menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge api (without plugins)

Browse Source
This commit is contained in:
James R. Barlow
2019-07-27 02:04:01 -07:00
parent 423324af01 f83de20c37
commit 6fbeb6347d
249 changed files with 12915 additions and 11491 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.coveragerc
+12 -3
View File
@@ -1,9 +1,18 @@
# Coverage isn't really compatible with subprocesses so results are unreliable
[paths]
source =
src
*/site-packages
[run]
branch = True
#concurrency = multiprocessing
source = ocrmypdf/
branch = true
parallel = true
source =
src/ocrmypdf
tests
omit =
tests/spoof/*
[report]
exclude_lines =
.docker/Dockerfile
+56 -48
View File
@@ -1,17 +1,61 @@
# OCRmyPDF
#
FROM ubuntu:18.04
FROM ubuntu:19.04 as base
FROM base as builder
ENV LANG=C.UTF-8
RUN apt-get update && apt-get install -y --no-install-recommends \
build-essential autoconf automake libtool \
libleptonica-dev \
zlib1g-dev \
libexempi3 \
ocrmypdf \
pngquant \
python3-pip \
python3-venv \
tesseract-ocr \
unpaper \
wget \
git
# Compile and install jbig2
# Needs libleptonica-dev, zlib1g-dev
RUN \
mkdir jbig2 \
&& wget -q https://github.com/agl/jbig2enc/archive/0.29.tar.gz -O - | \
tar xz -C jbig2 --strip-components=1 \
&& cd jbig2 \
&& ./autogen.sh && ./configure && make && make install \
&& cd .. \
&& rm -rf jbig2
RUN python3 -m venv /appenv
COPY . /app
WORKDIR /app
RUN . /appenv/bin/activate; \
pip install --upgrade pip \
&& pip install .
