Re-enable macos

Reorganize and see if this is better now

.travis.yml

@@ -44,7 +44,8 @@ install:
 script:
 - mv ocrmypdf dont_import_this_ocrmypdf
 - tesseract --version
-- pytest -n auto
+- qpdf --version
+- pytest
 - mv dont_import_this_ocrmypdf ocrmypdf
 
 deploy:

README.rst

@@ -122,6 +122,13 @@ If you detect an issue, please:
 -  If possible provide your input PDF file as well as the content of the
    temporary folder (using a file sharing service like Dropbox)
 
+Requirements
+------------
+
+Runs on CPython 3.6, and requires external program installations of Ghostscript, Tesseract OCR, QPDF, and Leptonica. ocrmypdf is pure Python, but uses CFFI to portably generate library bindings.
+
+Python 3.5 is fully supported, but support will end at the end of 2017, to allow f-strings in the code base.
+
 Press & Media
 -------------
 

docs/batch.rst

@@ -147,7 +147,7 @@ This is only possible for x86-based Synology products. Some Synology products us
 	# use this uid:gid in -u flag
 	# rw rights for docker:docker at source dir are also necessary
 	# the script is processed as root user via chron 
-	            cmd = ['docker', 'run', '--rm', '-v', docker_mount, '-u="1030:65538"', 'jbarlow83/ocrmypdf', , '--deskew' , filename, filename_OCR]
+	            cmd = ['docker', 'run', '--rm', '-v', docker_mount, '-u=1030:65538', 'jbarlow83/ocrmypdf', , '--deskew' , filename, filename_OCR]
 	            logging.info(cmd)
 	            proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
 	            result = proc.stdout.read()

docs/installation.rst

@@ -125,6 +125,9 @@ Or in the wonderful `fish shell <https://fishshell.com/>`_:
     alias ocrmypdf 'docker run --rm -v (pwd):/home/docker ocrmypdf'
     funcsave ocrmypdf
 
+.. note::
+
+    The ocrmypdf Docker containers are designed to be used for a single OCR job. The ``docker run --rm`` argument tells Docker to delete temporary storage associated with container when it is done executing.
 
 Manual installation on macOS
 ----------------------------
@@ -171,7 +174,13 @@ Update the homebrew pip and install Pillow:
     pip3 install --upgrade pip
     pip3 install --upgrade pillow
 
-You can then install OCRmyPDF from PyPI:
+You can then install OCRmyPDF from PyPI, for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf
+
+or system-wide:
 
 .. code-block:: bash
 
@@ -200,8 +209,13 @@ No package is currently available for Ubuntu 16.04, but you can install the depe
         python3-pip \
         python3-cffi
 
-If you wish install OCRmyPDF to the system Python, then install as follows (note this installs new packages
-into your system Python, which could interfere with other programs):
+If you wish install OCRmyPDF for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf
+
+Alternately, system-wide. Note that this may modify the system Python environment:
 
 .. code-block:: bash
 
@@ -221,7 +235,7 @@ Installing on Ubuntu 14.04 LTS
 ------------------------------
 
 Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other options,
-because it is older.
+because it is older and does not provide ``pip``.
 
 Update apt-get:
 

docs/release_notes.rst

@@ -5,6 +5,18 @@ OCRmyPDF uses `semantic versioning <http://semver.org/>`_ for its command line i
 
 The OCRmyPDF package itself does not contain a public API, although it is fairly stable and breaking changes are usually timed with a major release. A future release will clearly define the stable public API.
 
