Merge branch 'release/v3.2'

Now checks input image to ensure the implied page size of its .hocr file
matches the rest of the PDF.

.dockerignore

@@ -1,21 +1,31 @@
-bin/
-build/
-dist/
-include/
-lib/
-ocrmypdf.egg-info/
-staging/
-.git/
-.ruffus_history.sqlite
-MANIFEST.in
-*.sublime*
+*.ipynb
 *.pdf
-*.rst
 *.pyc
+*.rst
+*.sublime*
 */*.pyc
 */*/*.pyc
 */*/*/*.pyc
 */*/*/*/*.pyc
 */*/*/*/*/*.pyc
 */*/*/*/*/*/*.pyc
-*/*/*/*/*/*/*/*.pyc
+*/*/*/*/*/*/*/*.pyc
+.cache/
+.git/
+.ipynb_checkpoints/
+.ruffus_history.sqlite
+bin/
+build/
+dist/
+htmlcov/
+include/
+lib/
+MANIFEST.in
+ocrmypdf.egg-info/
+staging/
+tests/cache/
+tests/output/
+tests/resources/private/
+tmp/
+venv-3.4/
+venv-3.5/

.gitignore

@@ -1,17 +1,31 @@
-tmp/
-log/
+# Development environment
 *.pyc
-tests/output/
-.ruffus_history.sqlite
 *.sublime-*
-/*.pdf
+venv-3.4/
+venv-3.5/
+venv/
+pyvenv.cfg
+
+# Package building
+*.egg-info/
+.cache/
+.eggs/
 build/
 dist/
-*.egg-info/
-venv/
-*/test/output
-bin/
-include/
-lib/
-pip-selfcheck.json
-pyvenv.cfg
+
+# Automatically generated files
+ocrmypdf/lib/_*.py
+ocrmypdf/version.py
+
+# Code coverage
+.coverage
+htmlcov/
+
+# Testing
+log/
+/*.pdf
+.ipynb_checkpoints/
+tests/cache/
+tests/output/
+tests/resources/private
+tmp/

.travis.yml

@@ -0,0 +1,42 @@
+language: python
+
+cache:
+    directories:
+        - $HOME/.cache/pip
+        - $HOME/.ccache
+        - tarballs
+        - tests/cache
+
+python:
+    - 3.4
+
+before_install:
+    # Ubuntu packages
+    - sudo add-apt-repository ppa:evl.ms/precise -y  # for Ghostscript 9.15
+    - sudo add-apt-repository ppa:lyrasis/precise-backports -y  # for Tesseract 3.03
+    - sudo add-apt-repository ppa:b-eltzner/qpdfview-exp -y  # for QPDF 5
+    - sudo add-apt-repository ppa:itachi-san/ffmpeg -y  # for libav 11.2 (for unpaper)
+    - sudo apt-get update -qq  # must go after all add-apt-repo
+    - sudo apt-get install -y ghostscript tesseract-ocr tesseract-ocr-deu tesseract-ocr-eng tesseract-ocr-fra qpdf poppler-utils gcc libavformat-dev libavcodec-dev libavutil-dev automake make pkg-config xsltproc 
+
+    # pip
+    - pip install --upgrade pip
+
+    # Download, make and install unpaper (using ccache)
+    - mkdir -p tarballs
+    - "[ -f tarballs/unpaper-6.1.tar.xz ] || wget -q https://www.flameeyes.eu/files/unpaper-6.1.tar.xz -O tarballs/unpaper-6.1.tar.xz"
+    - tar -xvf tarballs/unpaper-6.1.tar.xz
+    - export PATH="/usr/lib/ccache:$PATH"
+    - pushd unpaper-6.1 && ./configure --prefix=/usr && make -j && sudo make install && popd
+
+install:
+    - pip install -r requirements.txt
+    - pip install -r test_requirements.txt
+
+script:
+    - python setup.py clean
+    - python setup.py install
+    - py.test
+
+os:
+    - linux

Dockerfile

@@ -1,6 +1,6 @@
 # OCRmyPDF
 #
-# VERSION               3.0.0
+# VERSION               3.0.2
 FROM      debian:stretch
 MAINTAINER James R. Barlow <jim@purplerock.ca>
 
@@ -21,7 +21,9 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
   python3-pip \
   python3-venv \
   python3-reportlab \
-  python3-pil
+  python3-pil \
+  python3-wheel \
+  unpaper
 
 # Enforce UTF-8
 # Borrowed from https://index.docker.io/u/crosbymichael/python/ 
@@ -30,36 +32,9 @@ RUN dpkg-reconfigure locales && \
   /usr/sbin/update-locale LANG=C.UTF-8
 ENV LC_ALL C.UTF-8
 
-# Build unpaper 6.1
-RUN apt-get install -y \
-  wget \
-  gcc \
-  libavformat-dev \
-  libavcodec-dev \
-  libavutil-dev \
-  autoconf \
-  automake \
-  make \
-  pkg-config \
-  xsltproc
-
-WORKDIR /root
-RUN wget https://github.com/Flameeyes/unpaper/archive/unpaper-6.1.tar.gz
-RUN tar xf unpaper-6.1.tar.gz
-WORKDIR /root/unpaper-unpaper-6.1
-RUN autoreconf -i
-RUN ./configure CFLAGS="-O2 -march=native -pipe -flto"
-RUN make -j install
-
-RUN apt-get remove -y \
-  gcc \
-  autoconf \
-  automake \
-  pkg-config \
-  xsltproc \
-  make
+# Remove the junk
 RUN apt-get autoremove -y && apt-get clean -y
-RUN rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
+RUN rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* /root/*
 
 # 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
@@ -78,7 +53,6 @@ RUN . /appenv/bin/activate; \
 USER docker
 WORKDIR /home/docker
 
-ENV DEFAULT_RUFFUS_HISTORY_FILE=/tmp/.{basename}.ruffus_history.sqlite
 ENV OCRMYPDF_TEST_OUTPUT=/tmp/test-output
 ENV OCRMYPDF_IN_DOCKER=1
 

Dockerfile.polyglot

@@ -0,0 +1,13 @@
+# OCRmyPDF polyglot
+#
+# VERSION               3.0.2
+FROM      jbarlow83/ocrmypdf:latest
+MAINTAINER James R. Barlow <jim@purplerock.ca>
+
+# Update system and install our dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  tesseract-ocr-all 
+
+# Must use array form of ENTRYPOINT
+# Non-array form does not append other arguments, because that is "intuitive"
+ENTRYPOINT ["/application/docker-wrapper.sh"]

README.rst

@@ -9,24 +9,20 @@ Main features
 
 -  Generates a searchable
    `PDF/A <https://en.wikipedia.org/?title=PDF/A>`__ file from a regular PDF
-   only containing images
--  Places OCRed text accurately below the image to ease copy / paste
+-  Places OCR text accurately below the image to ease copy / paste
 -  Keeps the exact resolution of the original embedded images
-
-   -  or if requested oversamples the images before OCRing so as to get
-      better results
-
--  When possible, copies input images directly to output without transcoding them,
-   to preserve image quality
+-  When possible, inserts OCR information as a "lossless" operation without rendering vector information
 -  Keeps file size about the same
 -  If requested deskews and/or cleans the image before performing OCR
 -  Validates input and output files
 -  Provides debug mode to enable easy verification of the OCR results
--  Processes several pages in parallel when more than one CPU core is
+-  Processes pages in parallel when more than one CPU core is
    available
--  Uses Tesseract OCR engine
+-  Uses `Tesseract OCR <https://github.com/tesseract-ocr/tesseract>`__ engine
+-  Supports 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
 ----------
@@ -90,16 +86,18 @@ To execute the OCRmyPDF on a local file, you must `provide a writable volume to
 
    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``:: 
 
    docker run -v "$(pwd):/home/docker"   ocrmypdf --skip-text test.pdf output.pdf
 
-Note that `ocrmypdf` has its own separate -v argument to control debug verbosity. All Docker arguments should before the `ocrmypdf` container name and all arguments to `ocrmypdf` should be listed after.
+Note that ``ocrmypdf`` has its own separate ``-v VERBOSITYLEVEL`` argument to control debug verbosity. All Docker arguments should before the ``ocrmypdf`` container name and all arguments to ``ocrmypdf`` should be listed after.
 
-Installing on Mac OS X Yosemite
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Installing on Mac OS X
+~~~~~~~~~~~~~~~~~~~~~~
 
-If it's not already present, `install Homebrew <http://brew.sh/>`__
+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::
 
@@ -120,7 +118,7 @@ It is also recommended that install Pillow and confirm it can read and write JPE
    pip3 install --upgrade pip
    pip3 install --upgrade pillow
 
-To test that your Python imaging library (Pillow) can access JPEG and PNG files, try this command::
+Sometimes, the Python imaging library (Pillow) can end up being compiled and installed without support for JPEG and PNG files. (Arguably, this is an unfixed bug in Pillow's installer.) To confirm that Pillow is compiled correctly and can access JPEG and PNG files, try this command::
 
    python3 -c "from PIL import Image; im = Image.new('1', (1, 1)); im.save('test.png'); im.save('test.jpg')"
 
@@ -137,7 +135,7 @@ The command line program should now be available::
 Installing on Ubuntu 14.04 LTS
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-Installing on Ubuntu 14.04 LTS (trusty) is more difficult than other options, because of certain bugs in package installation.
+Installing on Ubuntu 14.04 LTS (trusty) is more difficult than other options, because of certain bugs in Python package installation.
 
 Update apt-get::
 
@@ -176,22 +174,38 @@ package <http://www.thefourtheye.in/2014/12/Python-venv-problem-with-ensurepip-i
    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.
+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
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-To install the HEAD revision from sources in development mode::
+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.
+
+First, clone the HEAD revision::
 
    git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
    cd OCRmyPDF
+
+To install the HEAD revision from sources::
+
+   pip3 install .
+
+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 .
    
-On certain Linux/UNIX platforms such as Ubuntu, you may need to use 
+On certain Linux distributions such as Ubuntu, you may need to use 
 run the install command as superuser::
 
-   sudo pip3 install -e .
+   sudo pip3 install [-e] .
    
 Note that this will alter your system's Python distribution. If you prefer 
 to not install as superuser, you can install the package in a Python virtual environment::
@@ -200,10 +214,10 @@ to not install as superuser, you can install the package in a Python virtual env
    pyvenv venv
    source venv/bin/activate
    cd OCRmyPDF
-   pip3 install -e .
+   pip3 install .
 
-If your platform does not have ``pip3``, make sure that Python 3.4+ and the `pip` 
-package are installed.
+However, ``ocrmypdf`` will only be accessible on the system PATH after
+you activate the virtual environment.
 
 To run the program::
    
@@ -221,13 +235,12 @@ In case you detect an issue, please:
 
 -  Check if your issue is already known
 -  If no problem report exists on github, please create one here:
-   https://github.com/fritz-hh/OCRmyPDF/issues
+   https://github.com/jbarlow83/OCRmyPDF/issues
 -  Describe your problem thoroughly
 -  Append the console output of the script when running the debug mode
-   (-v 1 option)
+   (``-v 1`` option)
 -  If possible provide your input PDF file as well as the content of the
-   temporary folder (using a file sharing service like
-   www.file-upload.net)
+   temporary folder (using a file sharing service like Dropbox)
 