FROM base
ENV LANG=C.UTF-8
RUN apt-get update && apt-get install -y --no-install-recommends \
ghostscript \
img2pdf \
liblept5 \
libsm6 libxext6 libxrender-dev \
zlib1g \
pngquant \
python3 \
python3-venv \
qpdf \
tesseract-ocr \
tesseract-ocr-chi-sim \
tesseract-ocr-deu \
tesseract-ocr-eng \
@@ -21,52 +65,16 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
unpaper \
wget
# Copy
COPY --from=builder /app/misc/webservice.py /app/
ENV LANG=C.UTF-8
# Copy minimal project files to get the test suite.
COPY --from=builder /app/setup.cfg /app/setup.py /app/README.md /app/
COPY --from=builder /app/requirements /app/requirements
COPY --from=builder /app/tests /app/tests
COPY --from=builder /app/src /app/src
# Compile and install jbig2
# Needs libleptonica-dev, zlib1g-dev
RUN \
mkdir jbig2 \
&& wget -q https://github.com/agl/jbig2enc/archive/0.29.tar.gz -O - | \
tar xz -C jbig2 --strip-components=1 \
&& cd jbig2 \
&& ./autogen.sh && ./configure && make && make install \
&& cd .. \
&& rm -rf jbig2
COPY --from=builder /appenv /appenv
COPY --from=builder /usr/local /usr/local
RUN apt-get remove -y autoconf automake libtool
RUN python3 -m venv --system-site-packages /appenv
# This installs the latest binary wheel instead of the code in the current
# folder. Installing from source will fail, apparently because cffi needs
# build-essentials (gcc) to do a source installation
# (i.e. "pip install ."). It's unclear to me why this is the case.
RUN . /appenv/bin/activate; \
pip install --upgrade pip \
&& pip install --upgrade ocrmypdf
# Now copy the application in, mainly to get the test suite.
# Do this now to make the best use of Docker cache.
COPY . /application
RUN . /appenv/bin/activate; \
pip install -r /application/requirements/test.txt
# Remove the junk, including the source version of application since it was
# already installed
RUN rm -rf /tmp/* /var/tmp/* /root/* /application/ocrmypdf \
&& apt-get remove -y build-essential \
&& apt-get autoremove -y \
&& apt-get autoclean -y
RUN useradd docker \
&& mkdir /home/docker \
&& chown docker:docker /home/docker
USER docker
WORKDIR /home/docker
# Must use array form of ENTRYPOINT
# Non-array form does not append other arguments, because that is "intuitive"
ENTRYPOINT ["/application/.docker/docker-wrapper.sh"]
ENTRYPOINT ["/appenv/bin/ocrmypdf"]
.docker/alpine.dockerfile
+46 -39
View File
@@ -4,33 +4,38 @@ FROM base as builder
ENV LANG=C.UTF-8
# Normally:
# echo '@testing http://nl.alpinelinux.org/alpine/edge/testing' >> /etc/apk/repositories
RUN \
echo '@testing http://nl.alpinelinux.org/alpine/edge/testing' >> /etc/apk/repositories \
echo -e '@testing http://nl.alpinelinux.org/alpine/edge/testing\n@community http://nl.alpinelinux.org/alpine/edge/community'\
>> /etc/apk/repositories \
# Add runtime dependencies
&& apk add --update \
python3-dev \
py3-setuptools \
jbig2enc@testing \
ghostscript \
qpdf \
tesseract-ocr \
unpaper \
pngquant \
libxml2-dev \
libxslt-dev \
zlib-dev \
qpdf-dev \
libffi-dev \
leptonica-dev \
binutils \
python3-dev \
py3-setuptools \
jbig2enc@testing \
ghostscript \
qpdf@community \
qpdf-dev@community \
tesseract-ocr \
unpaper \
pngquant \
libxml2-dev \
libxslt-dev \
zlib-dev \
libffi-dev \
leptonica-dev \
binutils \
&& pip3 install --upgrade pip \
# Install pybind11 for pikepdf
&& pip3 install pybind11 \
# Install flask for the webservice
&& pip3 install flask \
# Add build dependencies
&& apk add --virtual build-dependencies \
build-base \
git
build-base \
git
COPY . /app
@@ -42,43 +47,45 @@ FROM base
ENV LANG=C.UTF-8
# Normally:
# echo '@testing http://nl.alpinelinux.org/alpine/edge/testing' >> /etc/apk/repositories
RUN \
echo '@testing http://nl.alpinelinux.org/alpine/edge/testing' >> /etc/apk/repositories \
echo -e '@testing http://nl.alpinelinux.org/alpine/edge/testing\n@community http://nl.alpinelinux.org/alpine/edge/community'\
>> /etc/apk/repositories \
# Add runtime dependencies
&& apk add --update \
python3 \
jbig2enc@testing \
ghostscript \
qpdf \
tesseract-ocr \
tesseract-ocr-data-deu \
tesseract-ocr-data-chi_sim \
unpaper \
pngquant \
libxml2 \
libxslt \
zlib \
qpdf \
libffi \
leptonica-dev \
binutils \
python3 \
jbig2enc@testing \
ghostscript \
qpdf@community \
qpdf-dev@community \
tesseract-ocr \
tesseract-ocr-data-deu \
tesseract-ocr-data-chi_sim \
unpaper \
pngquant \
libxml2 \
libxslt \
zlib \
libffi \
leptonica-dev \
binutils \
&& mkdir /app
WORKDIR /app
# Copy build artifacts (python site-packages9
# Copy build artifacts (python site-packages)
COPY --from=builder /usr/lib/python3.6/site-packages /usr/lib/python3.6/site-packages
COPY --from=builder /usr/bin/ocrmypdf /usr/bin/dumppdf.py /usr/bin/latin2ascii.py /usr/bin/pdf2txt.py /usr/bin/img2pdf /usr/bin/chardetect /usr/bin/