+v5.4.4
+------
+
+-   Fix issue #181: fix final merge failure for PDFs with more pages than the system file handle limit (``ulimit -n``)
+-   Fix issue #200: an uncommon syntax for formatting decimal numbers in a PDF would cause qpdf to issue a warning, which ocrmypdf treated as an error. Now this the warning is relayed.
+-   Fix an issue where intermediate PDFs would be created at version 1.3 instead of the version of the original file. It's possible but unlikely this had side effects.
+-   A warning is now issued when older versions of qpdf are used since issues like #200 cause qpdf to infinite-loop
+-   Address issue #140: if Tesseract outputs invalid UTF-8, escape it and print its message instead of aborting with a Unicode error 
+-   Adding previously unlisted setup requirement, pytest-runner
+-   Update documentation: fix an error in the example script for Synology with Docker images, improved security guidance, advised ``pip install --user``
+
+
 v5.4.3
 ------
 

docs/security.rst

@@ -10,7 +10,7 @@ The disclaimer applies: this software has no warranties of any kind.
 PDFs may contain malware
 ------------------------
 
-PDF is a rich, complex file format. The official PDF 1.7 specification, ISO 32000:2008, is hundreds of packages long and references several annexes each of which are similar in length. PDFs can contain video, audio, JavaScript and other programming, and forms. In some cases, they can open internet connections to pre-selected URLs. All of these possible attack vectors.
+PDF is a rich, complex file format. The official PDF 1.7 specification, ISO 32000:2008, is hundreds of pages long and references several annexes each of which are similar in length. PDFs can contain video, audio, XML, JavaScript and other programming, and forms. In some cases, they can open internet connections to pre-selected URLs. All of these possible attack vectors.
 
 In short, PDFs `may contain viruses <https://security.stackexchange.com/questions/64052/can-a-pdf-file-contain-a-virus>`_.
 
@@ -21,18 +21,45 @@ How OCRmyPDF processes PDFs
 
 OCRmyPDF must open and interpret your PDF in order to insert an OCR layer. First, it runs all PDFs through `qpdf <https://github.com/qpdf/qpdf>`_, a program that repairs PDFs with syntax errors. This is done because, in the author's experience, a significant number of PDFs in the wild especially those created by scanners are not well-formed files. qpdf makes it more likely that OCRmyPDF will succeed, but offers no security guarantees. qpdf is also used to split the PDF into single page PDFs.
 
-After qpdf, OCRmyPDF examines each page using `PyPDF2 <https://github.com/mstamy2/PyPDF2>`_. This library also has no warranties or guarantees.
+After qpdf, OCRmyPDF examines each page using `PyPDF2 <https://github.com/mstamy2/PyPDF2>`_. This library also has no warranties or guarantees. OCRmyPDF works with qpdf 5.0 and up, but version 7.0 is recommended because of known security vulnerabilities in early versions.
 
-Finally, OCRmyPDF rasterizes each page of the PDF using `Ghostscript <http://ghostscript.com/>`_ in ``-dSAFER`` mode. 
+Finally, OCRmyPDF rasterizes each page of the PDF using `Ghostscript <http://ghostscript.com/>`_ in ``-dSAFER`` mode.
 
 Depending on the options specified, OCRmyPDF may graft the OCR layer into the existing PDF or it may essentially reconstruct ("re-fry") a visually identical PDF that may be quite different at the binary level. That said, OCRmyPDF is not a tool designed for sanitizing PDFs.
 
-Using OCRmyPDF online
----------------------
+Using OCRmyPDF online or as a service
+-------------------------------------
 