 Press & Media
 -------------

RELEASE_NOTES.rst

@@ -5,6 +5,55 @@ Please always read this file before installing the package
 
 Download software here: https://github.com/jbarlow83/OCRmyPDF/tags
 
+
+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:
 =====
 
@@ -36,6 +85,8 @@ 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

ocrmypdf/__init__.py

@@ -1,4 +1,5 @@
 from enum import IntEnum
+import os
 
 
 class ExitCode(IntEnum):
@@ -10,3 +11,8 @@ class ExitCode(IntEnum):
     file_access_error = 5
     already_done_ocr = 6
     other_error = 15
+
+
+def get_program(name):
+    envvar = 'OCRMYPDF_' + name.upper()
+    return os.environ.get(envvar, name)

ocrmypdf/ghostscript.py

@@ -4,12 +4,13 @@
 from tempfile import NamedTemporaryFile
 from subprocess import Popen, PIPE, check_call
 from shutil import copy
+from . import get_program
 
 
 def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log):
     with NamedTemporaryFile(delete=True) as tmp:
         args_gs = [
-            'gs',
+            get_program('gs'),
             '-dQUIET',
             '-dBATCH',
             '-dNOPAUSE',
@@ -36,7 +37,7 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log):
 def generate_pdfa(pdf_pages, output_file, threads=1):
     with NamedTemporaryFile(delete=True) as gs_pdf:
         args_gs = [
-            "gs",
+            get_program("gs"),
             "-dQUIET",
             "-dBATCH",
             "-dNOPAUSE",
@@ -44,7 +45,8 @@ def generate_pdfa(pdf_pages, output_file, threads=1):
             "-sDEVICE=pdfwrite",
             "-sColorConversionStrategy=/RGB",
             "-sProcessColorModel=DeviceRGB",
-            "-dPDFA",
+            "-dJPEGQ=95",
+            "-dPDFA=2",
             "-sPDFACompatibilityPolicy=2",
             "-sOutputICCProfile=srgb.icc",
             "-sOutputFile=" + gs_pdf.name,

ocrmypdf/hocrtransform.py

@@ -1,4 +1,4 @@
-#!/usr/local/bin/python3
+#!/usr/bin/env python3
 ##############################################################################
 # Copyright (c) 2013-14: fritz-hh from Github
 #   (https://github.com/fritz-hh)

ocrmypdf/main.py

@@ -2,7 +2,7 @@
 # © 2015 James R. Barlow: github.com/jbarlow83
 
 from contextlib import suppress
-from tempfile import NamedTemporaryFile, mkdtemp
+from tempfile import mkdtemp
 import sys
 import os
 import re
@@ -11,17 +11,12 @@ import warnings
 import multiprocessing
 import atexit
 import textwrap
+import img2pdf
 
 import PyPDF2 as pypdf
 from PIL import Image
 
-from subprocess import Popen, check_call, PIPE, CalledProcessError, \
-    TimeoutExpired, check_output, STDOUT
-try:
-    from subprocess import DEVNULL
-except ImportError:
-    DEVNULL = open(os.devnull, 'wb')
-
+from functools import partial
 
 from ruffus import transform, suffix, merge, active_if, regex, jobs_limit, \
     formatter, follows, split, collate, check_if_uptodate
@@ -33,15 +28,17 @@ from .pageinfo import pdf_get_all_pageinfo
 from .pdfa import generate_pdfa_def
 from . import ghostscript
 from . import tesseract
+from . import qpdf
 from . import ExitCode
 
+import pkg_resources
+
+VERSION = pkg_resources.get_distribution('ocrmypdf').version
+
 warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
 
 
 BASEDIR = os.path.dirname(os.path.realpath(__file__))
-JHOVE_PATH = os.path.realpath(os.path.join(BASEDIR, 'jhove'))
-JHOVE_JAR = os.path.join(JHOVE_PATH, 'bin', 'JhoveApp.jar')
-JHOVE_CFG = os.path.join(JHOVE_PATH, 'conf', 'jhove.conf')
 
 
 # -------------
@@ -103,12 +100,12 @@ check_pil_encoder('zlib', 'PNG')
 parser = cmdline.get_argparse(
     prog="ocrmypdf",
     description="Generate searchable PDF file from an image-only PDF file.",
-    version='3.0',
+    version=VERSION,
     fromfile_prefix_chars='@',
     ignored_args=[
         'touch_files_only', 'recreate_database', 'checksum_file_name',
         'key_legend_in_graph', 'draw_graph_horizontally', 'flowchart_format',
-        'forced_tasks', 'target_tasks'])
+        'forced_tasks', 'target_tasks', 'use_threads', 'jobs'])
 
 parser.add_argument(
     'input_file',
@@ -119,6 +116,9 @@ parser.add_argument(
 parser.add_argument(
     '-l', '--language', action='append',
     help="languages of the file to be OCRed")
+parser.add_argument(
+    '-j', '--jobs', metavar='N', type=int,
+    help="Use up to N CPU cores simultaneously (default: use all)")
 
 metadata = parser.add_argument_group(
     "Metadata options",
@@ -173,13 +173,16 @@ advanced = parser.add_argument_group(
     "Advanced",
     "Advanced options for power users")
 advanced.add_argument(
-    '--tesseract-config', default=[], type=list, action='append',
+    '--tesseract-config', action='append', metavar='CFG', default=[],
     help="additional Tesseract configuration files")
 advanced.add_argument(
-    '--pdf-renderer', choices=['tesseract', 'hocr'], default='hocr',
+    '--tesseract-pagesegmode', action='store', type=int, metavar='PSM',
+    help="set Tesseract page segmentation mode (see tesseract --help)")
+advanced.add_argument(
+    '--pdf-renderer', choices=['auto', 'tesseract', 'hocr'], default='auto',
     help='choose OCR PDF renderer')
 advanced.add_argument(
-    '--tesseract-timeout', default=180.0, type=float,
+    '--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')
 
@@ -218,6 +221,8 @@ if not set(options.language).issubset(tesseract.languages()):
 # ----------
 # Arguments
 
+if options.pdf_renderer == 'auto':
+    options.pdf_renderer = 'hocr'
 
 if any((options.deskew, options.clean, options.clean_final)):
     try:
@@ -245,6 +250,10 @@ 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.")
 
+lossless_reconstruction = False
+if options.pdf_renderer == 'hocr':
+    if not options.deskew and not options.clean_final and not options.force_ocr:
+        lossless_reconstruction = True
 
 # ----------
 # Logging
@@ -341,9 +350,8 @@ def cleanup_working_files(*args):
 
 @transform(
     input=options.input_file,
-    filter=suffix('.pdf'),
-    output='.repaired.pdf',
-    output_dir=work_folder,
+    filter=formatter('(?i)\.pdf'),
+    output=work_folder + '{basename[0]}.repaired.pdf',
     extras=[_log, _pdfinfo, _pdfinfo_lock])
 def repair_pdf(
         input_file,
@@ -351,23 +359,8 @@ def repair_pdf(
         log,
         pdfinfo,
         pdfinfo_lock):
-    args_qpdf = [
-        'qpdf', input_file, output_file
-    ]
-    try:
-        out = check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 2:
-            print("{0}: not a valid PDF, and could not repair it.".format(
-                    options.input_file))
-            print("Details:")
-            print(e.output)
-        else:
-            print(e.output)
-        sys.exit(ExitCode.input_file)
-
-    log.debug(out)
 
+    qpdf.repair(input_file, output_file, log)
     with pdfinfo_lock:
         pdfinfo.extend(pdf_get_all_pageinfo(output_file))
         log.info(pdfinfo)
@@ -432,16 +425,8 @@ def split_pages(
         with suppress(FileNotFoundError):
             os.unlink(oo)
 
-    pages = check_output(['qpdf', '--show-npages', input_file],
-                         universal_newlines=True, close_fds=True)
-
-    for n in range(int(pages)):
-        args_qpdf = [
-            'qpdf', input_file,
-            '--pages', input_file, '{0}'.format(n + 1), '--',
-            os.path.join(work_folder, '{0:06d}.page.pdf'.format(n + 1))
-        ]
-        check_call(args_qpdf)
+    npages = qpdf.get_npages(input_file)
+    qpdf.split_pages(input_file, work_folder, npages)
 
     from glob import glob
     for filename in glob(os.path.join(work_folder, '*.page.pdf')):
@@ -468,7 +453,6 @@ def rasterize_with_ghostscript(
         log,
         pdfinfo,
         pdfinfo_lock):
-
     pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
 
     device = 'png16m'  # 24-bit
@@ -547,60 +531,17 @@ def ocr_tesseract_hocr(
         pdfinfo,
         pdfinfo_lock):
 
-    pageinfo = get_pageinfo(input_file, pdfinfo, pdfinfo_lock)
-
-    badxml = os.path.splitext(output_file)[0] + '.badxml'
-
-    args_tesseract = [
-        'tesseract',
-        '-l', '+'.join(options.language),
-        input_file,
-        badxml,
-        'hocr'
-    ] + options.tesseract_config
-    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
-              universal_newlines=True)
-    try:
-        stdout, stderr = p.communicate(timeout=options.tesseract_timeout)
-    except TimeoutExpired:
-        p.kill()
-        stdout, stderr = p.communicate()
-        # Generate a HOCR file with no recognized text if tesseract times out
-        # Temporary workaround to hocrTransform not being able to function if
-        # it does not have a valid hOCR file.
-        with open(output_file, 'w', encoding="utf-8") as f:
-            f.write(tesseract.HOCR_TEMPLATE.format(
-                pageinfo['width_pixels'],
-                pageinfo['height_pixels']))
-    else:
-        if stdout:
-            log.info(stdout)
-        if stderr:
-            log.error(stderr)
-
-        if p.returncode != 0:
-            raise CalledProcessError(p.returncode, args_tesseract)
-
-        if os.path.exists(badxml + '.html'):
-            # Tesseract 3.02 appends suffix ".html" on its own (.badxml.html)
-            shutil.move(badxml + '.html', badxml)
-        elif os.path.exists(badxml + '.hocr'):
-            # Tesseract 3.03 appends suffix ".hocr" on its own (.badxml.hocr)
-            shutil.move(badxml + '.hocr', badxml)
-
-        # Tesseract 3.03 inserts source filename into hocr file without
-        # escaping it, creating invalid XML and breaking the parser.
-        # As a workaround, rewrite the hocr file, replacing the filename
-        # with a space.  Don't know if Tesseract 3.02 does the same.
-
-        regex_nested_single_quotes = re.compile(
-            r"""title='image "([^"]*)";""")
-        with open(badxml, mode='r', encoding='utf-8') as f_in, \
-                open(output_file, mode='w', encoding='utf-8') as f_out:
-            for line in f_in:
-                line = regex_nested_single_quotes.sub(
-                    r"""title='image " ";""", line)
-                f_out.write(line)
+    tesseract.generate_hocr(
+        input_file=input_file,
+        output_hocr=output_file,
+        language=options.language,
+        tessconfig=options.tesseract_config,
+        timeout=options.tesseract_timeout,
+        pageinfo_getter=partial(get_pageinfo, input_file, pdfinfo,
+                                pdfinfo_lock),
+        pagesegmode=options.tesseract_pagesegmode,
+        log=log
+        )
 
 
 @active_if(options.pdf_renderer == 'hocr')
@@ -633,24 +574,47 @@ def select_image_for_pdf(
 
 @active_if(options.pdf_renderer == 'hocr')
 @collate(
-    input=[select_image_for_pdf, ocr_tesseract_hocr],
-    filter=regex(r".*/(\d{6})(?:\.image|\.hocr)"),
-    output=os.path.join(work_folder, r'\1.rendered.pdf'),
+    input=[select_image_for_pdf, split_pages],
+    filter=regex(r".*/(\d{6})(?:\.image|\.ocr\.page\.pdf)"),
+    output=os.path.join(work_folder, r'\1.image-layer.pdf'),
     extras=[_log, _pdfinfo, _pdfinfo_lock])
-def render_hocr_page(
+def select_image_layer(
         infiles,
         output_file,
         log,
         pdfinfo,
         pdfinfo_lock):
-    hocr = next(ii for ii in infiles if ii.endswith('.hocr'))
+
+    page_pdf = next(ii for ii in infiles if ii.endswith('.page.pdf'))
     image = next(ii for ii in infiles if ii.endswith('.image'))
 
-    pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
+    if lossless_reconstruction:
+        re_symlink(page_pdf, output_file)
+    else:
+        pageinfo = get_pageinfo(image, pdfinfo, pdfinfo_lock)
+        dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
+        with open(output_file, 'wb') as pdf:
+            img2pdf.convert([image], dpi=dpi, outputstream=pdf)
+
+
+@active_if(options.pdf_renderer == 'hocr')
+@transform(
+    input=ocr_tesseract_hocr,
+    filter=suffix('.hocr'),
+    output='.hocr.pdf',
+    extras=[_log, _pdfinfo, _pdfinfo_lock])
+def render_hocr_page(
+        input_file,
+        output_file,
+        log,
+        pdfinfo,
+        pdfinfo_lock):
+    hocr = input_file
+    pageinfo = get_pageinfo(hocr, pdfinfo, pdfinfo_lock)
     dpi = round(max(pageinfo['xres'], pageinfo['yres'], options.oversample))
 
     hocrtransform = HocrTransform(hocr, dpi)
-    hocrtransform.to_pdf(output_file, imageFileName=image,
+    hocrtransform.to_pdf(output_file, imageFileName=None,
                          showBoundingboxes=False, invisibleText=True)
 
 
@@ -678,6 +642,35 @@ def render_hocr_debug_page(
                          showBoundingboxes=True, invisibleText=False)
 
