, jpeg_progressive)
+
+    def topil(self):
+        "Returns a PIL.Image version of this Pix"
+        from PIL import Image
+
+        with LeptonicaErrorTrap():
+            pix_swapped = Pix(lept.pixEndianByteSwapNew(self._pix))
+
+        size = (pix_swapped._pix.wpl * 4, pix_swapped._pix.h)
+        buf = ffi.buffer(pix_swapped._pix.data, size[0] * size[1])
+
+        im_raw = Image.frombytes(self.mode, size, buf, 'raw')
+
+        # Leptonica stores images in 32-bit words
+        # Need to crop the any trailing amount
+        box = (0, 0, self.width, self.height)
+        im = im_raw.crop(box)
+
+        return im
+
+    def show(self):
+        return self.topil().show()
 
     def deskew(self, reduction_factor=0):
         """Returns the deskewed pix object.
@@ -170,16 +229,21 @@ 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):
         "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, scalex, scaley))
 
     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 +253,105 @@ 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))
+
     @staticmethod
     @lru_cache(maxsize=1)
     def make_pixel_sum_tab8():
@@ -224,16 +362,17 @@ 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 +383,7 @@ 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 +391,48 @@ 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.

ocrmypdf/lib/compile_leptonica.py

@@ -42,6 +42,18 @@ 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("""
@@ -50,10 +62,17 @@ 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 );
 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 +81,10 @@ 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,
@@ -72,6 +95,45 @@ pixOtsuAdaptiveThreshold(PIX       *pixs,
                          PIX      **ppixth,
                          PIX      **ppixd);
 
+PIX *
+pixOtsuThreshOnBackgroundNorm(PIX       *pixs,
+                              PIX       *pixim,
+                              l_int32    sx,
+                              l_int32    sy,
+                              l_int32    thresh,
+                              l_int32    mincount,
+                              l_int32    bgval,
+                              l_int32    smoothx,
+                              l_int32    smoothy,
+                              l_float32  scorefract,
+                              l_int32   *pthresh);
+
+PIX *
+pixCleanBackgroundToWhite(PIX       *pixs,
+                          PIX       *pixim,
+                          PIX       *pixg,
+                          l_float32  gamma,
+                          l_int32    blackval,
+                          l_int32    whiteval);
+
+BOX *
+pixFindPageForeground(PIX         *pixs,
+                      l_int32      threshold,
+                      l_int32      mindist,
+                      l_int32      erasedist,
+                      l_int32      pagenum,
+                      l_int32      showmorph,
+                      l_int32      display,
+                      const char  *pdfdir);
+
+PIX *
+pixClipRectangle(PIX   *pixs,
+                 BOX   *box,
+                 BOX  **pboxc);
+
+void
+boxDestroy(BOX  **pbox);
+
 void lept_free(void *ptr);
 """)
 

ocrmypdf/main.py

@@ -27,11 +27,12 @@ import ruffus.proxy_logger as proxy_logger
 
 from .hocrtransform import HocrTransform
 from .pageinfo import pdf_get_all_pageinfo
-from .pdfa import generate_pdfa_def
+from .pdfa import generate_pdfa_def, file_claims_pdfa
 from . import ghostscript
 from . import tesseract
 from . import qpdf
-from . import ExitCode, page_number
+from . import ExitCode, page_number, is_iterable_notstr
+from collections.abc import Sequence
 
 import pkg_resources
 
@@ -42,6 +43,8 @@ warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
 
 BASEDIR = os.path.dirname(os.path.realpath(__file__))
 
+VECTOR_PAGE_DPI = 400
+
 
 # -------------
 # External dependencies
@@ -120,6 +123,16 @@ parser.add_argument(
 parser.add_argument(
     '-j', '--jobs', metavar='N', type=int,
     help="Use up to N CPU cores simultaneously (default: use all)")
+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'], 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.")
 
 metadata = parser.add_argument_group(
     "Metadata options",
@@ -269,6 +282,13 @@ if options.clean and not options.clean_final \
         "Tesseract PDF renderer cannot render --clean pages without "
         "also performing --clean-final, so --clean-final is assumed.")
 
+if set(options.language) & {'chi_sim', 'chi_tra'} \
+        and (options.pdf_renderer == 'hocr' or options.output_type == 'pdfa'):
+    complain(
+        "Your settings are known to cause problems with OCR of Chinese text. "
+        "Try adding these arguments: "
+        "    ocrmypdf --pdf-renderer tesseract --output-type pdf")
+
 lossless_reconstruction = False
 if options.pdf_renderer == 'hocr':
     if not options.deskew and not options.clean_final and not options.force_ocr:
@@ -384,10 +404,87 @@ def cleanup_working_files(*args):
             shutil.rmtree(work_folder)
 