-OCRmyPDF is not designed to be deployed "as a service", in a setting where a user/attacker could upload a file for OCR processing online. It is not designed to be secure in this case.
+OCRmyPDF should not be deployed as a public-facing service, like a website where a potential attacker could upload a PDF of their choice for OCR. OCRmyPDF is not designed to be secure against PDF malware. Another concern is PDFs specifically designed to be a denial of service attack: PDFs can contain recursive data structures that sometimes send parsers into infinite loops, and issue complex graphics drawing commands.
+
+OCRmyPDF should be relatively safe to use in a trusted intranet, with some considerations:
+
+Limiting CPU usage
+^^^^^^^^^^^^^^^^^^
+
+OCRmyPDF will attempt to use all available CPUs and storage, so executing ``nice ocrmypdf`` or limiting the number of jobs with the ``-j`` argument may ensure the server remains available. Another option would be run OCRmyPDF jobs inside a Docker container or virtual machine, which can impose its own limits on CPU usage.
+
+Temporary storage requirements
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+OCRmyPDF will use a large amount of temporary storage for its work, proportional to the total number of pixels needed to rasterize the PDF. The raster image of a 8.5×11" color page at 300 DPI takes 25 MB uncompressed; OCRmyPDF saves its intermediates as PNG, but that still means it requires about 9 MB per intermediate based on average compression ratios. Multiple intermediates per page are also required, depending on the command line given. A rule of thumb would be to allow 100 MB of temporary storage per page in a file – meaning that a small cloud servers or small VM partitions should be provisioned with plenty of extra space, if say, a 500 page file might be sent.
+
+To check temporary storage usage on actual files, run ``ocrmypdf -k ...`` which will preserve and print the path to temporary storage when the job is done.
+
+To change where temporary files are stored, change the ``TMPDIR`` environment variable for ocrmypdf's environment. (Python's ``tempfile.gettempdir()`` returns the root directory in which temporary files will be stored.)
+
+Timeouts
+^^^^^^^^
+
+To prevent excessively long OCR jobs consider setting ``--tesseract-timeout`` and/or ``--skip-big`` arguments. ``--skip-big`` is particularly helpful if your PDFs include documents such as reports on standard page sizes with large images attached - often large images are not worth OCR'ing anyway.
+
+Commercial alternatives
+^^^^^^^^^^^^^^^^^^^^^^^
+
+The author also provides professional services that include OCR and building databases around PDFs, and is happy to provide consultation.
+
+Abbyy Cloud OCR is a viable commercial alternative with a web services API. 
 
-Abbyy Cloud OCR is a viable commercial alternative with a web services API. The author also provides professional services that include OCR and building databases around PDFs, and is happy to provide consultation.
 
 Password protection, digital signatures and certification
 ---------------------------------------------------------

ocrmypdf/__main__.py

@@ -28,7 +28,7 @@ from . import PROGRAM_NAME, VERSION
 
 from .exceptions import *
 from . import exceptions as ocrmypdf_exceptions
-from . import _unicodefun
+from ._unicodefun import verify_python3_env
 
 warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
 
@@ -50,7 +50,10 @@ def complain(message):
 if 'IDE_PROJECT_ROOTS' in os.environ:
     os.environ['PATH'] = '/usr/local/bin:' + os.environ['PATH']
 
-_unicodefun._verify_python3_env()
+# -------- 
+# Critical environment tests
+
+verify_python3_env()
 
 if tesseract.version() < MINIMUM_TESS_VERSION:
     complain(
@@ -679,9 +682,19 @@ def run_pipeline():
         logging_factory, __name__, logger_args)
     _log.debug('ocrmypdf ' + VERSION)
     _log.debug('tesseract ' + tesseract.version())
+    _log.debug('qpdf ' + qpdf.version())
 
     check_options(options, _log)
 
+    # Complain about qpdf version < 7.0.0
+    # Suppress the warning if in the test suite, since there are no PPAs
+    # for qpdf 7.0.0 for Ubuntu trusty (i.e. Travis)
+    if qpdf.version() < '7.0.0' and not os.environ.get('PYTEST_CURRENT_TEST'):
+        complain(
+            "You are using qpdf version {0} which has known issues including "
+            "security vulnerabilities with certain malformed PDFs. Consider "
+            "upgrading to version 7.0.0 or newer.".format(qpdf.version()))
+
     # Any changes to options will not take effect for options that are already
     # bound to function parameters in the pipeline. (For example
     # options.input_file, options.pdf_renderer are already bound.)

ocrmypdf/_unicodefun.py

@@ -38,7 +38,7 @@ import sys
 import codecs
 
 
-def _verify_python3_env():
+def verify_python3_env():
     """Ensures that the environment is good for unicode on Python 3."""
     try:
         import locale

ocrmypdf/exec/qpdf.py

@@ -1,15 +1,17 @@
 #!/usr/bin/env python3
 # © 2017 James R. Barlow: github.com/jbarlow83
 