 
+@active_if(options.pdf_renderer == 'hocr')
+@collate(
+    input=[render_hocr_page, select_image_layer],
+    filter=regex(r".*/(\d{6})(?:\.hocr\.pdf|\.image-layer\.pdf)"),
+    output=os.path.join(work_folder, r'\1.rendered.pdf'),
+    extras=[_log, _pdfinfo, _pdfinfo_lock])
+def add_text_layer(
+        infiles,
+        output_file,
+        log,
+        pdfinfo,
+        pdfinfo_lock):
+    text = next(ii for ii in infiles if ii.endswith('.hocr.pdf'))
+    image = next(ii for ii in infiles if ii.endswith('.image-layer.pdf'))
+
+    pdf_output = pypdf.PdfFileWriter()
+
+    pdf_text = pypdf.PdfFileReader(open(text, "rb"))
+    pdf_image = pypdf.PdfFileReader(open(image, "rb"))
+
+    page = pdf_text.getPage(0)
+    page.mergePage(pdf_image.getPage(0))
+
+    pdf_output.addPage(page)
+
+    with open(output_file, "wb") as out:
+        pdf_output.write(out)
+
+
 @active_if(options.pdf_renderer == 'tesseract')
 @collate(
     input=[preprocess_clean, split_pages],
@@ -698,33 +691,21 @@ def tesseract_ocr_and_render_pdf(
         re_symlink(input_pdf, output_file)
         return
 
-    args_tesseract = [
-        'tesseract',
-        '-l', '+'.join(options.language),
-        input_image,
-        os.path.splitext(output_file)[0],  # Tesseract appends suffix
-        'pdf'
-    ] + options.tesseract_config
-    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
-              universal_newlines=True)
-
-    try:
-        stdout, stderr = p.communicate(timeout=options.tesseract_timeout)
-        if stdout:
-            log.info(stdout)
-        if stderr:
-            log.error(stderr)
-    except TimeoutExpired:
-        p.kill()
-        log.info("Tesseract - page timed out")
-        re_symlink(input_pdf, output_file)
+    tesseract.generate_pdf(
+        input_image=input_image,
+        skip_pdf=input_pdf,
+        output_pdf=output_file,
+        language=options.language,
+        tessconfig=options.tesseract_config,
+        timeout=options.tesseract_timeout,
+        pagesegmode=options.tesseract_pagesegmode,
+        log=log)
 
 
 @transform(
     input=repair_pdf,
-    filter=suffix('.repaired.pdf'),
-    output='.pdfa_def.ps',
-    output_dir=work_folder,
+    filter=formatter(r'\.repaired\.pdf'),
+    output=os.path.join(work_folder, 'pdfa_def.ps'),
     extras=[_log])
 def generate_postscript_stub(
         input_file,
@@ -758,6 +739,11 @@ def generate_postscript_stub(
     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())
+
     generate_pdfa_def(output_file, pdfmark)
 
 
@@ -775,7 +761,7 @@ def skip_page(
 
 
 @merge(
-    input=[render_hocr_page, render_hocr_debug_page, skip_page,
+    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])
@@ -821,46 +807,7 @@ def copy_final(
 def validate_pdfa(
         input_file,
         log):
-
-    args_qpdf = [
-        'qpdf',
-        '--check',
-        input_file
-    ]
-
-    try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 2:
-            print("{0}: not a valid PDF, and could not repair it.".format(
-                    options.input_file))
-            print("Details:")
-            print(e.output)
-        elif e.returncode == 3:
-            log.info("qpdf --check returned warnings:")
-            log.info(e.output)
-        else:
-            print(e.output)
-        return False
-
-    return True
-
-
-# @active_if(ocr_required and options.exact_image)
-# @merge([render_hocr_blank_page, extract_single_page],
-#        os.path.join(work_folder, "%04i.merged.pdf") % pageno)
-# def merge_hocr_with_original_page(infiles, output_file):
-#     with open(infiles[0], 'rb') as hocr_input, \
-#             open(infiles[1], 'rb') as page_input, \
-#             open(output_file, 'wb') as output:
-#         hocr_reader = pypdf.PdfFileReader(hocr_input)
-#         page_reader = pypdf.PdfFileReader(page_input)
-#         writer = pypdf.PdfFileWriter()
-
-#         the_page = hocr_reader.getPage(0)
-#         the_page.mergePage(page_reader.getPage(0))
-#         writer.addPage(the_page)
-#         writer.write(output)
+    return qpdf.check(input_file, log)
 
 
 def available_cpu_count():
@@ -881,11 +828,18 @@ def available_cpu_count():
     return 1
 
 
-def run_pipeline():
-    if not options.jobs or options.jobs == 1:
-        options.jobs = available_cpu_count()
+def cleanup_ruffus_error_message(msg):
+    msg = re.sub(r'\s+', r' ', msg, re.MULTILINE)
+    msg = re.sub(r"\((.+?)\)", r'\1', msg)
+    msg = msg.strip()
+    return msg
 
+
+def run_pipeline():
+    if not options.jobs:
+        options.jobs = available_cpu_count()
     try:
+        options.history_file = os.path.join(work_folder, 'ruffus_history.sqlite')
         cmdline.run(options)
     except ruffus_exceptions.RethrownJobError as e:
         if options.verbose:
@@ -896,9 +850,21 @@ def run_pipeline():
         for exc in e.args:
             task_name, job_name, exc_name, exc_value, exc_stack = exc
             if exc_name == 'builtins.SystemExit':
-                return eval(
-                    exc_value,
-                    {'ExitCode': ExitCode}, {'exc_value': exc_value})
+                match = re.search(r"\.(.+?)\)", exc_value)
+                exit_code_name = match.groups()[0]
+                exit_code = getattr(ExitCode, exit_code_name, 'other_error')
+                return exit_code
+            elif exc_name == 'ruffus.ruffus_exceptions.MissingInputFileError':
+                print(cleanup_ruffus_error_message(exc_value))
+                return 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':
+                    print("Input file '{0}' is not a valid PDF".format(
+                        options.input_file))
+                    return ExitCode.input_file
+
         return ExitCode.other_error
 
     if not validate_pdfa(options.output_file, _log):

ocrmypdf/pageinfo.py

@@ -104,16 +104,37 @@ def _find_page_images(page, pageinfo):
         yield image
 
 
-def _pdf_get_pageinfo(infile, page: int):
+def _page_has_text(pdf, page):
+    # 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(infile, pageno: int):
     pageinfo = {}
-    pageinfo['pageno'] = page
+    pageinfo['pageno'] = pageno
     pageinfo['images'] = []
 
     pdf = pypdf.PdfFileReader(infile)
-    page = pdf.pages[page - 1]
+    page = pdf.pages[pageno]
 
-    text = page.extractText()
-    pageinfo['has_text'] = (text.strip() != '')
+    pageinfo['has_text'] = _page_has_text(pdf, page)
 
     width_pt = page['/MediaBox'][2] - page['/MediaBox'][0]
     height_pt = page['/MediaBox'][3] - page['/MediaBox'][1]

ocrmypdf/pdfa.py

@@ -8,6 +8,7 @@ from string import Template
 from subprocess import Popen, PIPE
 import os
 import codecs
+from . import get_program
 
 
 # This is a template written in PostScript which is needed to create PDF/A
@@ -27,6 +28,7 @@ def
   /Author <$author>
   /Subject <$subject>
   /Keywords <$keywords>
+  /Creator <$creator>
   /DOCINFO pdfmark
 
 % Define an ICC profile :
@@ -86,6 +88,7 @@ def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
                           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
 
@@ -93,7 +96,8 @@ def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
 def _get_postscript_icc_path():
     "Parse Ghostscript's help message to find where iccprofiles are stored"
 
-    p_gs = Popen(['gs', '--help'], close_fds=True, universal_newlines=True,
+    p_gs = Popen([get_program('gs'), '--help'], close_fds=True,
+                 universal_newlines=True,
                  stdout=PIPE, stderr=PIPE)
     out, _ = p_gs.communicate()
     lines = out.splitlines()

ocrmypdf/qpdf.py

@@ -0,0 +1,84 @@
+#!/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:
+            print("{0}: not a valid PDF, and could not repair it.".format(
+                    input_file))
+            print("Details:")
+            print(e.output)
+        elif e.returncode == 3:
+            log.info("qpdf --check returned warnings:")
+            log.info(e.output)
+        else:
+            print(e.output)
+        return False
+    return True
+
+
+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:')
+            log.debug(e.output)
+            print(e.output)
+            return
+
+        if e.returncode == 2 and e.output.find("invalid password"):
+            print("{0}: this PDF is password-protected - password must "
+                  "be removed for OCR".format(input_file))
+            sys.exit(ExitCode.input_file)
+        elif e.returncode == 2:
+            print("{0}: not a valid PDF, and could not repair it.".format(
+                    input_file))
+            print("Details:")
+            print(e.output)
+            sys.exit(ExitCode.input_file)
+        else:
+            print("{0}: unknown error".format(
+                    input_file))
+            print(e.output)
+            sys.exit(ExitCode.unknown)
+
+
+def get_npages(input_file):
+    pages = check_output(
+        [get_program('qpdf'), '--show-npages', input_file],
+         universal_newlines=True, close_fds=True)
+    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

@@ -1,49 +1,19 @@
 #!/usr/bin/env python3
 # © 2015 James R. Barlow: github.com/jbarlow83
 
-from subprocess import STDOUT, CalledProcessError, check_output
 import sys
 import os
 import re
+import shutil
 from functools import lru_cache
-from . import ExitCode
+from . import ExitCode, get_program
 
-
-@lru_cache(maxsize=1)
-def version():
-    args_tess = [
-        'tesseract',
-        '--version'
-    ]
-    try:
-        versions = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
-    except CalledProcessError:
-        print("Could not find Tesseract executable on system PATH.")
-        sys.exit(ExitCode.missing_dependency)
-
-    tesseract_version = re.match(r'tesseract\s(.+)', versions).group(1)
-    return tesseract_version
-
-
-@lru_cache(maxsize=1)
-def languages():
-    args_tess = [
-        'tesseract',
-        '--list-langs'
-    ]
-    try:
-        langs = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
-    except CalledProcessError as e:
-        print("Tesseract failed to report available languages.")
-        print("Output from Tesseract:")
-        print("-" * 40)
-        print(e.output)
-        sys.exit(ExitCode.missing_dependency)
-    return set(lang.strip() for lang in langs.splitlines()[1:])
+from subprocess import Popen, PIPE, CalledProcessError, \
+    TimeoutExpired, check_output, STDOUT
+try:
+    from subprocess import DEVNULL
+except ImportError:
+    DEVNULL = open(os.devnull, 'wb')
 
 
 HOCR_TEMPLATE = '''<?xml version="1.0" encoding="UTF-8"?>
@@ -67,3 +37,145 @@ HOCR_TEMPLATE = '''<?xml version="1.0" encoding="UTF-8"?>
   </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 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)
+    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
+              universal_newlines=True)
+    try:
+        stdout, stderr = p.communicate(timeout=timeout)
+    except TimeoutExpired:
+        p.kill()
+        stdout, stderr = p.communicate()
+        # Generate a HOCR file with no recognized text if tesseract times out
+        # Temporary workaround to hocrTransform not being able to function if
+        # it does not have a valid hOCR file.
+        with open(output_hocr, 'w', encoding="utf-8") as f:
+            pageinfo = pageinfo_getter()
+            f.write(HOCR_TEMPLATE.format(
+                pageinfo['width_pixels'],
+                pageinfo['height_pixels']))
+    else:
+        if stdout:
+            log.info(stdout)
+        if stderr:
+            log.error(stderr)
+
+        if p.returncode != 0:
+            raise CalledProcessError(p.returncode, args_tesseract)
+
+        if os.path.exists(badxml + '.html'):
+            # Tesseract 3.02 appends suffix ".html" on its own (.badxml.html)
+            shutil.move(badxml + '.html', badxml)
+        elif os.path.exists(badxml + '.hocr'):
+            # Tesseract 3.03 appends suffix ".hocr" on its own (.badxml.hocr)
+            shutil.move(badxml + '.hocr', badxml)
+
+        # Tesseract 3.03 inserts source filename into hocr file without
+        # escaping it, creating invalid XML and breaking the parser.
+        # As a workaround, rewrite the hocr file, replacing the filename
+        # with a space.  Don't know if Tesseract 3.02 does the same.
+
+        regex_nested_single_quotes = re.compile(
+            r"""title='image "([^"]*)";""")
+        with open(badxml, mode='r', encoding='utf-8') as f_in, \
+                open(output_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)
+    p = Popen(args_tesseract, close_fds=True, stdout=PIPE, stderr=PIPE,
+              universal_newlines=True)
+
+    try:
+        stdout, stderr = p.communicate(timeout=timeout)
+        if stdout:
+            log.info(stdout)
+        if stderr:
+            log.error(stderr)
+    except TimeoutExpired:
+        p.kill()
+        log.info("Tesseract - page timed out")
+        shutil.copy(skip_pdf, output_pdf)
+

ocrmypdf/unpaper.py

@@ -8,13 +8,13 @@ from tempfile import NamedTemporaryFile
 import sys
 import os
 from functools import lru_cache
-from . import ExitCode
+from . import ExitCode, get_program
 