# Copy
COPY --from=builder /app/.docker/webservice.py /app/
COPY --from=builder /app/misc/webservice.py /app/
# Copy minimal project files to get the test suite.
COPY --from=builder /app/setup.cfg /app/setup.py /app/README.md /app/
COPY --from=builder /app/requirements /app/requirements
COPY --from=builder /app/tests /app/tests
COPY --from=builder /app/src /app/src
# Copy PKG-INFO from build artifact in app dir to make setuptools-scm happy
RUN cp /usr/lib/python3.6/site-packages/ocrmypdf-*.egg-info/PKG-INFO /app
ENTRYPOINT ["/usr/bin/ocrmypdf"]
.docker/docker-wrapper.sh
-5
View File
@@ -1,5 +0,0 @@
#!/bin/bash
. /appenv/bin/activate
cd /home/docker
exec ocrmypdf "$@"
.docker/polyglot.dockerfile
-17
View File
@@ -1,17 +0,0 @@
# OCRmyPDF polyglot
#
FROM jbarlow83/ocrmypdf:latest
USER root
# Update system and install our dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
tesseract-ocr-all
RUN apt-get autoremove -y && apt-get clean -y
USER docker
# Must use array form of ENTRYPOINT
# Non-array form does not append other arguments, because that is "intuitive"
ENTRYPOINT ["/application/.docker/docker-wrapper.sh"]
.docker/webservice.dockerfile
-24
View File
@@ -1,24 +0,0 @@
# OCRmyPDF webservice
#
FROM jbarlow83/ocrmypdf-polyglot:latest
USER root
# Update system and install our dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
python3-flask
RUN apt-get autoremove -y && apt-get clean -y
EXPOSE 5000
COPY .docker/webservice.py /application
USER docker
VOLUME ["/config"]
# This config file is optional
ENV OCRMYPDF_WEBSERVICE_SETTINGS "/config/config.py"
ENTRYPOINT ["python3", "/application/webservice.py"]
.dockerignore
-1
View File
@@ -6,7 +6,6 @@
**/*.pyc
.*/
!.git/
!.docker/
.ruffus_history.sqlite
bin/
build/
README.md
+1 -2
View File
@@ -1,5 +1,4 @@
OCRmyPDF
========
<img src="docs/images/logo.svg" width="240" alt="OCRmyPDF">
[![Travis build status][travis]](https://travis-ci.org/jbarlow83/OCRmyPDF) [![PyPI version][pypi]](https://pypi.org/project/ocrmypdf/) ![Homebrew version][homebrew] ![ReadTheDocs][docs]
docs/advanced.rst
+162 -65
View File
@@ -1,16 +1,34 @@
=================
Advanced features
=================
Control of unpaper
------------------
==================
OCRmyPDF uses ``unpaper`` to provide the implementation of the ``--clean`` and ``--clean-final`` arguments. `unpaper <https://github.com/Flameeyes/unpaper/blob/master/doc/basic-concepts.md>`_ provides a variety of image processing filters to improve images.
OCRmyPDF uses ``unpaper`` to provide the implementation of the
``--clean`` and ``--clean-final`` arguments.
`unpaper <https://github.com/Flameeyes/unpaper/blob/master/doc/basic-concepts.md>`__
provides a variety of image processing filters to improve images.
By default, OCRmyPDF uses only ``unpaper`` arguments that were found to be safe to use on almost all files without having to inspect every page of the file afterwards. This is particularly true when only ``--clean`` is used, since that instructs OCRmyPDF to only clean the image before OCR and not the final image.
By default, OCRmyPDF uses only ``unpaper`` arguments that were found to
be safe to use on almost all files without having to inspect every page
of the file afterwards. This is particularly true when only ``--clean``
is used, since that instructs OCRmyPDF to only clean the image before
OCR and not the final image.
However, if you wish to use the more aggressive options in ``unpaper``, you may use ``--unpaper-args '...'`` to override the OCRmyPDF's defaults and forward other arguments to unpaper. This option will forward arguments to ``unpaper`` without any knowledge of what that program considers to be valid arguments. The string of arguments must be quoted as shown in the examples below. No filename arguments may be included. OCRmyPDF will assume it can append input and output filename of intermediate images to the ``--unpaper-args`` string.