-from subprocess import CalledProcessError, STDOUT, PIPE, run, check_output
+from subprocess import CalledProcessError, STDOUT, PIPE, run
 from functools import lru_cache
 import sys
 import os
 import re
+import resource
 
 from ..exceptions import InputFileError, SubprocessOutputError, \
     MissingDependencyError, EncryptedPdfError
 from . import get_program, get_version
+from ..helpers import re_symlink
 
 
 @lru_cache(maxsize=1)
@@ -28,7 +30,8 @@ def check(input_file, log=None):
         import logging as log
 
     try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
+        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
+            check=True)
     except CalledProcessError as e:
         if e.returncode == 2:
             log.error("{0}: not a valid PDF, and could not repair it.".format(
@@ -57,11 +60,11 @@ def repair(input_file, output_file, log):
         get_program('qpdf'), input_file, output_file
     ]
     try:
-        check_output(args_qpdf, stderr=STDOUT, universal_newlines=True)
+        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
+            check=True)
     except CalledProcessError as e:
         if e.returncode == 3 and e.output.find("operation succeeded"):
             log.debug('qpdf found and fixed errors: ' + e.output)
-            log.debug(e.output)
             return
 
         if _probably_encrypted(e):
@@ -82,9 +85,9 @@ def repair(input_file, output_file, log):
 
 def get_npages(input_file, log):
     try:
-        pages = check_output(
+        pages = run(
             [get_program('qpdf'), '--show-npages', input_file],
-             universal_newlines=True, close_fds=True)
+            universal_newlines=True, check=True, stdout=PIPE, stderr=STDOUT)
     except CalledProcessError as e:
         if e.returncode == 2 and e.output.find('No such file'):
             log.error(e.output)
@@ -107,18 +110,97 @@ def split_pages(input_file, work_folder, npages):
         run(args_qpdf, check=True)
 
 
-def merge(input_files, output_file, min_version=None):
+def _merge_inner(input_files, output_file, min_version=None, log=None):
     """Merge the list of input files (all filenames) into the output file.
 
     The input files may contain one or more pages.
     """
+
+    # Single page 'merges' should still be attempted to that the same error
+    # checking is applied to single page case
+
     version_arg = ['--min-version={}'.format(min_version)] \
                   if min_version else []
 
+    if log is None:
+        import logging as log
+
     args_qpdf = [
         get_program('qpdf')
     ] + version_arg + [
         input_files[0], '--pages'
     ] + input_files + ['--', output_file]
-    run(args_qpdf, check=True)
-    
+
+    try:
+        run(args_qpdf, check=True, stderr=PIPE, universal_newlines=True)
+    except CalledProcessError as e:
+        if e.returncode == 3 and \
+                e.stderr.find("unknown token while reading object") and \
+                e.stderr.find("operation succeeded"):
+            # Only whitelist the 'unknown token' problem (decimal/string issue)
+            # qpdf issue #165
+            log.warning('qpdf found and fixed errors: ' + e.stderr)
+            return
+        raise e from e
+
+
+def merge(input_files, output_file, min_version=None, log=None, max_files=None):
+    """Merge the list of input files (all filenames) into the output file.
+
+    The input files may contain one or more pages.
+
+    """
+    # qpdf requires that every file that contributes to the output has a file 
+    # descriptor that remains open. That means, given our approach of one 
+    # intermediate PDF per, we can practically hit the number of file 
+    # descriptors. 
+
+    if max_files is None or max_files < 2:
+        # Find out how many open file descriptors we can get away with
+        ulimits = resource.getrlimit(resource.RLIMIT_NOFILE)
+        max_open_files = ulimits[0]
+        max_files = max_open_files // 2  # Conservative guess
+
+    # We'll write things alongside the output file
+    output_dir = os.path.dirname(output_file)
+
+    import random
+    import string    
+
+    def randstr():
+        return ''.join(random.sample(string.ascii_lowercase, 6))
+
+    # How many files to grab at once, merging all their contents
+    step_size = max_files
+
+    workqueue = input_files.copy()
+    counter = 1
+    next_workqueue = []
+    while len(workqueue) > 1 or len(next_workqueue) > 0:
+        # Take n files out of the queue
+        n = min(step_size, len(workqueue))
+        job = workqueue[0:n]
+        del workqueue[0:n]
+        print('merging ' + repr(job))
+
+        # Merge them into 1 file, which will contain n^depth pages
+        merge_file = os.path.join(
+            output_dir, "merge-{:06d}-{}.pdf".format(counter, randstr()))
+        counter += 1
+        _merge_inner(job, merge_file, min_version=min_version, log=log)
+
+        # On the next 
+        next_workqueue.append(merge_file)
+        print('next_workqueue ' + repr(next_workqueue))
+
+        # If we're out of things to do in this queue, move on to the next
+        # queue. On the counter-th pass of the workqueue we can chew through
+        # (step_size)**N pages, so on most systems the second pass finishes
+        # the job.
+        if len(workqueue) == 0:
+            workqueue = next_workqueue
+            next_workqueue = []
+
+    re_symlink(workqueue.pop(), output_file)
+
+