 
 @lru_cache(maxsize=1)
 def version():
     args_unpaper = [
-        'unpaper',
+        get_program('unpaper'),
         '--version'
     ]
     p_unpaper = Popen(args_unpaper, close_fds=True, universal_newlines=True,
@@ -33,7 +33,7 @@ except ImportError:
 
 def run(input_file, output_file, dpi, log, mode_args):
     args_unpaper = [
-        'unpaper',
+        get_program('unpaper'),
         '-v',
         '--dpi', str(dpi)
     ] + mode_args

pipeline.svg

@@ -4,232 +4,263 @@
 <!-- Generated by graphviz version 2.38.0 (20140413.2041)
  -->
 <!-- Title: Pipeline: Pages: 1 -->
-<svg width="728pt" height="651pt"
- viewBox="0.00 0.00 728.00 650.53" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
-<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 646.53)">
+<svg width="1132pt" height="708pt"
+ viewBox="0.00 0.00 1132.00 708.08" 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(4 704.083)">
 <title>Pipeline:</title>
-<polygon fill="white" stroke="none" points="-4,4 -4,-646.53 724,-646.53 724,4 -4,4"/>
+<polygon fill="white" stroke="none" points="-4,4 -4,-704.083 1128,-704.083 1128,4 -4,4"/>
 <g id="clust1" class="cluster"><title>clustertasks</title>
-<polygon fill="none" stroke="black" points="8,-8 8,-634.53 712,-634.53 712,-8 8,-8"/>
-<text text-anchor="middle" x="360" y="-606.53" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
+<polygon fill="none" stroke="black" points="8,-8 8,-692.083 1116,-692.083 1116,-8 8,-8"/>
+<text text-anchor="middle" x="562" y="-664.083" 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="#006000" points="481.791,-588.53 386.209,-588.53 382.209,-584.53 382.209,-552.53 477.791,-552.53 481.791,-556.53 481.791,-588.53"/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 382.209,-584.53 "/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 477.791,-552.53 "/>
-<polyline fill="none" stroke="#006000" points="477.791,-584.53 481.791,-588.53 "/>
-<text text-anchor="middle" x="432" y="-564.53" font-family="Times,serif" font-size="20.00" fill="#006000">repair_pdf</text>
+<polygon fill="#efa03b" stroke="black" points="936.535,-646.083 713.465,-646.083 709.465,-642.083 709.465,-610.083 932.535,-610.083 936.535,-614.083 936.535,-646.083"/>
+<polyline fill="none" stroke="black" points="932.535,-642.083 709.465,-642.083 "/>
+<polyline fill="none" stroke="black" points="932.535,-642.083 932.535,-610.083 "/>
+<polyline fill="none" stroke="black" points="932.535,-642.083 936.535,-646.083 "/>
+<text text-anchor="middle" x="823" y="-622.083" 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="466.782,-509.564 374,-526.497 281.218,-509.564 281.304,-482.165 466.696,-482.165 466.782,-509.564"/>
-<polygon fill="none" stroke="black" points="470.799,-512.902 374,-530.569 277.201,-512.902 277.311,-478.159 470.689,-478.159 470.799,-512.902"/>
-<text text-anchor="middle" x="374" y="-495.991" font-family="Times,serif" font-size="20.00">split_pages</text>
+<polygon fill="#efa03b" stroke="black" points="914.112,-567.155 710,-584.057 505.888,-567.155 506.078,-539.806 913.922,-539.806 914.112,-567.155"/>
+<polygon fill="none" stroke="black" points="918.134,-570.834 710,-588.069 501.866,-570.834 502.11,-535.808 917.89,-535.808 918.134,-570.834"/>
+<text text-anchor="middle" x="710" y="-553.596" 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="gray" d="M417.064,-552.394C412.296,-546.925 406.86,-540.689 401.493,-534.532"/>
-<polygon fill="gray" stroke="gray" points="403.996,-532.077 394.787,-526.838 398.719,-536.676 403.996,-532.077"/>
+<path fill="none" stroke="#0044a0" d="M793.9,-609.961C783.582,-603.89 771.656,-596.873 760.092,-590.069"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="761.747,-586.982 751.353,-584.927 758.197,-593.015 761.747,-586.982"/>
 </g>
-<!-- t10 -->
-<g id="node12" class="node"><title>t10</title>
-<polygon fill="#efa03b" stroke="#006000" points="704.338,-451.452 493.662,-451.452 489.662,-447.452 489.662,-415.452 700.338,-415.452 704.338,-419.452 704.338,-451.452"/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 489.662,-447.452 "/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 700.338,-415.452 "/>
-<polyline fill="none" stroke="#006000" points="700.338,-447.452 704.338,-451.452 "/>
-<text text-anchor="middle" x="597" y="-427.452" font-family="Times,serif" font-size="20.00" fill="#006000">generate_postscript_stub</text>
+<!-- t12 -->
+<g id="node14" class="node"><title>t12</title>
+<polygon fill="#efa03b" stroke="black" points="1108.08,-509.109 769.918,-509.109 765.918,-505.109 765.918,-473.109 1104.08,-473.109 1108.08,-477.109 1108.08,-509.109"/>
+<polyline fill="none" stroke="black" points="1104.08,-505.109 765.918,-505.109 "/>
+<polyline fill="none" stroke="black" points="1104.08,-505.109 1104.08,-473.109 "/>
+<polyline fill="none" stroke="black" points="1104.08,-505.109 1108.08,-509.109 "/>
+<text text-anchor="middle" x="937" y="-485.109" font-family="Times,serif" font-size="20.00">generate_postscript_stub</text>
 </g>
-<!-- t0&#45;&gt;t10 -->
-<g id="edge16" class="edge"><title>t0&#45;&gt;t10</title>
-<path fill="none" stroke="gray" d="M453.048,-552.468C461.441,-545.654 471.185,-537.73 480,-530.53 510.227,-505.84 544.753,-477.467 568.421,-457.99"/>
-<polygon fill="gray" stroke="gray" points="570.818,-460.55 576.315,-451.493 566.369,-455.146 570.818,-460.55"/>
+<!-- t0&#45;&gt;t12 -->
+<g id="edge19" class="edge"><title>t0&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M899.32,-610.004C909.979,-604.592 919.748,-597.466 927,-588.083 941.916,-568.78 943.099,-540.385 941.345,-519.486"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="944.8,-518.881 940.218,-509.328 937.843,-519.653 944.8,-518.881"/>
 </g>
 <!-- t2 -->
 <g id="node3" class="node"><title>t2</title>
-<polygon fill="#efa03b" stroke="black" points="451.555,-451.452 228.445,-451.452 224.445,-447.452 224.445,-415.452 447.555,-415.452 451.555,-419.452 451.555,-451.452"/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 224.445,-447.452 "/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 447.555,-415.452 "/>
-<polyline fill="none" stroke="black" points="447.555,-447.452 451.555,-451.452 "/>
-<text text-anchor="middle" x="338" y="-427.452" font-family="Times,serif" font-size="20.00">rasterize_with_ghostscript</text>
+<polygon fill="#efa03b" stroke="black" points="592.299,-509.109 241.701,-509.109 237.701,-505.109 237.701,-473.109 588.299,-473.109 592.299,-477.109 592.299,-509.109"/>
+<polyline fill="none" stroke="black" points="588.299,-505.109 237.701,-505.109 "/>
+<polyline fill="none" stroke="black" points="588.299,-505.109 588.299,-473.109 "/>
+<polyline fill="none" stroke="black" points="588.299,-505.109 592.299,-509.109 "/>
+<text text-anchor="middle" x="415" y="-485.109" font-family="Times,serif" font-size="20.00">rasterize_with_ghostscript</text>
 </g>
 <!-- t1&#45;&gt;t2 -->
 <g id="edge2" class="edge"><title>t1&#45;&gt;t2</title>
-<path fill="none" stroke="#0044a0" d="M361.611,-478.092C358.514,-472.369 355.171,-466.19 352.003,-460.333"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="355.064,-458.636 347.227,-451.506 348.907,-461.967 355.064,-458.636"/>
+<path fill="none" stroke="#0044a0" d="M608.89,-535.808C573.672,-527.87 534.524,-519.048 500.679,-511.42"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="501.345,-507.982 490.82,-509.198 499.806,-514.811 501.345,-507.982"/>
+</g>
+<!-- t7 -->
+<g id="node9" class="node"><title>t7</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>
+<!-- t1&#45;&gt;t7 -->
+<g id="edge11" class="edge"><title>t1&#45;&gt;t7</title>
+<path fill="none" stroke="#0044a0" d="M501.985,-548.028C416.651,-540.897 317.351,-528.985 229,-509.109 131.492,-487.174 17,-534.054 17,-434.109 17,-434.109 17,-434.109 17,-374.109 17,-340.481 4.97908,-324.525 27,-299.109 32.8004,-292.415 39.6539,-286.828 47.1559,-282.17"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="49.1993,-285.038 56.2361,-277.118 45.7959,-278.921 49.1993,-285.038"/>
+</g>
+<!-- t13 -->
+<g id="node12" class="node"><title>t13</title>
+<polygon fill="#efa03b" stroke="black" points="1029.34,-451.109 808.662,-451.109 804.662,-447.109 804.662,-415.109 1025.34,-415.109 1029.34,-419.109 1029.34,-451.109"/>
+<polyline fill="none" stroke="black" points="1025.34,-447.109 804.662,-447.109 "/>
+<polyline fill="none" stroke="black" points="1025.34,-447.109 1025.34,-415.109 "/>
+<polyline fill="none" stroke="black" points="1025.34,-447.109 1029.34,-451.109 "/>
+<text text-anchor="middle" x="917" y="-427.109" font-family="Times,serif" font-size="20.00">skip_page</text>
+</g>
+<!-- t1&#45;&gt;t13 -->
+<g id="edge16" class="edge"><title>t1&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M716.849,-535.484C723.825,-515.965 736.562,-488.763 757,-473.109 768.342,-464.422 781.352,-457.641 794.947,-452.352"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="796.223,-455.612 804.44,-448.923 793.845,-449.029 796.223,-455.612"/>
 </g>
 <!-- t11 -->
-<g id="node10" class="node"><title>t11</title>
-<polygon fill="#efa03b" stroke="black" points="644.594,-393.452 551.406,-393.452 547.406,-389.452 547.406,-357.452 640.594,-357.452 644.594,-361.452 644.594,-393.452"/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 547.406,-389.452 "/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 640.594,-357.452 "/>
-<polyline fill="none" stroke="black" points="640.594,-389.452 644.594,-393.452 "/>
-<text text-anchor="middle" x="596" y="-369.452" font-family="Times,serif" font-size="20.00">skip_page</text>
+<g id="node13" class="node"><title>t11</title>
+<polygon fill="#efa03b" stroke="black" points="1068.24,-335.109 687.76,-335.109 683.76,-331.109 683.76,-299.109 1064.24,-299.109 1068.24,-303.109 1068.24,-335.109"/>
+<polyline fill="none" stroke="black" points="1064.24,-331.109 683.76,-331.109 "/>
+<polyline fill="none" stroke="black" points="1064.24,-331.109 1064.24,-299.109 "/>
+<polyline fill="none" stroke="black" points="1064.24,-331.109 1068.24,-335.109 "/>
+<text text-anchor="middle" x="876" y="-311.109" font-family="Times,serif" font-size="20.00">tesseract_ocr_and_render_pdf</text>
 </g>
 <!-- t1&#45;&gt;t11 -->
-<g id="edge13" class="edge"><title>t1&#45;&gt;t11</title>
-<path fill="none" stroke="#0044a0" d="M425.139,-478.007C437.834,-470.731 450.731,-461.83 461,-451.452 473.874,-438.442 466.946,-427.178 481,-415.452 490.212,-407.766 513.965,-399.205 537.508,-392.067"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="538.592,-395.396 547.189,-389.203 536.607,-388.683 538.592,-395.396"/>
-</g>
-<!-- t9 -->
-<g id="node11" class="node"><title>t9</title>
-<polygon fill="#efa03b" stroke="black" points="272.496,-277.452 19.5039,-277.452 15.5039,-273.452 15.5039,-241.452 268.496,-241.452 272.496,-245.452 272.496,-277.452"/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 15.5039,-273.452 "/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 268.496,-241.452 "/>
-<polyline fill="none" stroke="black" points="268.496,-273.452 272.496,-277.452 "/>
-<text text-anchor="middle" x="144" y="-253.452" font-family="Times,serif" font-size="20.00">tesseract_ocr_and_render_pdf</text>
-</g>
-<!-- t1&#45;&gt;t9 -->
-<g id="edge15" class="edge"><title>t1&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M277.341,-486.044C254.708,-478.702 232.21,-467.759 215,-451.452 168.207,-407.114 152.056,-329.127 146.633,-287.855"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="150.084,-287.235 145.422,-277.721 143.133,-288.065 150.084,-287.235"/>
+<g id="edge18" class="edge"><title>t1&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M710.717,-535.657C712.098,-517.744 715.879,-492.693 726,-473.109 754.651,-417.673 809.56,-368.731 844.403,-341.318"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="846.607,-344.038 852.372,-335.148 842.322,-338.503 846.607,-344.038"/>
 </g>
 <!-- t3 -->
 <g id="node4" class="node"><title>t3</title>
-<polygon fill="#efa03b" stroke="black" points="458.999,-393.452 291.001,-393.452 287.001,-389.452 287.001,-357.452 454.999,-357.452 458.999,-361.452 458.999,-393.452"/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 287.001,-389.452 "/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 454.999,-357.452 "/>
-<polyline fill="none" stroke="black" points="454.999,-389.452 458.999,-393.452 "/>
-<text text-anchor="middle" x="373" y="-369.452" font-family="Times,serif" font-size="20.00">preprocess_deskew</text>
+<polygon fill="#efa03b" stroke="black" points="564.742,-451.109 269.258,-451.109 265.258,-447.109 265.258,-415.109 560.742,-415.109 564.742,-419.109 564.742,-451.109"/>
+<polyline fill="none" stroke="black" points="560.742,-447.109 265.258,-447.109 "/>
+<polyline fill="none" stroke="black" points="560.742,-447.109 560.742,-415.109 "/>
+<polyline fill="none" stroke="black" points="560.742,-447.109 564.742,-451.109 "/>
+<text text-anchor="middle" x="415" y="-427.109" font-family="Times,serif" font-size="20.00">preprocess_deskew</text>
 </g>
 <!-- t2&#45;&gt;t3 -->
 <g id="edge3" class="edge"><title>t2&#45;&gt;t3</title>
-<path fill="none" stroke="#0044a0" d="M348.691,-415.346C351.337,-411.112 354.23,-406.485 357.066,-401.946"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="360.042,-403.788 362.374,-393.453 354.106,-400.078 360.042,-403.788"/>
+<path fill="none" stroke="#0044a0" d="M415,-473.003C415,-469.312 415,-465.322 415,-461.352"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="418.5,-461.111 415,-451.111 411.5,-461.111 418.5,-461.111"/>
 </g>
 <!-- t6 -->
-<g id="node6" class="node"><title>t6</title>
-<polygon fill="#efa03b" stroke="black" points="664.375,-277.452 477.625,-277.452 473.625,-273.452 473.625,-241.452 660.375,-241.452 664.375,-245.452 664.375,-277.452"/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 473.625,-273.452 "/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 660.375,-241.452 "/>
-<polyline fill="none" stroke="black" points="660.375,-273.452 664.375,-277.452 "/>
-<text text-anchor="middle" x="569" y="-253.452" font-family="Times,serif" font-size="20.00">select_image_for_pdf</text>
+<g id="node8" class="node"><title>t6</title>
+<polygon fill="#efa03b" stroke="black" points="354.119,-335.109 39.8808,-335.109 35.8808,-331.109 35.8808,-299.109 350.119,-299.109 354.119,-303.109 354.119,-335.109"/>
+<polyline fill="none" stroke="black" points="350.119,-331.109 35.8808,-331.109 "/>
+<polyline fill="none" stroke="black" points="350.119,-331.109 350.119,-299.109 "/>
+<polyline fill="none" stroke="black" points="350.119,-331.109 354.119,-335.109 "/>
+<text text-anchor="middle" x="195" y="-311.109" font-family="Times,serif" font-size="20.00">select_image_for_pdf</text>
 </g>
 <!-- t2&#45;&gt;t6 -->
-<g id="edge7" class="edge"><title>t2&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M423.032,-415.374C438.812,-409.925 454.543,-402.785 468,-393.452 508.201,-365.571 539.287,-316.662 555.808,-286.558"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="559.024,-287.965 560.653,-277.496 552.851,-284.664 559.024,-287.965"/>
+<g id="edge9" class="edge"><title>t2&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M294.315,-473.06C280.461,-467.585 267.283,-460.434 256,-451.109 223.245,-424.039 207.251,-375.559 200.099,-345.207"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="203.486,-344.31 197.925,-335.292 196.648,-345.809 203.486,-344.31"/>
 </g>
 <!-- t4 -->
 <g id="node5" class="node"><title>t4</title>
-<polygon fill="#efa03b" stroke="black" points="449.705,-335.452 300.295,-335.452 296.295,-331.452 296.295,-299.452 445.705,-299.452 449.705,-303.452 449.705,-335.452"/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 296.295,-331.452 "/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 445.705,-299.452 "/>
-<polyline fill="none" stroke="black" points="445.705,-331.452 449.705,-335.452 "/>
-<text text-anchor="middle" x="373" y="-311.452" font-family="Times,serif" font-size="20.00">preprocess_clean</text>
+<polygon fill="#efa03b" stroke="black" points="587.95,-393.109 310.05,-393.109 306.05,-389.109 306.05,-357.109 583.95,-357.109 587.95,-361.109 587.95,-393.109"/>
+<polyline fill="none" stroke="black" points="583.95,-389.109 306.05,-389.109 "/>
+<polyline fill="none" stroke="black" points="583.95,-389.109 583.95,-357.109 "/>
+<polyline fill="none" stroke="black" points="583.95,-389.109 587.95,-393.109 "/>
+<text text-anchor="middle" x="447" y="-369.109" font-family="Times,serif" font-size="20.00">preprocess_clean</text>
 </g>
 <!-- t3&#45;&gt;t4 -->
 <g id="edge4" class="edge"><title>t3&#45;&gt;t4</title>