 
+def triage_image_file(input_file, output_file, log):
+    try:
+        log.info("Input file is not a PDF, checking if it is an image...")
+        im = Image.open(input_file)
+    except EnvironmentError as e:
+        log.error(e)
+        sys.exit(ExitCode.input_file)
+        return
+    else:
+        log.info("Input file is an image")
+
+        if 'dpi' in im.info:
+            if im.info['dpi'] <= (96, 96) and not options.image_dpi:
+                log.info("Image size: (%d, %d)" % im.size)
+                log.info("Image resolution: (%d, %d)" % im.info['dpi'])
+                log.error(
+                    "Input file is an image, but the resolution (DPI) is "
+                    "not credible.  Estimate the resolution at which the "
+                    "image was scanned and specify it using --image-dpi.")
+                sys.exit(ExitCode.input_file)
+        elif not options.image_dpi:
+            log.info("Image size: (%d, %d)" % im.size)
+            log.error(
+                "Input file is an image, but has no resolution (DPI) "
+                "in its metadata.  Estimate the resolution at which "
+                "image was scanned and specify it using --image-dpi.")
+            sys.exit(ExitCode.input_file)
+
+        if 'iccprofile' not in im.info:
+            if im.mode == 'RGB':
+                log.info('Input image has no ICC profile, assuming sRGB')
+            elif im.mode == 'CMYK':
+                log.info('Input CMYK image has no ICC profile, not usable')
+                sys.exit(ExitCode.input_file)
+        im.close()
+
+    try:
+        log.info("Image seems valid. Try converting to PDF...")
+        layout_fun = img2pdf.default_layout_fun
+        if options.image_dpi:
+            layout_fun = img2pdf.get_fixed_dpi_layout_fun(
+                (options.image_dpi, options.image_dpi))
+        with open(output_file, 'wb') as outf:
+            img2pdf.convert(
+                input_file,
+                layout_fun=layout_fun,
+                with_pdfrw=False,
+                outputstream=outf)
+        log.info("Successfully converted to PDF, processing...")
+    except img2pdf.ImageOpenError as e:
+        log.error(e)
+        sys.exit(ExitCode.input_file)
+
+
 @transform(
     input=options.input_file,
-    filter=formatter('(?i)\.pdf'),
-    output=os.path.join(work_folder, '{basename[0]}.repaired.pdf'),
+    filter=formatter('(?i)'),
+    output=os.path.join(work_folder, '{basename[0]}.pdf'),
+    extras=[_log])
+def triage(
+        input_file,
+        output_file,
+        log):
+    try:
+        with open(input_file, 'rb') as f:
+            signature = f.read(4)
+            if signature == b'%PDF':
+                re_symlink(input_file, output_file)
+                return
+    except EnvironmentError as e:
+        log.error(e)
+        sys.exit(ExitCode.input_file)
+
+    triage_image_file(input_file, output_file, log)
+
+
+@transform(
+    input=triage,
+    filter=suffix('.pdf'),
+    output='.repaired.pdf',
+    output_dir=work_folder,
     extras=[_log, _pdfinfo, _pdfinfo_lock])
 def repair_pdf(
         input_file,
@@ -409,31 +506,60 @@ def get_pageinfo(input_file, pdfinfo, pdfinfo_lock):
     return pageinfo
 
 
+def get_page_dpi(pageinfo):
+    "Get the DPI when nonsquare DPI is tolerable"
+    xres = max(pageinfo.get('xres', VECTOR_PAGE_DPI), options.oversample or 0)
+    yres = max(pageinfo.get('yres', VECTOR_PAGE_DPI), options.oversample or 0)
+    return (float(xres), float(yres))
+
+
+def get_page_square_dpi(pageinfo):
+    "Get the DPI when we require xres == yres"
+    return float(max(
+        pageinfo.get('xres', VECTOR_PAGE_DPI),
+        pageinfo.get('yres', VECTOR_PAGE_DPI),
+        options.oversample or 0))
+
+
 def is_ocr_required(pageinfo, log):
     page = pageinfo['pageno'] + 1
     ocr_required = True
     if not pageinfo['images']:
-        # If the page has no images, then it contains vector content or text
-        # or both. It seems quite unlikely that one would find meaningful text
-        # from rasterizing vector content. So skip the page.
-        log.info(
-            "{0:4d}: page has no images - skipping OCR".format(page)
-        )
-        ocr_required = False
+        if options.force_ocr and options.oversample:
+            # The user really wants to reprocess this file
+            log.info(
+                "{0:4d}: page has no images - "
+                "rasterizing at {1} DPI because "
+                "--force-ocr --oversample was specified".format(
+                    page, options.oversample))
+        elif options.force_ocr:
+            # Warn the user they might not want to do this
+            log.warning(
+                "{0:4d}: page has no images - "
+                "all vector content will be "
+                "rasterized at {1} DPI, losing some resolution and likely "
+                "increasing file size. Use --oversample to adjust the "
+                "DPI.".format(page, VECTOR_PAGE_DPI))
+        else:
+            log.info(
+                "{0:4d}: page has no images - "
+                "skipping all processing on this page".format(page))
+            ocr_required = False
+
     elif pageinfo['has_text']:
-        s = "{0:4d}: page already has text! – {1}"
+        msg = "{0:4d}: page already has text! – {1}"
 