ocrmypdf/exec/tesseract.py

@@ -62,8 +62,7 @@ def has_textonly_pdf():
     params = ''
     try:
         params = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
+                args_tess, universal_newlines=True, stderr=STDOUT)
     except CalledProcessError as e:
         print("Could not --print-parameters from tesseract",
               file=sys.stderr)
@@ -86,8 +85,7 @@ def languages():
     ]
     try:
         langs = check_output(
-                args_tess, close_fds=True, universal_newlines=True,
-                stderr=STDOUT)
+                args_tess, universal_newlines=True, stderr=STDOUT)
     except CalledProcessError as e:
         msg = dedent("""Tesseract failed to report available languages.
         Output from Tesseract:
@@ -120,19 +118,18 @@ def get_orientation(input_file, language: list, engine_mode, timeout: float,
 
     try:
         stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, timeout=timeout)
+            args_tesseract, stderr=STDOUT, timeout=timeout)
     except TimeoutExpired:
         return OrientationConfidence(angle=0, confidence=0.0)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_file)
-        if ('Too few characters. Skipping this page' in e.output or
-                'Image too large' in e.output):
+        if (b'Too few characters. Skipping this page' in e.output or
+                b'Image too large' in e.output):
             return OrientationConfidence(0, 0)
         raise e from e
     else:
         osd = {}
-        for line in stdout.splitlines():
+        for line in stdout.decode().splitlines():
             line = line.strip()
             parts = line.split(':', maxsplit=2)
             if len(parts) == 2:
@@ -157,8 +154,17 @@ def get_orientation(input_file, language: list, engine_mode, timeout: float,
 
 
 def tesseract_log_output(log, stdout, input_file):
-    lines = stdout.splitlines()
     prefix = "{0:4d}: [tesseract] ".format(page_number(input_file))
+
+    try:
+        text = stdout.decode()
+    except UnicodeDecodeError as e:
+        log.error(prefix + "command line output was not utf-8. " +
+            "This usually means Tesseract's language packs do not match "
+            "the installed version of Tesseract.")
+        text = stdout.decode('utf-8', 'backslashreplace')
+
+    lines = text.splitlines()
     for line in lines:
         if line.startswith("Tesseract Open Source"):
             continue
@@ -231,8 +237,8 @@ def generate_hocr(input_file, output_files, language: list, engine_mode,
     try:
         log.debug(args_tesseract)
         stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, timeout=timeout)
+            args_tesseract, stderr=STDOUT,
+            timeout=timeout)
     except TimeoutExpired:
         # Generate a HOCR file with no recognized text if tesseract times out
         # Temporary workaround to hocrTransform not being able to function if
@@ -241,9 +247,9 @@ def generate_hocr(input_file, output_files, language: list, engine_mode,
         _generate_null_hocr(output_hocr, output_sidecar, input_file)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_file)
-        if 'read_params_file: parameter not found' in e.output:
+        if b'read_params_file: parameter not found' in e.output:
             raise TesseractConfigError() from e
-        if 'Image too large' in e.output:
+        if b'Image too large' in e.output:
             _generate_null_hocr(output_hocr, output_sidecar, input_file)
             return
 