-<path fill="none" stroke="#0044a0" d="M373,-357.346C373,-353.655 373,-349.665 373,-345.695"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="376.5,-345.453 373,-335.453 369.5,-345.453 376.5,-345.453"/>
+<path fill="none" stroke="#0044a0" d="M424.775,-415.003C427.132,-410.878 429.703,-406.379 432.232,-401.952"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="435.362,-403.53 437.285,-393.111 429.285,-400.057 435.362,-403.53"/>
 </g>
 <!-- t3&#45;&gt;t6 -->
-<g id="edge6" class="edge"><title>t3&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M414.399,-357.412C428.785,-351.039 444.852,-343.405 459,-335.452 466.909,-331.007 506.49,-303.805 535.921,-283.431"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="538.073,-286.198 544.299,-277.625 534.086,-280.444 538.073,-286.198"/>
-</g>
-<!-- t4&#45;&gt;t6 -->
-<g id="edge5" class="edge"><title>t4&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M432.605,-299.422C453.701,-293.395 477.617,-286.561 499.465,-280.319"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="500.687,-283.61 509.341,-277.497 498.764,-276.879 500.687,-283.61"/>
+<g id="edge8" class="edge"><title>t3&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M351.041,-415.02C333.086,-409.151 313.871,-401.82 297,-393.109 269.828,-379.08 242.065,-358.149 222.377,-341.96"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="224.378,-339.071 214.46,-335.347 219.891,-344.444 224.378,-339.071"/>
 </g>
 <!-- t5 -->
-<g id="node7" class="node"><title>t5</title>
-<polygon fill="#efa03b" stroke="black" points="455.922,-277.452 294.078,-277.452 290.078,-273.452 290.078,-241.452 451.922,-241.452 455.922,-245.452 455.922,-277.452"/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 290.078,-273.452 "/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 451.922,-241.452 "/>
-<polyline fill="none" stroke="black" points="451.922,-273.452 455.922,-277.452 "/>
-<text text-anchor="middle" x="373" y="-253.452" font-family="Times,serif" font-size="20.00">ocr_tesseract_hocr</text>
+<g id="node6" class="node"><title>t5</title>
+<polygon fill="#efa03b" stroke="black" points="665.666,-335.109 376.334,-335.109 372.334,-331.109 372.334,-299.109 661.666,-299.109 665.666,-303.109 665.666,-335.109"/>
+<polyline fill="none" stroke="black" points="661.666,-331.109 372.334,-331.109 "/>
+<polyline fill="none" stroke="black" points="661.666,-331.109 661.666,-299.109 "/>
+<polyline fill="none" stroke="black" points="661.666,-331.109 665.666,-335.109 "/>
+<text text-anchor="middle" x="519" y="-311.109" font-family="Times,serif" font-size="20.00">ocr_tesseract_hocr</text>
 </g>
 <!-- t4&#45;&gt;t5 -->
-<g id="edge8" class="edge"><title>t4&#45;&gt;t5</title>
-<path fill="none" stroke="#0044a0" d="M373,-299.346C373,-295.655 373,-291.665 373,-287.695"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="376.5,-287.453 373,-277.453 369.5,-287.453 376.5,-287.453"/>
+<g id="edge5" class="edge"><title>t4&#45;&gt;t5</title>
+<path fill="none" stroke="#0044a0" d="M468.994,-357.003C475.274,-352.118 482.229,-346.709 488.905,-341.516"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="491.396,-344.013 497.141,-335.111 487.099,-338.487 491.396,-344.013"/>
 </g>
-<!-- t4&#45;&gt;t9 -->
-<g id="edge14" class="edge"><title>t4&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M303.36,-299.422C278.157,-293.259 249.509,-286.253 223.521,-279.898"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="224.249,-276.473 213.704,-277.497 222.586,-283.273 224.249,-276.473"/>
+<!-- t4&#45;&gt;t6 -->
+<g id="edge7" class="edge"><title>t4&#45;&gt;t6</title>
+<path fill="none" stroke="#0044a0" d="M370.365,-357.079C342.386,-350.862 310.55,-343.787 281.752,-337.388"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="282.225,-333.907 271.704,-335.155 280.707,-340.741 282.225,-333.907"/>
 </g>
-<!-- t7 -->
-<g id="node8" class="node"><title>t7</title>
-<polygon fill="#efa03b" stroke="black" points="426.366,-219.452 269.634,-219.452 265.634,-215.452 265.634,-183.452 422.366,-183.452 426.366,-187.452 426.366,-219.452"/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 265.634,-215.452 "/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 422.366,-183.452 "/>
-<polyline fill="none" stroke="black" points="422.366,-215.452 426.366,-219.452 "/>
-<text text-anchor="middle" x="346" y="-195.452" font-family="Times,serif" font-size="20.00">render_hocr_page</text>
-</g>
-<!-- t6&#45;&gt;t7 -->
-<g id="edge9" class="edge"><title>t6&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M501.184,-241.422C476.75,-235.286 448.99,-228.315 423.772,-221.982"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="424.429,-218.538 413.877,-219.497 422.724,-225.328 424.429,-218.538"/>
+<!-- t4&#45;&gt;t11 -->
+<g id="edge17" class="edge"><title>t4&#45;&gt;t11</title>
+<path fill="none" stroke="#0044a0" d="M577.461,-357.079C627.38,-350.563 684.512,-343.106 735.342,-336.47"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="735.957,-339.92 745.42,-335.155 735.051,-332.979 735.957,-339.92"/>
 </g>
 <!-- t8 -->
-<g id="node9" class="node"><title>t8</title>
-<polygon fill="#efa03b" stroke="black" points="663.742,-219.452 448.258,-219.452 444.258,-215.452 444.258,-183.452 659.742,-183.452 663.742,-187.452 663.742,-219.452"/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 444.258,-215.452 "/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 659.742,-183.452 "/>
-<polyline fill="none" stroke="black" points="659.742,-215.452 663.742,-219.452 "/>
-<text text-anchor="middle" x="554" y="-195.452" font-family="Times,serif" font-size="20.00">render_hocr_debug_page</text>
-</g>
-<!-- t6&#45;&gt;t8 -->
-<g id="edge11" class="edge"><title>t6&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M564.418,-241.346C563.4,-237.546 562.298,-233.43 561.203,-229.345"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="564.522,-228.207 558.554,-219.453 557.761,-230.018 564.522,-228.207"/>
-</g>
-<!-- t5&#45;&gt;t7 -->
-<g id="edge10" class="edge"><title>t5&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M364.752,-241.346C362.816,-237.329 360.708,-232.958 358.629,-228.645"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="361.693,-226.941 354.197,-219.453 355.387,-229.981 361.693,-226.941"/>
+<g id="node7" class="node"><title>t8</title>
+<polygon fill="#efa03b" stroke="black" points="986.109,-277.109 701.891,-277.109 697.891,-273.109 697.891,-241.109 982.109,-241.109 986.109,-245.109 986.109,-277.109"/>
+<polyline fill="none" stroke="black" points="982.109,-273.109 697.891,-273.109 "/>
+<polyline fill="none" stroke="black" points="982.109,-273.109 982.109,-241.109 "/>
+<polyline fill="none" stroke="black" points="982.109,-273.109 986.109,-277.109 "/>
+<text text-anchor="middle" x="842" y="-253.109" font-family="Times,serif" font-size="20.00">render_hocr_page</text>
 </g>
 <!-- t5&#45;&gt;t8 -->
-<g id="edge12" class="edge"><title>t5&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M428.289,-241.346C447.587,-235.375 469.416,-228.622 489.404,-222.437"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="490.53,-225.753 499.049,-219.453 488.461,-219.065 490.53,-225.753"/>
+<g id="edge6" class="edge"><title>t5&#45;&gt;t8</title>
+<path fill="none" stroke="#0044a0" d="M617.226,-299.079C654.028,-292.699 696.036,-285.416 733.699,-278.886"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="734.429,-282.312 743.685,-277.155 733.234,-275.415 734.429,-282.312"/>
 </g>
-<!-- t12 -->
-<g id="node13" class="node"><title>t12</title>
-<polygon fill="#efa03b" stroke="black" points="473.845,-105.456 554,-78.0208 634.155,-105.456 634.08,-149.848 473.92,-149.848 473.845,-105.456"/>
-<polygon fill="none" stroke="black" points="469.836,-102.581 554,-73.7729 638.164,-102.581 638.078,-153.869 469.922,-153.869 469.836,-102.581"/>
-<text text-anchor="middle" x="554" y="-111.726" font-family="Times,serif" font-size="20.00">merge_pages</text>
+<!-- t9 -->
+<g id="node11" class="node"><title>t9</title>
+<polygon fill="#efa03b" stroke="black" points="679.486,-277.109 336.514,-277.109 332.514,-273.109 332.514,-241.109 675.486,-241.109 679.486,-245.109 679.486,-277.109"/>
+<polyline fill="none" stroke="black" points="675.486,-273.109 332.514,-273.109 "/>
+<polyline fill="none" stroke="black" points="675.486,-273.109 675.486,-241.109 "/>
+<polyline fill="none" stroke="black" points="675.486,-273.109 679.486,-277.109 "/>
+<text text-anchor="middle" x="506" y="-253.109" font-family="Times,serif" font-size="20.00">render_hocr_debug_page</text>
 </g>
-<!-- t7&#45;&gt;t12 -->
-<g id="edge20" class="edge"><title>t7&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M389.35,-183.419C409.937,-175.33 435.398,-165.326 460.092,-155.624"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="461.479,-158.839 469.507,-151.924 458.919,-152.324 461.479,-158.839"/>
+<!-- t5&#45;&gt;t9 -->
+<g id="edge15" class="edge"><title>t5&#45;&gt;t9</title>
+<path fill="none" stroke="#0044a0" d="M515.029,-299.003C514.147,-295.204 513.191,-291.087 512.243,-287.002"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="515.617,-286.06 509.947,-277.111 508.799,-287.643 515.617,-286.06"/>
 </g>
-<!-- t8&#45;&gt;t12 -->
-<g id="edge19" class="edge"><title>t8&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M554,-183.12C554,-177.585 554,-171.177 554,-164.592"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="557.5,-164.201 554,-154.201 550.5,-164.201 557.5,-164.201"/>
+<!-- t10 -->
+<g id="node10" class="node"><title>t10</title>
+<polygon fill="#efa03b" stroke="black" points="973.082,-219.109 714.918,-219.109 710.918,-215.109 710.918,-183.109 969.082,-183.109 973.082,-187.109 973.082,-219.109"/>
+<polyline fill="none" stroke="black" points="969.082,-215.109 710.918,-215.109 "/>
+<polyline fill="none" stroke="black" points="969.082,-215.109 969.082,-183.109 "/>
+<polyline fill="none" stroke="black" points="969.082,-215.109 973.082,-219.109 "/>
+<text text-anchor="middle" x="842" y="-195.109" font-family="Times,serif" font-size="20.00">add_text_layer</text>
 </g>
-<!-- t11&#45;&gt;t12 -->
-<g id="edge17" class="edge"><title>t11&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M615.939,-357.13C634.767,-339.333 661.745,-309.733 673,-277.452 686.754,-238.003 694.347,-219.364 673,-183.452 666.345,-172.256 656.999,-162.875 646.44,-155.047"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="648.378,-152.132 638.152,-149.36 644.417,-157.904 648.378,-152.132"/>
+<!-- t8&#45;&gt;t10 -->
+<g id="edge12" class="edge"><title>t8&#45;&gt;t10</title>
+<path fill="none" stroke="#0044a0" d="M842,-241.003C842,-237.312 842,-233.322 842,-229.352"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="845.5,-229.111 842,-219.111 838.5,-229.111 845.5,-229.111"/>
 </g>
-<!-- t9&#45;&gt;t12 -->
-<g id="edge18" class="edge"><title>t9&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M164.711,-241.235C186.394,-224.08 222.088,-198.224 257,-183.452 321.895,-155.994 399.893,-139.571 459.67,-130.141"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="460.504,-133.554 469.855,-128.574 459.439,-126.635 460.504,-133.554"/>
+<!-- t6&#45;&gt;t7 -->
+<g id="edge10" class="edge"><title>t6&#45;&gt;t7</title>
+<path fill="none" stroke="#0044a0" d="M185.836,-299.003C183.626,-294.878 181.216,-290.379 178.845,-285.952"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="181.915,-284.273 174.108,-277.111 175.745,-287.578 181.915,-284.273"/>
 </g>
-<!-- t10&#45;&gt;t12 -->
-<g id="edge22" class="edge"><title>t10&#45;&gt;t12</title>
-<path fill="none" stroke="gray" d="M629.379,-415.337C660.552,-396.255 703,-362.233 703,-318.452 703,-318.452 703,-318.452 703,-258.452 703,-224.066 705.581,-210.223 684,-183.452 673.976,-171.017 660.952,-160.773 647.026,-152.396"/>
-<polygon fill="gray" stroke="gray" points="648.682,-149.312 638.258,-147.419 645.227,-155.4 648.682,-149.312"/>
+<!-- t6&#45;&gt;t9 -->
+<g id="edge14" class="edge"><title>t6&#45;&gt;t9</title>
+<path fill="none" stroke="#0044a0" d="M289.577,-299.079C324.86,-292.726 365.115,-285.478 401.259,-278.969"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="402.116,-282.372 411.337,-277.155 400.875,-275.482 402.116,-282.372"/>
 </g>
-<!-- t13 -->
-<g id="node14" class="node"><title>t13</title>
-<polygon fill="#efa03b" stroke="black" points="616.338,-52 495.662,-52 491.662,-48 491.662,-16 612.338,-16 616.338,-20 616.338,-52"/>
-<polyline fill="none" stroke="black" points="612.338,-48 491.662,-48 "/>
-<polyline fill="none" stroke="black" points="612.338,-48 612.338,-16 "/>
-<polyline fill="none" stroke="black" points="612.338,-48 616.338,-52 "/>
-<text text-anchor="middle" x="554" y="-28" font-family="Times,serif" font-size="20.00">validate_pdfa</text>
+<!-- t7&#45;&gt;t10 -->
+<g id="edge13" class="edge"><title>t7&#45;&gt;t10</title>
+<path fill="none" stroke="#0044a0" d="M314.366,-242.002C317.606,-241.697 320.821,-241.399 324,-241.109 451.228,-229.521 596.262,-218.816 700.44,-211.565"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="700.976,-215.036 710.71,-210.852 700.491,-208.053 700.976,-215.036"/>
 </g>
-<!-- t12&#45;&gt;t13 -->
-<g id="edge21" class="edge"><title>t12&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M554,-73.9482C554,-69.9654 554,-66.007 554,-62.2247"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="557.5,-62.1573 554,-52.1573 550.5,-62.1574 557.5,-62.1573"/>
+<!-- t14 -->
+<g id="node15" class="node"><title>t14</title>
+<polygon fill="#efa03b" stroke="black" points="774.472,-105.333 939,-78.005 1103.53,-105.333 1103.37,-149.551 774.625,-149.551 774.472,-105.333"/>
+<polygon fill="none" stroke="black" points="770.46,-101.94 939,-73.9453 1107.54,-101.94 1107.36,-153.556 770.639,-153.556 770.46,-101.94"/>
+<text text-anchor="middle" x="939" y="-111.555" font-family="Times,serif" font-size="20.00">merge_pages</text>
+</g>
+<!-- t10&#45;&gt;t14 -->
+<g id="edge23" class="edge"><title>t10&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M862.571,-182.814C870.479,-176.165 879.888,-168.254 889.356,-160.293"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="891.89,-162.736 897.292,-153.622 887.385,-157.378 891.89,-162.736"/>
+</g>
+<!-- t9&#45;&gt;t14 -->
+<g id="edge24" class="edge"><title>t9&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M546.87,-241.103C586.226,-225.054 647.626,-200.869 702,-183.109 730.59,-173.771 761.415,-164.679 791.074,-156.405"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="792.38,-159.675 801.082,-153.632 790.511,-152.93 792.38,-159.675"/>
+</g>
+<!-- t13&#45;&gt;t14 -->
+<g id="edge20" class="edge"><title>t13&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M986.636,-414.956C1033.58,-398.748 1087,-369.103 1087,-318.109 1087,-318.109 1087,-318.109 1087,-258.109 1087,-215.948 1054.81,-182.761 1020.46,-159.346"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1021.98,-156.158 1011.7,-153.604 1018.15,-162.013 1021.98,-156.158"/>
+</g>
+<!-- t11&#45;&gt;t14 -->
+<g id="edge22" class="edge"><title>t11&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M969.056,-299.033C979.177,-293.579 988.239,-286.44 995,-277.109 1020.03,-242.568 998.243,-195.803 974.929,-162.021"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="977.661,-159.825 968.996,-153.728 971.968,-163.897 977.661,-159.825"/>
+</g>
+<!-- t12&#45;&gt;t14 -->
+<g id="edge21" class="edge"><title>t12&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M1006.64,-472.956C1053.58,-456.748 1107,-427.103 1107,-376.109 1107,-376.109 1107,-376.109 1107,-258.109 1107,-220.742 1095.53,-209.426 1069,-183.109 1059.8,-173.988 1049.06,-165.927 1037.76,-158.869"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1039.41,-155.777 1029.03,-153.665 1035.83,-161.791 1039.41,-155.777"/>
+</g>
+<!-- t15 -->
+<g id="node16" class="node"><title>t15</title>
+<polygon fill="#efa03b" stroke="black" points="1053.18,-52 828.822,-52 824.822,-48 824.822,-16 1049.18,-16 1053.18,-20 1053.18,-52"/>
+<polyline fill="none" stroke="black" points="1049.18,-48 824.822,-48 "/>
+<polyline fill="none" stroke="black" points="1049.18,-48 1049.18,-16 "/>
+<polyline fill="none" stroke="black" points="1049.18,-48 1053.18,-52 "/>
+<text text-anchor="middle" x="939" y="-28" font-family="Times,serif" font-size="20.00">copy_final</text>
+</g>
+<!-- t14&#45;&gt;t15 -->
+<g id="edge25" class="edge"><title>t14&#45;&gt;t15</title>
+<path fill="none" stroke="#0044a0" d="M939,-73.8665C939,-69.8921 939,-65.942 939,-62.1676"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="942.5,-62.1213 939,-52.1214 935.5,-62.1214 942.5,-62.1213"/>
 </g>
 </g>
 </svg>