         if not options.force_ocr and not options.skip_text:
-            log.error(s.format(page,
-                               "aborting (use --force-ocr to force OCR)"))
+            log.error(msg.format(page,
+                                 "aborting (use --force-ocr to force OCR)"))
             sys.exit(ExitCode.already_done_ocr)
         elif options.force_ocr:
-            log.info(s.format(page,
-                              "rasterizing text and running OCR anyway"))
+            log.info(msg.format(page,
+                                "rasterizing text and running OCR anyway"))
             ocr_required = True
         elif options.skip_text:
-            log.info(s.format(page,
-                              "skipping all processing on this page"))
+            log.info(msg.format(page,
+                                "skipping all processing on this page"))
             ocr_required = False
 
     if ocr_required and options.skip_big:
@@ -452,12 +578,17 @@ def is_ocr_required(pageinfo, log):
     os.path.join(work_folder, '*.page.pdf'),
     extras=[_log, _pdfinfo, _pdfinfo_lock])
 def split_pages(
-        input_file,
+        input_files,
         output_files,
         log,
         pdfinfo,
         pdfinfo_lock):
 
+    if is_iterable_notstr(input_files):
+        input_file = input_files[0]
+    else:
+        input_file = input_files
+
     for oo in output_files:
         with suppress(FileNotFoundError):
             os.unlink(oo)
@@ -606,11 +737,14 @@ def rasterize_with_ghostscript(
             device = 'pnggray'
 
     log.debug("Rasterize {0} with {1}".format(
-            os.path.basename(input_file), device))
-    xres = max(pageinfo['xres'], options.oversample or 0)
-    yres = max(pageinfo['yres'], options.oversample or 0)
+              os.path.basename(input_file), device))
 
-    ghostscript.rasterize_pdf(input_file, output_file, xres, yres, device, log)
+    # Produce the page image with square resolution or else deskew and OCR
+    # will not work properly
+    dpi = get_page_square_dpi(pageinfo)
+    ghostscript.rasterize_pdf(
+        input_file, output_file, xres=dpi, yres=dpi, raster_device=device,
+        log=log)
 
 
 @transform(
@@ -630,7 +764,7 @@ def preprocess_deskew(
         return
 
     pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-    dpi = int(pageinfo['xres'])
+    dpi = get_page_square_dpi(pageinfo)
 
     from . import leptonica
     leptonica.deskew(input_file, output_file, dpi)
@@ -653,7 +787,7 @@ def preprocess_clean(
         return
 
     pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-    dpi = int(pageinfo['xres'])
+    dpi = get_page_square_dpi(pageinfo)
 
     unpaper.clean(input_file, output_file, dpi, log)
 
@@ -704,13 +838,20 @@ def select_image_for_pdf(
     image = next(ii for ii in infiles if ii.endswith(image_suffix))
 
     pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-    if all(image['enc'] == 'jpeg' for image in pageinfo['images']):
+    if all(orig_image['enc'] == 'jpeg' for orig_image in pageinfo['images']):
         # If all images were JPEGs originally, produce a JPEG as output
         im = Image.open(image)
-        dpi = im.info.get(
-            'dpi',
-            (int(pageinfo['xres']), int(pageinfo['yres']))
-        )
+
+        # At this point the image should be a .png, but deskew, unpaper might
+        # have removed the DPI information. In this case, fall back to square
+        # DPI used to rasterize. When the preview image was rasterized, it
+        # was also converted to square resolution, which is what we want to
+        # give tesseract, so keep it square.
+        fallback_dpi = get_page_square_dpi(pageinfo)
+        dpi = im.info.get('dpi', (fallback_dpi, fallback_dpi))
+
+        # Pillow requires integer DPI
+        dpi = round(dpi[0]), round(dpi[1])
         im.save(output_file, format='JPEG', dpi=dpi)
     else:
         re_symlink(image, output_file)
@@ -738,18 +879,16 @@ def select_image_layer(
         re_symlink(page_pdf, output_file)
     else:
         pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-        dpi = round(max(pageinfo['xres'], pageinfo['yres'],
-                        options.oversample))
-        imgsize = ((img2pdf.ImgSize.dpi, dpi), (img2pdf.ImgSize.dpi, dpi))
-
-        layout_fun = img2pdf.get_layout_fun(None, imgsize, None, None, None)
+        dpi = get_page_dpi(pageinfo)
+        dpi = float(dpi[0]), float(dpi[1])
+        layout_fun = img2pdf.get_fixed_dpi_layout_fun(dpi)
 
         with open(image, 'rb') as imfile, \
                 open(output_file, 'wb') as pdf:
             rawdata = imfile.read()
-            pdf.write(img2pdf.convert(
-                rawdata, producer="img2pdf", with_pdfrw=False,
-                layout_fun=layout_fun))
+            img2pdf.convert(
+                rawdata, with_pdfrw=False,
+                layout_fun=layout_fun, outputstream=pdf)
 