@@ -324,8 +330,8 @@ def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
     try:
         log.debug(args_tesseract)
         stdout = check_output(
-            args_tesseract, close_fds=True, stderr=STDOUT,
-            universal_newlines=True, timeout=timeout)
+            args_tesseract, stderr=STDOUT,
+            timeout=timeout)
         if os.path.exists(prefix + '.txt'):
             shutil.move(prefix + '.txt', output_text)
     except TimeoutExpired:
@@ -333,10 +339,10 @@ def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
         use_skip_page(text_only, skip_pdf, output_pdf, output_text)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_image)
-        if 'read_params_file: parameter not found' in e.output:
+        if b'read_params_file: parameter not found' in e.output:
             raise TesseractConfigError() from e
 
-        if 'Image too large' in e.output:
+        if b'Image too large' in e.output:
             use_skip_page(text_only, skip_pdf, output_pdf, output_text)
             return
         raise e from e

ocrmypdf/hocrtransform.py

@@ -1,12 +1,33 @@
 #!/usr/bin/env python3
-##############################################################################
-# Copyright (c) 2013-14: fritz-hh from Github
-#   (https://github.com/fritz-hh)
 #
-# Copyright (c) 2010: Jonathan Brinley from Github
-#   (https://github.com/jbrinley/HocrConverter)
-#   Initial version by Jonathan Brinley, jonathanbrinley@gmail.com
-##############################################################################
+# Copyright (c) 2010, Jonathan Brinley
+#   Original version from: https://github.com/jbrinley/HocrConverter
+#
+# Copyright (c) 2013-14, Julien Pfefferkorn
+#   Modifications
+#
+# Copyright (c) 2015-16, James R. Barlow
+#   Set text to transparent
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+# 
+# The above copyright notice and this permission notice shall be included
+# in all copies or substantial portions of the Software.
+# 
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+# OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+# CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
 from reportlab.pdfgen.canvas import Canvas
 from reportlab.lib.units import inch
 from xml.etree import ElementTree

ocrmypdf/pipeline.py

@@ -6,7 +6,7 @@ import sys
 import os
 import shutil
 import img2pdf
-
+import re
 import PyPDF2 as pypdf
 from PIL import Image
 
@@ -149,6 +149,22 @@ def triage_image_file(input_file, output_file, log, options):
         raise UnsupportedImageFormatError() from e
 
 
+def _pdf_guess_version(input_file, search_window=1024):
+    """Try to find version signature at start of file.
+    
+    Not robust enough to deal with appended files.
+
+    Returns empty string if not found, indicating file is probably not PDF.
+    """
+
+    with open(input_file, 'rb') as f:
+        signature = f.read(1024)
+    m = re.search(b'%PDF-(\d\.\d)', signature)
+    if m:
+        return m.group(1)
+    return ''
+
+
 def triage(
         input_file,
         output_file,
@@ -157,14 +173,12 @@ def triage(
 
     options = context.get_options()
     try:
-        with open(input_file, 'rb') as f:
-            signature = f.read(1024)
-            if b'%PDF' in signature:
-                if options.image_dpi:
-                    log.warning("Argument --image-dpi ignored because the "
-                                "input file is a PDF, not an image.")
-                re_symlink(input_file, output_file, log)
-                return
+        if _pdf_guess_version(input_file):
+            if options.image_dpi:
+                log.warning("Argument --image-dpi ignored because the "
+                            "input file is a PDF, not an image.")
+            re_symlink(input_file, output_file, log)
+            return
     except EnvironmentError as e:
         log.error(e)
         raise InputFileError() from e
@@ -914,6 +928,7 @@ def merge_pages_qpdf(
     first_page = pypdf.PdfFileReader(pdf_pages[0])
 