requirements.txt

@@ -1,4 +1,5 @@
 ruffus>=2.6.3
 Pillow>=2.4.0
 reportlab>=3.1.44
-PyPDF2>=1.25.1
+PyPDF2>=1.25.1
+git+https://github.com/jbarlow83/img2pdf.git@e9bcce0afc3720752ca53a991db93f911a1df709#egg=img2pdf-0.1.5.dev

setup.cfg

@@ -1,2 +1,4 @@
 [bdist_wheel]
-python-tag = py34
+python-tag = py34
+[pytest]
+norecursedirs = lib 

setup.py

@@ -1,14 +1,18 @@
 #!/usr/bin/env python3
 # © 2015 James R. Barlow: github.com/jbarlow83
 
+from __future__ import print_function, unicode_literals
 from setuptools import setup
-from subprocess import Popen, STDOUT, check_output, CalledProcessError
-from string import Template
+from subprocess import STDOUT, check_output, CalledProcessError
 from collections.abc import Mapping
 import re
 import sys
 
 
+if sys.version_info < (3, 4):
+    print("Python 3.4 or newer is required")
+    sys.exit(1)
+
 missing_program = '''
 The program '{program}' could not be executed or was not found on your
 system PATH.
@@ -174,12 +178,10 @@ if 'upload' in sys.argv[1:]:
     print('Use twine to upload the package - setup.py upload is insecure')
     sys.exit(1)
 