 
 @active_if(options.pdf_renderer == 'hocr')
@@ -766,7 +905,7 @@ def render_hocr_page(
         pdfinfo_lock):
     hocr = input_file
     pageinfo = get_pageinfo(hocr, pdfinfo, pdfinfo_lock)
-    dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
+    dpi = get_page_square_dpi(pageinfo)
 
     hocrtransform = HocrTransform(hocr, dpi)
     hocrtransform.to_pdf(output_file, imageFileName=None,
@@ -790,13 +929,17 @@ def render_hocr_debug_page(
     image = next(ii for ii in infiles if ii.endswith('.image'))
 
     pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
-    dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
+    dpi = get_page_square_dpi(pageinfo)
 
     hocrtransform = HocrTransform(hocr, dpi)
     hocrtransform.to_pdf(output_file, imageFileName=None,
                          showBoundingboxes=True, invisibleText=False)
 
 
+class PdfMergeFailedError(Exception):
+    pass
+
+
 @active_if(options.pdf_renderer == 'hocr')
 @collate(
     input=[render_hocr_page, select_image_layer],
@@ -836,7 +979,7 @@ def add_text_layer(
     y2 = page_image.mediaBox.getUpperRight_y()
 
     # Rotation occurs about the page's (0, 0). Most pages will have the media
-    # box at (0, 0) will all content in the first quadrant but some cropped
+    # box at (0, 0) with all content in the first quadrant but some cropped
     # files may have an offset mediabox. We translate the page so that its
     # bottom left corner after rotation is pinned to (0, 0) with the image
     # in the first quadrant.
@@ -855,8 +998,12 @@ def add_text_layer(
         log.info("{0:4d}: rotating image layer {1} degrees".format(
             page_number(image), rotation, tx, ty))
 
-    page_text.mergeRotatedScaledTranslatedPage(
-        page_image, rotation, 1.0, tx, ty, expand=False)
+    try:
+        page_text.mergeRotatedScaledTranslatedPage(
+            page_image, rotation, 1.0, tx, ty, expand=False)
+    except (AttributeError, ValueError) as e:
+        if 'writeToStream' in str(e) or 'invalid literal' in str(e):
+            raise PdfMergeFailedError() from e
 
     pdf_output = pypdf.PdfFileWriter()
     pdf_output.addPage(page_text)
@@ -896,6 +1043,40 @@ def tesseract_ocr_and_render_pdf(
         log=log)
 
 
+def get_pdfmark(base_pdf):
+    def from_document_info(key):
+        # pdf.documentInfo.get() DOES NOT behave as expected for a dict-like
+        # object, so call with precautions.  TypeError may occur if the PDF
+        # is missing the optional document info section.
+        try:
+            s = base_pdf.documentInfo[key]
+            return str(s)
+        except (KeyError, TypeError):
+            return ''
+
+    pdfmark = {
+        '/Title': from_document_info('/Title'),
+        '/Author': from_document_info('/Author'),
+        '/Keywords': from_document_info('/Keywords'),
+        '/Subject': from_document_info('/Subject'),
+    }
+    if options.title:
+        pdfmark['/Title'] = options.title
+    if options.author:
+        pdfmark['/Author'] = options.author
+    if options.keywords:
+        pdfmark['/Keywords'] = options.keywords
+    if options.subject:
+        pdfmark['/Subject'] = options.subject
+
+    pdfmark['/Creator'] = '{0} {1} / Tesseract OCR{2} {3}'.format(
+            parser.prog, VERSION,
+            '+PDF' if options.pdf_renderer == 'tesseract' else '',
+            tesseract.version())
+    return pdfmark
+
+
+@active_if(options.output_type == 'pdfa')
 @transform(
     input=repair_pdf,
     filter=formatter(r'\.repaired\.pdf'),
@@ -907,37 +1088,7 @@ def generate_postscript_stub(
         log):
 
     pdf = pypdf.PdfFileReader(input_file)
-
-    def from_document_info(key):
-        # pdf.documentInfo.get() DOES NOT behave as expected for a dict-like
-        # object, so call with precautions.  TypeError may occur if the PDF
-        # is missing the optional document info section.
-        try:
-            s = pdf.documentInfo[key]
-            return str(s)
-        except (KeyError, TypeError):
-            return ''
-
-    pdfmark = {
-        'title': from_document_info('/Title'),
-        'author': from_document_info('/Author'),
-        'keywords': from_document_info('/Keywords'),
-        'subject': from_document_info('/Subject'),
-    }
-    if options.title:
-        pdfmark['title'] = options.title
-    if options.author:
-        pdfmark['author'] = options.author
-    if options.keywords:
-        pdfmark['keywords'] = options.keywords
-    if options.subject:
-        pdfmark['subject'] = options.subject
-
-    pdfmark['creator'] = '{0} {1} / Tesseract OCR{2} {3}'.format(
-            parser.prog, VERSION,
-            '+PDF' if options.pdf_renderer == 'tesseract' else '',
-            tesseract.version())
-
+    pdfmark = get_pdfmark(pdf)
     generate_pdfa_def(output_file, pdfmark)
 