     writer = pypdf.PdfFileWriter()
+    writer._header = b'%PDF-' + _pdf_guess_version(pdf_pages[0])  # copy version from source
     writer.appendPagesFromReader(first_page)
     writer.addMetadata(pdfmark)
     writer_file = pdf_pages[0].replace('.pdf', '.metadata.pdf')
@@ -923,7 +938,8 @@ def merge_pages_qpdf(
     pdf_pages[0] = writer_file
 
     qpdf.merge(input_files=pdf_pages, output_file=output_file,
-               min_version=context.get_pdfinfo().min_version)
+               min_version=context.get_pdfinfo().min_version,
+               log=log)
 
 
 def merge_sidecars(

setup.py

@@ -220,7 +220,8 @@ setup(
     python_requires='>=3.5',
     setup_requires=[
         'setuptools_scm',       # so that version will work
-        'cffi>=1.9.1'           # to build the leptonica module
+        'cffi>=1.9.1',          # to build the leptonica module
+        'pytest-runner'         # to enable python setup.py test
     ],
     use_scm_version={'version_scheme': 'post-release'},
     cffi_modules=[

test_requirements.txt

@@ -1,4 +1,5 @@
-pytest >= 3.0
+pytest >= 3.2
 pytest-helpers-namespace
 pytest-xdist
-pytest-cov
+pytest-cov
+pytest-timeout

tests/resources/README.rst

@@ -61,6 +61,9 @@ under the terms of the license in LICENSE.rst.
     *   - cmyk.pdf
         - @jbarlow83
         - a CMYK image created in Photoshop
+    *   - crom.png
+        - @jbarlow83
+        - test for non-dictionary words        
     *   - enormous.pdf
         - @jbarlow83
         - very large PDF page
@@ -69,7 +72,7 @@ under the terms of the license in LICENSE.rst.
         - a linearized PDF containing some unusual indirect objects, created by an Epson printer; printout of a Wikipedia article (CC BY-SA)
     *   - formxobject.pdf
         - @jbarlow83
-        - hand-crafted exotic PDF containing an image inside a Form XObject
+        - hand-crafted PDF containing an image inside a Form XObject
     *   - francais.pdf
         - @jbarlow83
         - a page containing French accents (diacritics)  
@@ -88,12 +91,15 @@ under the terms of the license in LICENSE.rst.
     *   - overlay.pdf
         - @maxandersen
         - PDF file generated by PDFPen pro that triggered content stream parse errors
+    *   - negzero.pdf
+        - @jbarlow83
+        - copy of formxobject.pdf with token that qpdf doesn't like
     *   - no_contents.pdf
         - @jbarlow83
         - synthetic PDF with a blank page that has no /Contents entry
-    *   - crom.png
+    *   - trivial.pdf
         - @jbarlow83
-        - test for non-dictionary words
+        - smallest possible valid PDF-1.3 with all required fields
 
 
 Assemblies

tests/resources/trivial.pdf

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

tests/spoof/qpdf_dummy_return2.py

@@ -5,6 +5,9 @@ import sys
 
 
 def main():
+    if sys.argv[-1] == '--version':
+        print('qpdf version 7.0.0')
+        sys.exit(0)
     print('qpdf dummy')
     sys.exit(2)
 

tests/spoof/tesseract_badutf8.py

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

tests/test_main.py

@@ -1,16 +1,16 @@
-#!/usr/bin/env python3
-# © 2015 James R. Barlow: github.com/jbarlow83
+# © 2015-17 James R. Barlow: github.com/jbarlow83
 
 from subprocess import Popen, PIPE, check_output, check_call, DEVNULL
 import os
 import shutil
+import resource
 import pytest
 from ocrmypdf.pdfinfo import PdfInfo, Colorspace, Encoding
 import PyPDF2 as pypdf
 from ocrmypdf.exceptions import ExitCode
 from ocrmypdf import leptonica
 from ocrmypdf.pdfa import file_claims_pdfa
-from ocrmypdf.exec import ghostscript, tesseract
+from ocrmypdf.exec import ghostscript, tesseract, qpdf
 import logging
 from math import isclose
 
@@ -55,6 +55,11 @@ def spoof_no_tess_gs_raster_fail():
     return spoof(tesseract='tesseract_noop.py', gs='gs_raster_failure.py')
 