-install_requires = open('requirements.txt').read().splitlines()
 tests_require = open('test_requirements.txt').read().splitlines()
 
 setup(
     name='ocrmypdf',
-    version='3.0',  # also update: release notes, main.py
     description='OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to be searched',
     url='https://github.com/jbarlow83/OCRmyPDF',
     author='James R. Barlow',
@@ -203,7 +205,17 @@ setup(
         "Topic :: Text Processing :: Indexing",
         "Topic :: Text Processing :: Linguistic",
         ],
-    install_requires=install_requires,
+    setup_requires=[
+        'setuptools_scm'
+    ],
+    use_scm_version={'version_scheme': 'post-release'},
+    install_requires=[
+        'ruffus',
+        'Pillow',
+        'reportlab',
+        'PyPDF2',
+        'img2pdf'
+    ],
     tests_require=tests_require,
     entry_points={
         'console_scripts': [

test_requirements.txt

@@ -1,2 +1 @@
-img2pdf>=0.1.5
 pytest>=2.7.2

tests/replace_ghostscript_nopdfa.py

@@ -1,72 +0,0 @@
-#!/usr/bin/env python3
-import sys
-import os
-
-'''
-For testing PDF/A generation failures, this wrapper calls ghostscript
-but filters out all PDF/A information so that a regular PDF will be
-created instead.
-
-It assumes that it is called from a staged system PATH where the first
-item on the PATH contains a file named 'gs' which is a symlink to this
-file.  It will strip out the first item on path to invoke the real
-'gs'.  That is, it expects this when called
-
-1.   tests/output/[...test name...]/bin/gs is a symlink to this file
-2.   tests/output/bin is the first item on PATH
-3.   The real executable is on the path
-
-OCRmyPDF also calls "gs --version" and "gs --help".  gs answers both
-on stdout so, this wrapper prints to stderr.
-
-'''
-
-
-def pdfa_param(arg):
-    if arg.startswith('-sPDFA'):
-        return True
-    if arg.startswith('-dPDFA'):
-        return True
-    if arg.endswith('.ps'):
-        return True
-    if 'ColorConversionStrategy' in arg:
-        return True
-    if 'ProcessColorModel' in arg:
-        return True
-    if 'OutputICCProfile' in arg:
-        return True
-    return False
-
-
-if __name__ == '__main__':
-    sys_args = sys.argv[1:]
-
-    print("Fake Ghostscript wrapper", file=sys.stderr)
-
-    if any(pdfa_param(arg) for arg in sys_args):
-        # We were asked to produce a PDF/A
-        # Filter out PDF/A arguments
-        args = [arg for arg in sys.argv[1:]
-                if not pdfa_param(arg)]
-
-        # Tell Ghostscript to create a PDF 1.3 instead so JHOVE won't
-        # think it's a PDF/A
-        indexof_pdfwrite = next(n for n, item in enumerate(args)
-                                if 'pdfwrite' in item)
-        args.insert(indexof_pdfwrite + 1, '-dCompatibilityLevel=1.3')
-        print("Rewrote arguments", file=sys.stderr)
-    else:
-        args = sys_args
-        print("Keeping arguments", file=sys.stderr)
-
-    exec_path = os.environ['PATH'].split(os.pathsep)
-    env = os.environ.copy()
-    env['PATH'] = os.pathsep.join(exec_path[1:])
-
-    print("Calling real Ghostscript with modified args and env",
-          file=sys.stderr)
-    print(args, file=sys.stderr)
-    print(env['PATH'], file=sys.stderr)
-
-    sys.stderr.flush()
-    os.execvpe('gs', ['gs'] + args, env)

tests/spoof/qpdf_dummy_return2.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python3
+import sys
+
+
+def main():
+    print('qpdf dummy')
+    sys.exit(2)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/tesseract_cache.py

@@ -0,0 +1,78 @@
+#!/usr/bin/env python3
+import sys
+import os
+import hashlib
+import shutil
+import subprocess
+
+
+CACHE_PATH = os.path.abspath(os.path.join(
+        os.path.dirname(__file__), '..', 'cache'))
+
+
+def main():
+    operation = sys.argv[-1]
+    # For anything except a hocr or pdf, defer to real tesseract
+    if operation != 'hocr' and operation != 'pdf':
+        tess_args = ['tesseract'] + sys.argv[1:]
+        os.execvp("tesseract", tess_args)
+        return  # Not reachable
+
+    try:
+        os.makedirs(CACHE_PATH)
+    except FileExistsError:
+        pass
+
+    m = hashlib.sha1()
+
+    version = subprocess.check_output(
+        ['tesseract', '--version'],
+        stderr=subprocess.STDOUT)
+
+    m.update(version)
+    m.update(operation.encode())
+
+    try:
+        lang = sys.argv[sys.argv.index('-l') + 1]
+        m.update(lang.encode())
+    except ValueError:
+        pass
+    try:
+        psm = sys.argv[sys.argv.index('-psm') + 1]
+        m.update(psm.encode())
+    except ValueError:
+        pass
+
+    input_file = sys.argv[-3]
+    output_file = sys.argv[-2]
+
+    if operation == 'hocr':
+        output_file += '.hocr'
+    elif operation == 'pdf':
+        output_file += '.pdf'
+
+    with open(input_file, 'rb') as f:
+        m.update(f.read())
+
+    cache_name = os.path.join(CACHE_PATH, m.hexdigest())
+    if os.path.exists(cache_name):
+        # Cache hit
+        print("Tesseract cache hit", file=sys.stderr)
+        shutil.copy(cache_name, output_file)
+        sys.exit(0)
+
+    # Cache miss
+    print("Tesseract cache miss", file=sys.stderr)
+
+    # Call tesseract
+    subprocess.check_call(['tesseract'] + sys.argv[1:])
+
+    # Insert file into cache
+    if os.path.exists(output_file):
+        shutil.copy(output_file, cache_name)
+    else:
+        print("Could not find output file", file=sys.stderr)
+
+
+if __name__ == '__main__':
+    main()

tests/spoof/tesseract_noop.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+import sys
+import img2pdf
+from PIL import Image
+
+
+VERSION_STRING = '''tesseract 3.04.00
+ leptonica-1.72
+  libjpeg 8d : libpng 1.6.19 : libtiff 4.0.6 : zlib 1.2.5
+SPOOFED
+'''
+
+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>'''
+
+
+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] == 'hocr':
+        inputf = sys.argv[-3]
+        output = sys.argv[-2]
+        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)
+    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/test_hocrtransform.py

@@ -0,0 +1,61 @@
+#!/usr/bin/env python3
+# © 2015 James R. Barlow: github.com/jbarlow83
+
+from ocrmypdf import hocrtransform
+from ocrmypdf.tesseract import HOCR_TEMPLATE
+from reportlab.pdfgen.canvas import Canvas
+from PIL import Image
+from tempfile import NamedTemporaryFile
+from contextlib import suppress
+import os
+import shutil
+import pytest
+import img2pdf
+import pytest
+import sys
+
+
+if sys.version_info.major < 3:
+    print("Requires Python 3.4+")
+    sys.exit(1)
+
+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 = os.path.join(PROJECT_ROOT, 'OCRmyPDF.sh')
+TEST_RESOURCES = os.path.join(PROJECT_ROOT, 'tests', 'resources')
+TEST_OUTPUT = os.environ.get(
+    'OCRMYPDF_TEST_OUTPUT',
+    default=os.path.join(PROJECT_ROOT, 'tests', 'output', 'hocrtransform'))
+
+
+def setup_module():
+    with suppress(FileNotFoundError):
+        shutil.rmtree(TEST_OUTPUT)
+    with suppress(FileExistsError):
+        os.makedirs(TEST_OUTPUT)
+    with open(_make_output('blank.hocr'), 'w') as f:
+        f.write(HOCR_TEMPLATE)
+
+
+def _make_input(input_basename):
+    return os.path.join(TEST_RESOURCES, input_basename)
+
+
+def _make_output(output_basename):
+    return os.path.join(TEST_OUTPUT, output_basename)
+
+
+def test_mono_image():
+    im = Image.new('1', (8, 8), 0)
+    for n in range(8):
+        im.putpixel((n, n), 1)
+    im.save(_make_output('mono.tif'), format='TIFF')
+
+    hocr = hocrtransform.HocrTransform(_make_output('blank.hocr'), 300)
+    hocr.to_pdf(_make_output('mono.pdf'), imageFileName=_make_output('mono.tif'))
+
+
+
+
+

tests/test_main.py

@@ -2,7 +2,7 @@
 # © 2015 James R. Barlow: github.com/jbarlow83
 
 from __future__ import print_function
-from subprocess import Popen, PIPE, check_output
+from subprocess import Popen, PIPE, check_output, check_call
 import os
 import shutil
 from contextlib import suppress
@@ -18,27 +18,27 @@ if sys.version_info.major < 3:
     sys.exit(1)
 
 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 = os.path.join(PROJECT_ROOT, 'OCRmyPDF.sh')
 TEST_RESOURCES = os.path.join(PROJECT_ROOT, 'tests', 'resources')
 TEST_OUTPUT = os.environ.get(
     'OCRMYPDF_TEST_OUTPUT',
-    default=os.path.join(PROJECT_ROOT, 'tests', 'output'))
-TEST_BINARY_PATH = os.path.join(TEST_OUTPUT, 'fakebin')
+    default=os.path.join(PROJECT_ROOT, 'tests', 'output', 'main'))
 
 
 def setup_module():
     with suppress(FileNotFoundError):
         shutil.rmtree(TEST_OUTPUT)
     with suppress(FileExistsError):
-        os.mkdir(TEST_OUTPUT)
+        os.makedirs(TEST_OUTPUT)
 
 
-def run_ocrmypdf_sh(input_file, output_file, *args):
+def run_ocrmypdf_sh(input_file, output_file, *args, env=None):
     sh_args = ['sh', OCRMYPDF] + list(args) + [input_file, output_file]
     sh = Popen(
         sh_args, close_fds=True, stdout=PIPE, stderr=PIPE,
-        universal_newlines=True)
+        universal_newlines=True, env=env)
     out, err = sh.communicate()
     return sh, out, err
 
@@ -51,12 +51,12 @@ def _make_output(output_basename):
     return os.path.join(TEST_OUTPUT, output_basename)
 
 
-def check_ocrmypdf(input_basename, output_basename, *args):
+def check_ocrmypdf(input_basename, output_basename, *args, env=None):
     input_file = _make_input(input_basename)
     output_file = _make_output(output_basename)
 
-    sh, _, err = run_ocrmypdf_sh(input_file, output_file, *args)
-    assert sh.returncode == 0, err
+    sh, out, err = run_ocrmypdf_sh(input_file, output_file, *args, env=env)
+    assert sh.returncode == 0, dict(stdout=out, stderr=err)
     assert os.path.exists(output_file), "Output file not created"
     assert os.stat(output_file).st_size > 100, "PDF too small or empty"
     return output_file
@@ -77,13 +77,32 @@ def run_ocrmypdf_env(input_basename, output_basename, *args, env=None):
     return p, out, err
 
 
-def test_quick():
-    check_ocrmypdf('c02-22.pdf', 'test_quick.pdf')
+@pytest.fixture
+def spoof_tesseract_noop():
+    env = os.environ.copy()
+    program = os.path.join(SPOOF_PATH, 'tesseract_noop.py')
+    check_call(['chmod', "+x", program])
+    env['OCRMYPDF_TESSERACT'] = program
+    return env
 
 
-def test_deskew():
+@pytest.fixture
+def spoof_tesseract_cache():
+    env = os.environ.copy()
+    program = os.path.join(SPOOF_PATH, "tesseract_cache.py")
+    check_call(['chmod', '+x', program])
+    env['OCRMYPDF_TESSERACT'] = program
+    return env
+
+
+def test_quick(spoof_tesseract_noop):
+    check_ocrmypdf('c02-22.pdf', 'test_quick.pdf', env=spoof_tesseract_noop)
+
+
+def test_deskew(spoof_tesseract_noop):
     # Run with deskew
-    deskewed_pdf = check_ocrmypdf('skew.pdf', 'test_deskew.pdf', '-d')
+    deskewed_pdf = check_ocrmypdf(
+        'skew.pdf', 'test_deskew.pdf', '-d', env=spoof_tesseract_noop)
 
     # Now render as an image again and use Leptonica to find the skew angle
     # to confirm that it was deskewed
@@ -110,29 +129,34 @@ def test_deskew():
     assert -0.5 < skew_angle < 0.5, "Deskewing failed"
 
 
-def test_clean():
-    check_ocrmypdf('skew.pdf', 'test_clean.pdf', '-c')
+def test_clean(spoof_tesseract_noop):
+    check_ocrmypdf('skew.pdf', 'test_clean.pdf', '-c', env=spoof_tesseract_noop)
 