 
@@ -958,12 +1109,13 @@ def skip_page(
     re_symlink(input_file, output_file, log)
 
 
+@active_if(options.output_type == 'pdfa')
 @merge(
     input=[add_text_layer, render_hocr_debug_page, skip_page,
            tesseract_ocr_and_render_pdf, generate_postscript_stub],
     output=os.path.join(work_folder, 'merged.pdf'),
     extras=[_log, _pdfinfo, _pdfinfo_lock])
-def merge_pages(
+def merge_pages_ghostscript(
         input_files,
         output_file,
         log,
@@ -988,26 +1140,66 @@ def merge_pages(
     ghostscript.generate_pdfa(pdf_pages, output_file, options.jobs or 1)
 
 
-@transform(
-    input=merge_pages,
-    filter=formatter(),
-    output=options.output_file,
+@active_if(options.output_type == 'pdf')
+@merge(
+    input=[add_text_layer, render_hocr_debug_page, skip_page,
+           tesseract_ocr_and_render_pdf, repair_pdf],
+    output=os.path.join(work_folder, 'merged.pdf'),
     extras=[_log, _pdfinfo, _pdfinfo_lock])
-def copy_final(
-        input_file,
+def merge_pages_qpdf(
+        input_files,
         output_file,
         log,
         pdfinfo,
         pdfinfo_lock):
+
+    metadata_file = next(
+        (ii for ii in input_files if ii.endswith('.repaired.pdf')))
+    input_files.remove(metadata_file)
+
+    def input_file_order(s):
+        '''Sort order: All rendered pages followed
+        by their debug page.'''
+        key = int(os.path.basename(s)[0:6]) * 10
+        if 'debug' in os.path.basename(s):
+            key += 1
+        return key
+
+    pdf_pages = sorted(input_files, key=input_file_order)
+    log.debug("Final pages: " + "\n".join(pdf_pages))
+
+    reader_metadata = pypdf.PdfFileReader(metadata_file)
+    pdfmark = get_pdfmark(reader_metadata)
+    pdfmark['/Producer'] = 'qpdf ' + qpdf.version()
+
+    first_page = pypdf.PdfFileReader(pdf_pages[0])
+
+    writer = pypdf.PdfFileWriter()
+    writer.appendPagesFromReader(first_page)
+    writer.addMetadata(pdfmark)
+    writer_file = pdf_pages[0].replace('.pdf', '.metadata.pdf')
+    with open(writer_file, 'wb') as f:
+        writer.write(f)
+
+    pdf_pages[0] = writer_file
+
+    qpdf.merge(pdf_pages, output_file)
+
+
+@merge(
+    input=[merge_pages_ghostscript, merge_pages_qpdf],
+    output=options.output_file,
+    extras=[_log, _pdfinfo, _pdfinfo_lock])
+def copy_final(
+        input_files,
+        output_file,
+        log,
+        pdfinfo,
+        pdfinfo_lock):
+    input_file = next((ii for ii in input_files if ii.endswith('.pdf')))
     shutil.copy(input_file, output_file)
 
 
-def validate_pdfa(
-        input_file,
-        log):
-    return qpdf.check(input_file, log)
-
-
 def available_cpu_count():
     try:
         return multiprocessing.cpu_count()
@@ -1053,26 +1245,55 @@ def do_ruffus_exception(ruffus_five_tuple):
             _log.error("Input file '{0}' is not a valid PDF".format(
                 options.input_file))
             return ExitCode.input_file
+    elif exc_name == 'builtins.KeyboardInterrupt':
+        _log.error("Interrupted by user")
+        return 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)
         return ExitCode.child_process_error
-    elif not options.verbose:
+    elif exc_name == 'ocrmypdf.main.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..]
+            """))
+        return ExitCode.input_file
+    elif exc_name == 'PyPDF2.utils.PdfReadError' and \
+            'not been decrypted' in exc_value:
+        _log.error(textwrap.dedent("""\
+            Input PDF uses either an encryption algorithm or a PDF security
+            handler that is not supported by ocrmypdf.
+
+            For information about this PDF's security use
+                qpdf --show-encryption [...input PDF...]
+
+            (Only algorithms "R = 1" and "R = 2" are supported.)
+
+            """))
+        return ExitCode.encrypted_pdf
+
+    if not options.verbose:
         _log.error(exc_stack)
-        return ExitCode.other_error
+    return ExitCode.other_error
 
 
-def traverse_ruffus_exception(e):
+def traverse_ruffus_exception(e_args):
     """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[0], str) and len(e) == 5:
-        return do_ruffus_exception(e)
-    elif hasattr(e, '__iter__'):
-        for exc in e:
+    if isinstance(e_args, Sequence) and isinstance(e_args[0], str) and \
+            len(e_args) == 5:
+        return do_ruffus_exception(e_args)
+    elif is_iterable_notstr(e_args):
+        for exc in e_args:
             return traverse_ruffus_exception(exc)
 