 
+@pytest.fixture
+def spoof_tess_bad_utf8():
+    return spoof(tesseract='tesseract_badutf8.py')
+
+
 @pytest.fixture
 def spoof_qpdf_always_error():
     return spoof(qpdf='qpdf_dummy_return2.py')
@@ -64,6 +69,7 @@ def test_quick(spoof_tesseract_cache, resources, outpdf):
     check_ocrmypdf(resources / 'ccitt.pdf', outpdf, env=spoof_tesseract_cache)
 
 
+@pytest.mark.filterwarnings('ignore:Image size')
 def test_deskew(spoof_tesseract_noop, resources, outdir):
     # Run with deskew
     deskewed_pdf = check_ocrmypdf(
@@ -1016,4 +1022,18 @@ def test_bad_locale():
     )
     assert out == '', "stdout not clean"
     assert p.returncode != 0
-    assert 'configured to use ASCII as encoding' in err, "should whine"
+    assert 'configured to use ASCII as encoding' in err, "should whine"
+
+
+@pytest.mark.parametrize('renderer', RENDERERS)
+def test_bad_utf8(spoof_tess_bad_utf8, renderer, resources, no_outpdf):
+    p, out, err = run_ocrmypdf(
+        resources / 'ccitt.pdf', no_outpdf,
+        '--pdf-renderer', renderer,
+        env=spoof_tess_bad_utf8
+    )
+
+    assert out == '', "stdout not clean"
+    assert p.returncode != 0
+    assert 'not utf-8' in err, "should whine about utf-8"
+    assert '\\x96' in err, 'should repeat backslash encoded output'

tests/test_qpdf.py

@@ -0,0 +1,55 @@
+# © 2017 James R. Barlow: github.com/jbarlow83
+
+import logging
+import resource
+import pytest
+
+from ocrmypdf.exec import ghostscript, tesseract, qpdf
+from ocrmypdf.pdfinfo import PdfInfo
+
+
+@pytest.mark.skipif(
+    qpdf.version() < '7.0.0',
+    reason="negzero.pdf crashes earlier versions")
+def test_qpdf_negative_zero(resources, outpdf):
+    negzero = resources / 'negzero.pdf'
+    hugemono = resources / 'hugemono.pdf'
+    # raises exception on err
+    qpdf.merge([str(negzero), str(hugemono)], outpdf, log=logging.getLogger())
+    
+
+@pytest.mark.timeout(15)
+@pytest.mark.parametrize('max_files,skip', [
+    (2, 0),  # Can we merge correctly without opening more than 2 files at once?
+    (16, 0), # And does this work properly when we can one-shot it?
+    (2, 1),  # Or playing with even/odd
+    (3, 0)   # Or odd step size
+    ])
+def test_qpdf_merge_correctness(resources, outpdf, max_files, skip):
+    # All of these must be only one page long
+    inputs = [
+        '2400dpi.pdf', 'aspect.pdf', 'blank.pdf', 'ccitt.pdf', 
+        'linn.pdf', 'masks.pdf', 'poster.pdf', 'overlay.pdf',
+        'skew.pdf', 'trivial.pdf']
+    
+    input_files = [str(resources / f) for f in inputs]
+
+    qpdf.merge(
+        input_files[skip:], outpdf, log=logging.getLogger(), 
+        max_files=max_files)
+    assert len(PdfInfo(outpdf).pages) == len(input_files[skip:])
+    
+
+@pytest.mark.timeout(15)
+@pytest.mark.skipif(
+    True, 
+    reason='qpdf binary cannot open multiple files multiple times')
+def test_page_merge_ulimit(resources, outpdf):
+    # Ensure we can merge pages without opening one file descriptor per page
+    ulimits = resource.getrlimit(resource.RLIMIT_NOFILE)
+    page_count = ulimits[0]
+    print(page_count)
+    input_files = [str(resources / 'trivial.pdf')] * page_count
+
+    qpdf.merge(input_files, outpdf, log=logging.getLogger())
+