 
-def check_exotic_image(pdf, renderer):
+@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')
+])
+def test_exotic_image(spoof_tesseract_cache, pdf, renderer):
     check_ocrmypdf(
         pdf,
         'test_{0}_{1}.pdf'.format(pdf, renderer),
         '-dc',
-        '--pdf-renderer', renderer)
+        '-v', '1',
+        '--pdf-renderer', renderer, env=spoof_tesseract_cache)
 
 
-def test_exotic_image():
-    yield check_exotic_image, 'palette.pdf', 'hocr'
-    yield check_exotic_image, 'palette.pdf', 'tesseract'
-    yield check_exotic_image, 'cmyk.pdf', 'hocr'
-    yield check_exotic_image, 'cmyk.pdf', 'tesseract'
-
-
-def test_preserve_metadata():
+def test_preserve_metadata(spoof_tesseract_noop):
     pdf_before = pypdf.PdfFileReader(_make_input('graph.pdf'))
 
-    output = check_ocrmypdf('graph.pdf', 'test_metadata_preserve.pdf')
+    output = check_ocrmypdf('graph.pdf', 'test_metadata_preserve.pdf',
+                            env=spoof_tesseract_noop)
 
     pdf_after = pypdf.PdfFileReader(output)
 
@@ -140,7 +164,7 @@ def test_preserve_metadata():
         assert pdf_before.documentInfo[key] == pdf_after.documentInfo[key]
 
 
-def test_override_metadata():
+def test_override_metadata(spoof_tesseract_noop):
     input_file = _make_input('c02-22.pdf')
     output_file = _make_output('test_override_metadata.pdf')
 
@@ -152,7 +176,8 @@ def test_override_metadata():
         input_file, output_file,
         '--title', german,
         '--author', chinese,
-        '--subject', high_unicode)
+        '--subject', high_unicode,
+        env=spoof_tesseract_noop)
 
     assert p.returncode == ExitCode.ok
 
@@ -171,10 +196,15 @@ def test_override_metadata():
     assert pdfinfo.get('Keywords', '') == ''
 
 
-def check_oversample(renderer):
+@pytest.mark.parametrize('renderer', [
+    'hocr',
+    'tesseract',
+    ])
+def test_oversample(spoof_tesseract_cache, renderer):
     oversampled_pdf = check_ocrmypdf(
         'skew.pdf', 'test_oversample_%s.pdf' % renderer, '--oversample', '300',
-        '--pdf-renderer', renderer)
+        '-f',
+        '--pdf-renderer', renderer, env=spoof_tesseract_cache)
 
     pdfinfo = pdf_get_all_pageinfo(oversampled_pdf)
 
@@ -182,111 +212,61 @@ def check_oversample(renderer):
     assert abs(pdfinfo[0]['xres'] - 300) < 1
 
 
-def test_oversample():
-    yield check_oversample, 'hocr'
-    yield check_oversample, 'tesseract'
-
-
 def test_repeat_ocr():
     sh, _, _ = run_ocrmypdf_sh('graph_ocred.pdf', 'wontwork.pdf')
     assert sh.returncode != 0
 
 
-def test_force_ocr():
-    out = check_ocrmypdf('graph_ocred.pdf', 'test_force.pdf', '-f')
+def test_force_ocr(spoof_tesseract_cache):
+    out = check_ocrmypdf('graph_ocred.pdf', 'test_force.pdf', '-f',
+                         env=spoof_tesseract_cache)
     pdfinfo = pdf_get_all_pageinfo(out)
     assert pdfinfo[0]['has_text']
 
 
-def test_skip_ocr():
-    check_ocrmypdf('graph_ocred.pdf', 'test_skip.pdf', '-s')
+def test_skip_ocr(spoof_tesseract_cache):
+    check_ocrmypdf('graph_ocred.pdf', 'test_skip.pdf', '-s',
+                   env=spoof_tesseract_cache)
 
 
-def test_argsfile():
+def test_argsfile(spoof_tesseract_noop):
     with open(_make_output('test_argsfile.txt'), 'w') as argsfile:
         print('--title', 'ArgsFile Test', '--author', 'Test Cases',
               sep='\n', end='\n', file=argsfile)
     check_ocrmypdf('graph.pdf', 'test_argsfile.pdf',
-                   '@' + _make_output('test_argsfile.txt'))
+                   '@' + _make_output('test_argsfile.txt'),
+                   env=spoof_tesseract_noop)
 
 
-def check_ocr_timeout(renderer):
+@pytest.mark.parametrize('renderer', [
+    'hocr',
+    'tesseract',
+    ])
+def test_ocr_timeout(renderer):
     out = check_ocrmypdf('skew.pdf', 'test_timeout_%s.pdf' % renderer,
                          '--tesseract-timeout', '1.0')
     pdfinfo = pdf_get_all_pageinfo(out)
     assert pdfinfo[0]['has_text'] == False
 
 
-def test_ocr_timeout():
-    yield check_ocr_timeout, 'hocr'
-    yield check_ocr_timeout, 'tesseract'
-
-
-def test_skip_big():
+def test_skip_big(spoof_tesseract_cache):
     out = check_ocrmypdf('enormous.pdf', 'test_enormous.pdf',
-                         '--skip-big', '10')
+                         '--skip-big', '10', env=spoof_tesseract_cache)
     pdfinfo = pdf_get_all_pageinfo(out)
     assert pdfinfo[0]['has_text'] == False
 
 
-def check_maximum_options(renderer):
+@pytest.mark.parametrize('renderer', [
+    'hocr',
+    'tesseract',
+    ])
+def test_maximum_options(spoof_tesseract_cache, renderer):
     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)
-
-
-def test_maximum_options():
-    yield check_maximum_options, 'hocr'
-    yield check_maximum_options, 'tesseract'
-
-
-def override_binary(binary, replacement):
-    '''Create a directory that contains a symlink named 'binary' that
-    points to replacement, another program to use in place of the
-    regular binary for testing.
-
-    override_binary('gs', 'replace_gs.py') will create an environment
-    in which "gs" will invoke replace_gs.py.
-
-    Not thread-safe with other test at the moment.
-
-    Returns the os.environ["PATH"] string under which this binary will
-    be invoked.'''
-
-    replacement_path = os.path.abspath(os.path.join(TESTS_ROOT,
-                                                    replacement))
-    subdir = os.path.splitext(os.path.basename(replacement))[0]
-    binary_path = os.path.abspath(os.path.join(TEST_BINARY_PATH,
-                                               subdir,
-                                               binary))
-    with suppress(FileExistsError):
-        os.makedirs(os.path.dirname(binary_path))
-    assert os.path.isdir(os.path.dirname(binary_path))
-    assert not os.path.lexists(binary_path)
-    print("symlink %s -> %s" % (replacement_path, binary_path))
-    os.symlink(replacement_path, binary_path)
-
-    os.chmod(replacement_path, int('755', base=8))
-
-    return os.path.dirname(binary_path) + os.pathsep + os.environ["PATH"]
-
-
-@pytest.fixture
-def break_ghostscript_pdfa():
-    return override_binary('gs', 'replace_ghostscript_nopdfa.py')
-
-
-@pytest.mark.skipif(os.environ.get('OCRMYPDF_IN_DOCKER', False),
-                    reason="Requires writable filesystem")
-def test_ghostscript_pdfa_fails(break_ghostscript_pdfa):
-    env = os.environ.copy()
-    env['PATH'] = break_ghostscript_pdfa
-
-    p, out, err = run_ocrmypdf_env(
-        'graph_ocred.pdf', 'not_a_pdfa.pdf', '-v', '1', '--skip-text', env=env)
-    assert p.returncode == ExitCode.ok, err  # no longer using JHOVE PDFA check
+        '--author', 'py.test', '--pdf-renderer', renderer,
+        env=spoof_tesseract_cache)
 
 
 def test_tesseract_missing_tessdata():
@@ -310,9 +290,9 @@ def test_blank_input_pdf():
     assert p.returncode == ExitCode.ok
 
 
-def test_french():
+def test_french(spoof_tesseract_cache):
     p, out, err = run_ocrmypdf_env(
-        'francais.pdf', 'francais.pdf', '-l', 'fra')
+        'francais.pdf', 'francais.pdf', '-l', 'fra', env=spoof_tesseract_cache)
     assert p.returncode == ExitCode.ok, \
         "This test may fail if Tesseract language packs are missing"
 
@@ -323,8 +303,67 @@ def test_klingon():
     assert p.returncode == ExitCode.bad_args
 
 
-def test_missing_docinfo():
+def test_missing_docinfo(spoof_tesseract_noop):
     p, out, err = run_ocrmypdf_env(
-        'missing_docinfo.pdf', 'missing_docinfo.pdf', '-l', 'eng', '-c')
+        'missing_docinfo.pdf', 'missing_docinfo.pdf', '-l', 'eng', '-c',
+        env=spoof_tesseract_noop)
     assert p.returncode == ExitCode.ok, err
 
+
+def test_uppercase_extension(spoof_tesseract_noop):
+    shutil.copy(_make_input("skew.pdf"), _make_input("UPPERCASE.PDF"))
+    try:
+        check_ocrmypdf("UPPERCASE.PDF", "UPPERCASE_OUT.PDF",
+                       env=spoof_tesseract_noop)
+    finally:
+        os.unlink(_make_input("UPPERCASE.PDF"))
+
+
+def test_input_file_not_found():
+    input_file = "does not exist.pdf"
+    sh, out, err = run_ocrmypdf_sh(
+        _make_input(input_file),
+        _make_output("will not happen.pdf"))
+    assert sh.returncode == ExitCode.input_file
+    assert (input_file in out or input_file in err)
+
+
+def test_input_file_not_a_pdf():
+    input_file = __file__  # Try to OCR this file
+    sh, out, err = run_ocrmypdf_sh(
+        _make_input(input_file),
+        _make_output("will not happen.pdf"))
+    assert sh.returncode == ExitCode.input_file
+    assert (input_file in out or input_file in err)
+
+
+def test_qpdf_repair_fails():
+    env = os.environ.copy()
+    env['OCRMYPDF_QPDF'] = os.path.abspath('./spoof/qpdf_dummy_return2.py')
+    p, out, err = run_ocrmypdf_env(
+        '-v', '1',
+        'c02-22.pdf', 'wont_be_created.pdf', env=env)
+    print(out)
+    print(err)
+    assert p.returncode == ExitCode.input_file
+
+
+def test_encrypted():
+    p, out, err = run_ocrmypdf_env('skew-encrypted.pdf', 'wont_be_created.pdf')
+    assert p.returncode == ExitCode.input_file
+    assert out.find('password')
+
+
+@pytest.mark.parametrize('renderer', [
+    'hocr',
+    'tesseract',
+    ])
+def test_pagesegmode(renderer, spoof_tesseract_cache):
+    check_ocrmypdf(
+        'skew.pdf', 'test_psm_%s.pdf' % renderer,
+        '--tesseract-pagesegmode', '7',
+        '-v', '1',
+        '--pdf-renderer', renderer, env=spoof_tesseract_cache)
+
+
+

tests/test_pageinfo.py

@@ -10,20 +10,37 @@ import os
 import shutil
 import pytest
 import img2pdf
-from pkg_resources import Requirement, resource_filename
+import pytest
+import sys
 
-req = Requirement.parse('ocrmypdf')
 
+if sys.version_info.major < 3:
+    print("Requires Python 3.4+")
+    sys.exit(1)
+
+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 = os.path.join(PROJECT_ROOT, 'OCRmyPDF.sh')
+TEST_RESOURCES = os.path.join(PROJECT_ROOT, 'tests', 'resources')
 TEST_OUTPUT = os.environ.get(
     'OCRMYPDF_TEST_OUTPUT',
-    default=os.path.join(os.path.dirname(__file__), 'output'))
+    default=os.path.join(PROJECT_ROOT, 'tests', 'output', 'pageinfo'))
 
 
 def setup_module():
     with suppress(FileNotFoundError):
         shutil.rmtree(TEST_OUTPUT)
     with suppress(FileExistsError):
-        os.mkdir(TEST_OUTPUT)
+        os.makedirs(TEST_OUTPUT)
+
+
+def _make_input(input_basename):
+    return os.path.join(TEST_RESOURCES, input_basename)
+
+
+def _make_output(output_basename):
+    return os.path.join(TEST_OUTPUT, output_basename)
 
 
 def test_single_page_text():
@@ -100,7 +117,7 @@ def test_single_page_inline_image():
 
 
 def test_jpeg():
-    filename = resource_filename(req, 'tests/resources/c02-22.pdf')
+    filename = _make_input('c02-22.pdf')
 
     pdfinfo = pageinfo.pdf_get_all_pageinfo(filename)