 
@@ -1113,9 +1334,17 @@ def run_pipeline():
         _log.error(e)
         return ExitCode.other_error
 
-    if not validate_pdfa(options.output_file, _log):
-        _log.warning('Output file: The generated PDF/A file is INVALID')
-        return ExitCode.invalid_output_pdfa
+    if options.output_type == 'pdfa':
+        pdfa_info = file_claims_pdfa(options.output_file)
+        if pdfa_info['pass']:
+            _log.info(pdfa_info['message'])
+        else:
+            _log.warning(pdfa_info['message'])
+            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
 
     with _pdfinfo_lock:
         _log.debug(_pdfinfo)

ocrmypdf/pageinfo.py

@@ -93,7 +93,7 @@ def _interpret_contents(contentstream):
             ctm = stack.pop()
         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(
@@ -230,6 +230,14 @@ 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:

ocrmypdf/pdfa.py

@@ -7,8 +7,9 @@ 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)
@@ -74,8 +75,14 @@ def encode_text_string(s: str) -> str:
     Postscript file to be completely ASCII and no escaping of Postscript
     characters is necessary.
     '''
+
+    # Sometimes lazy C programmer leave their NULs at the end of strings
+    # 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()
@@ -88,11 +95,11 @@ def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
     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', ''))
+                          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', ''))
     return result
 
 
@@ -104,7 +111,48 @@ 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', 'message': 'No XMP metadata'}
+
+    pdfa_dict = {attr.localName: attr.value for attr in pdfa_nodes}
+    pdfa_dict['pass'] = False
+    pdfa_dict['output'] = 'pdf'
+    if pdfa_dict:
+        part_conformance = pdfa_dict['part'] + pdfa_dict['conformance']
+        valid_part_conforms = {'1A', '1B', '2A', '2B', '2U', '3A', '3B', '3U'}
+
+        message = 'File claims to be PDF/A-{}'.format(
+            part_conformance)
+
+        if part_conformance in valid_part_conforms:
+            pdfa_dict['pass'] = True
+            pdfa_dict['output'] = 'pdfa'
+        pdfa_dict['message'] = message
+    else:
+        pdfa_dict['message'] = 'File is a regular PDF'
+
+    return pdfa_dict
+

ocrmypdf/qpdf.py

@@ -2,12 +2,32 @@
 # © 2015 James R. Barlow: github.com/jbarlow83
 
 from subprocess import CalledProcessError, check_output, STDOUT, check_call
+from functools import lru_cache
 import sys
 import os
+import re
 
 from . import ExitCode, get_program
 
 
+@lru_cache(maxsize=1)
+def version():
+    args_qpdf = [
+        get_program('qpdf'),
+        '--version'
+    ]
+    try:
+        versions = check_output(
+            args_qpdf, close_fds=True, universal_newlines=True,
+            stderr=STDOUT)
+    except CalledProcessError:
+        print("Could not find qpdf executable on system PATH.")
+        sys.exit(ExitCode.missing_dependency)
+
+    qpdf_version = re.match(r'qpdf version (.+)', versions).group(1)
+    return qpdf_version
+
+
 def check(input_file, log):
     args_qpdf = [
         get_program('qpdf'),
@@ -20,7 +40,7 @@ def check(input_file, log):
     except CalledProcessError as e:
         if e.returncode == 2:
             log.error("{0}: not a valid PDF, and could not repair it.".format(
-                    input_file))
+                input_file))
             log.error("Details:")
             log.error(e.output)
         elif e.returncode == 3:
@@ -85,3 +105,15 @@ def split_pages(input_file, work_folder, npages):
             os.path.join(work_folder, '{0:06d}.page.pdf'.format(n + 1))
         ]
         check_call(args_qpdf)
+
+
+def merge(input_files, output_file):
+    """Merge the list of input files (all filenames) into the output file.
+
+    The input files may contain one or more pages.
+    """
+    args_qpdf = [
+        get_program('qpdf'), input_files[0], '--pages'
+    ] + input_files + ['--', output_file]
+    check_call(args_qpdf)
+

requirements.txt

@@ -1,6 +1,9 @@
+# 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==3.1.1
+Pillow==3.3.0
 reportlab==3.2.0
-PyPDF2==1.25.1
-img2pdf==0.2
+PyPDF2==1.26
+img2pdf==0.2.1
 cffi==1.5.2

setup.py

@@ -154,23 +154,23 @@ if command.startswith('install') or \
         command in ['check', 'test', 'nosetests', 'easy_install']:
     check_external_program(
         program='tesseract',
-        need_version='3.02.02',
+        need_version='3.03',  # limited by Travis CI / Ubuntu 12.04 backports
         package={'darwin': 'tesseract', 'linux': 'tesseract-ocr'}
     )
     check_external_program(
         program='gs',
-        need_version='9.14',
+        need_version='9.15',  # limited by Travis CI / Ubuntu 12.04 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='5.0.0',   # limited by Travis CI / Ubuntu 12.04 backports
         package='qpdf',
         version_check_args=['--version']
     )
@@ -217,12 +217,12 @@ setup(
         'ocrmypdf/lib/compile_leptonica.py:ffi'
     ],
     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>=3.1.0',        # Pillow is pretty stable
+        'reportlab>=3.2.0',     # oldest released version with sane image handling
+        'PyPDF2>=1.26',         # pure Python, so track HEAD closely
+        'img2pdf>=0.2.1',       # pure Python, so track HEAD closely
+        'cffi>=1.5.0'           # oldest version ever tested
     ],
     tests_require=tests_require,
     entry_points={
@@ -230,6 +230,6 @@ setup(
             '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)

tests/resources/README.rst

@@ -14,9 +14,9 @@ In some cases they were converted from one image format to another without other
 +=====================+================================================================================+
 | c02-22.pdf          | `Project Gutenberg`_, Adventures of Huckleberry Finn, page 22                  |
 +---------------------+--------------------------------------------------------------------------------+
-| congress.jpg        | `US Congressional Records`_                                                    |
+| congress.jpg        | `US Congressional Records`_ (Public Domain)                                    |
 +---------------------+--------------------------------------------------------------------------------+
-| graph.pdf           | `Wikimedia: Pandas text analysis.png`_                                         |
+| graph.pdf           | `Wikimedia: Pandas text analysis.png`_ (Public Domain)                         |
 +---------------------+--------------------------------------------------------------------------------+
 | lichtenstein.pdf    | `Wikimedia: JPEG2000 Lichtenstein`_ (Creative Commons BY-SA 3.0)               |
 +---------------------+--------------------------------------------------------------------------------+
@@ -31,6 +31,7 @@ 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.
 
+- aspect.pdf (a page with an image with 200 x 100 DPI resolution)
 - blank.pdf (a blank PDF page)
 - cmyk.pdf (a CMYK image created in Photoshop)
 - enormous.pdf (a very lage page)
@@ -47,12 +48,13 @@ These test resources are assemblies from other previously mentioned files, relea
 
 - cardinal.pdf (four cardinal directions, 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)
+- skew-encrypted.pdf (skew.pdf with encryption - access supported by PyPDF2)
 
 
 .. _`Wikimedia: LinnSequencer`: https://upload.wikimedia.org/wikipedia/en/b/b7/LinnSequencer_hardware_MIDI_sequencer_brochure_page_2_300dpi.jpg
@@ -63,4 +65,4 @@ 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

tests/test_main.py

@@ -12,6 +12,7 @@ from ocrmypdf.pageinfo import pdf_get_all_pageinfo
 import PyPDF2 as pypdf
 from ocrmypdf import ExitCode
 from ocrmypdf import leptonica
+from ocrmypdf.pdfa import file_claims_pdfa
 
 
 if sys.version_info.major < 3:
@@ -159,34 +160,34 @@ def test_deskew(spoof_tesseract_noop):
 
 
 def test_clean(spoof_tesseract_noop):
-    check_ocrmypdf('skew.pdf', 'test_clean.pdf', '-c', env=spoof_tesseract_noop)
+    check_ocrmypdf('skew.pdf', 'test_clean.pdf', '-c',
+                   env=spoof_tesseract_noop)
 
 
-@pytest.mark.parametrize("pdf,renderer", [
-    ('palette.pdf', 'hocr'),
-    ('palette.pdf', 'tesseract'),
-    ('cmyk.pdf', 'hocr'),
-    ('cmyk.pdf', 'tesseract'),
-    ('ccitt.pdf', 'hocr'),
-    ('ccitt.pdf', 'tesseract'),
-    ('jbig2.pdf', 'hocr'),
-    ('jbig2.pdf', 'tesseract'),
-    ('lichtenstein.pdf', 'hocr'),
-    ('lichtenstein.pdf', 'tesseract')
-])
-def test_exotic_image(spoof_tesseract_cache, pdf, renderer):
+# This will run 5 * 2 * 2 = 20 test cases
+@pytest.mark.parametrize(
+    "pdf",
+    ['palette.pdf', 'cmyk.pdf', 'ccitt.pdf', 'jbig2.pdf', 'lichtenstein.pdf'])
+@pytest.mark.parametrize("renderer", ['hocr', 'tesseract'])
+@pytest.mark.parametrize("output_type", ['pdf', 'pdfa'])
+def test_exotic_image(spoof_tesseract_cache, pdf, renderer, output_type):
     check_ocrmypdf(
         pdf,
         'test_{0}_{1}.pdf'.format(pdf, renderer),
         '-dc',
         '-v', '1',
+        '--output-type', output_type,
         '--pdf-renderer', renderer, env=spoof_tesseract_cache)
 
 
-def test_preserve_metadata(spoof_tesseract_noop):
+@pytest.mark.parametrize("output_type", [
+    'pdfa', 'pdf'
+    ])
+def test_preserve_metadata(spoof_tesseract_noop, output_type):
     pdf_before = pypdf.PdfFileReader(_infile('graph.pdf'))
 
     output = check_ocrmypdf('graph.pdf', 'test_metadata_preserve.pdf',
+                            '--output-type', output_type,
                             env=spoof_tesseract_noop)
 
     pdf_after = pypdf.PdfFileReader(output)
@@ -194,8 +195,14 @@ def test_preserve_metadata(spoof_tesseract_noop):
     for key in ('/Title', '/Author'):
         assert pdf_before.documentInfo[key] == pdf_after.documentInfo[key]
 
+    pdfa_info = file_claims_pdfa(output)
+    assert pdfa_info['output'] == output_type
 
-def test_override_metadata(spoof_tesseract_noop):
+
+@pytest.mark.parametrize("output_type", [
+    'pdfa', 'pdf'
+    ])
+def test_override_metadata(spoof_tesseract_noop, output_type):
     input_file = _infile('c02-22.pdf')
     output_file = _outfile('test_override_metadata.pdf')
 
@@ -208,6 +215,7 @@ def test_override_metadata(spoof_tesseract_noop):
         '--title', german,
         '--author', chinese,
         '--subject', high_unicode,
+        '--output-type', output_type,
         env=spoof_tesseract_noop)
 
     assert p.returncode == ExitCode.ok
@@ -226,6 +234,9 @@ def test_override_metadata(spoof_tesseract_noop):
     assert pdfinfo['Subject'] == high_unicode
     assert pdfinfo.get('Keywords', '') == ''
 
+    pdfa_info = file_claims_pdfa(output_file)
+    assert pdfa_info['output'] == output_type
+
 
 @pytest.mark.parametrize('renderer', [
     'hocr',
@@ -386,16 +397,15 @@ def test_skip_big(spoof_tesseract_cache):
     assert not pdfinfo[0]['has_text']
 
 
-@pytest.mark.parametrize('renderer', [
-    'hocr',
-    'tesseract',
-    ])
-def test_maximum_options(spoof_tesseract_cache, renderer):
+@pytest.mark.parametrize('renderer', ['hocr', 'tesseract'])
+@pytest.mark.parametrize('output_type', ['pdf', 'pdfa'])
+def test_maximum_options(spoof_tesseract_cache, renderer, output_type):
     check_ocrmypdf(
         'multipage.pdf', 'test_multipage%s.pdf' % renderer,
         '-d', '-c', '-i', '-g', '-f', '-k', '--oversample', '300',
         '--skip-big', '10', '--title', 'Too Many Weird Files',
         '--author', 'py.test', '--pdf-renderer', renderer,
+        '--output-type', output_type,
         env=spoof_tesseract_cache)
 
 
@@ -420,6 +430,17 @@ def test_blank_input_pdf():
     assert p.returncode == ExitCode.ok
 
 
+def test_force_ocr_on_pdf_with_no_images(spoof_tesseract_crash):
+    # As a correctness test, make sure that --force-ocr on a PDF with no
+    # content still triggers tesseract. If tesseract crashes, then it was
+    # called.
+    p, _, err = run_ocrmypdf_env(
+        'blank.pdf', 'wont_be_created.pdf', '--force-ocr',
+        env=spoof_tesseract_crash)
+    assert p.returncode == ExitCode.child_process_error, err
+    assert not os.path.exists(_outfile('wontwork.pdf'))
+
+
 def test_french(spoof_tesseract_cache):
     p, out, err = run_ocrmypdf_env(
         'francais.pdf', 'francais.pdf', '-l', 'fra', env=spoof_tesseract_cache)
@@ -547,3 +568,32 @@ def test_old_unpaper():
         'c02-22.pdf', 'wont_be_created.pdf', '--clean', env=env)
     assert sh.returncode == ExitCode.missing_dependency
 
+
+def test_algo4():
+    sh, _, _ = run_ocrmypdf_env('encrypted_algo4.pdf', 'wontwork.pdf')
+    assert sh.returncode == ExitCode.encrypted_pdf
+
+
+@pytest.mark.parametrize('renderer', [
+    'hocr'])  # tesseract cannot pass this test yet
+def test_non_square_resolution(renderer, spoof_tesseract_cache):
+    # Confirm input image is non-square resolution
+    in_pageinfo = pdf_get_all_pageinfo(_infile('aspect.pdf'))
+    assert in_pageinfo[0]['xres'] != in_pageinfo[0]['yres']
+
+    out = 'aspect_%s.pdf' % renderer
+    check_ocrmypdf(
+        'aspect.pdf', out,
+        '--pdf-renderer', renderer, env=spoof_tesseract_cache)
+
+    out_pageinfo = pdf_get_all_pageinfo(_outfile(out))
+
+    # Confirm resolution was kept the same
+    assert in_pageinfo[0]['xres'] == out_pageinfo[0]['xres']
+    assert in_pageinfo[0]['yres'] == out_pageinfo[0]['yres']
+
+
+def test_image_to_pdf(spoof_tesseract_noop):
+    check_ocrmypdf(
+        'LinnSequencer.jpg', 'image_to_pdf.pdf', '--image-dpi', '200',
+        env=spoof_tesseract_noop)