However, if you wish to use the more aggressive options in ``unpaper``,
you may use ``--unpaper-args '...'`` to override the OCRmyPDF's defaults
and forward other arguments to unpaper. This option will forward
arguments to ``unpaper`` without any knowledge of what that program
considers to be valid arguments. The string of arguments must be quoted
as shown in the examples below. No filename arguments may be included.
OCRmyPDF will assume it can append input and output filename of
intermediate images to the ``--unpaper-args`` string.
In this example, we tell ``unpaper`` to expect two pages of text on a sheet (image), such as occurs when two facing pages of a book are scanned. ``unpaper`` uses this information to deskew each independently and clean up the margins of both.
In this example, we tell ``unpaper`` to expect two pages of text on a
sheet (image), such as occurs when two facing pages of a book are
scanned. ``unpaper`` uses this information to deskew each independently
and clean up the margins of both.
.. code-block:: bash
@@ -19,40 +37,71 @@ In this example, we tell ``unpaper`` to expect two pages of text on a sheet (ima
.. warning::
Some ``unpaper`` features will reposition text within the image. ``--clean-final`` is recommended to avoid this issue.
Some ``unpaper`` features will reposition text within the image.
``--clean-final`` is recommended to avoid this issue.
.. warning::
Some ``unpaper`` features cause multiple input or output files to be consumed or produced. OCRmyPDF requires ``unpaper`` to consume one file and produce one file. An deviation from that condition will result in errors.
Some ``unpaper`` features cause multiple input or output files to be
consumed or produced. OCRmyPDF requires ``unpaper`` to consume one
file and produce one file. An deviation from that condition will
result in errors.
.. note::
``unpaper`` uses uncompressed PBM/PGM/PPM files for its intermediate files. For large images or documents, it can take a lot of temporary disk space.
``unpaper`` uses uncompressed PBM/PGM/PPM files for its intermediate
files. For large images or documents, it can take a lot of temporary
disk space.
Control of OCR options
----------------------
======================
OCRmyPDF provides many features to control the behavior of the OCR engine, Tesseract.
OCRmyPDF provides many features to control the behavior of the OCR
engine, Tesseract.
When OCR is skipped
"""""""""""""""""""
-------------------
If a page in a PDF seems to have text, by default OCRmyPDF will exit without modifying the PDF. This is to ensure that PDFs that were previously OCRed or were "born digital" rather than scanned are not processed.
If a page in a PDF seems to have text, by default OCRmyPDF will exit
without modifying the PDF. This is to ensure that PDFs that were
previously OCRed or were "born digital" rather than scanned are not
processed.
If ``--skip-text`` is issued, then no OCR will be performed on pages that already have text. The page will be copied to the output. This may be useful for documents that contain both "born digital" and scanned content, or to use OCRmyPDF to normalize and convert to PDF/A regardless of their contents.
If ``--skip-text`` is issued, then no OCR will be performed on pages
that already have text. The page will be copied to the output. This may
be useful for documents that contain both "born digital" and scanned
content, or to use OCRmyPDF to normalize and convert to PDF/A regardless
of their contents.
If ``--redo-ocr`` is issued, then a detailed text analysis is performed. Text is categorized as either visible or invisible. Invisible text (OCR) is stripped out. Then an image of each page is created with visible text masked out. The page image is sent for OCR, and any additional text is inserted as OCR. If a file contains a mix of text and bitmap images that contain text, OCRmyPDF will locate the additional text in images without disrupting the existing text.
If ``--redo-ocr`` is issued, then a detailed text analysis is performed.
Text is categorized as either visible or invisible. Invisible text (OCR)
is stripped out. Then an image of each page is created with visible text
masked out. The page image is sent for OCR, and any additional text is
inserted as OCR. If a file contains a mix of text and bitmap images that
contain text, OCRmyPDF will locate the additional text in images without
disrupting the existing text.
If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information. Any forms and vector graphics will be rasterized as well.
If ``--force-ocr`` is issued, then all pages will be rasterized to
images, discarding any hidden OCR text, and rasterizing any printable
text. This is useful for redoing OCR, for fixing OCR text with a damaged
character map (text is selectable but not searchable), and destroying
redacted information. Any forms and vector graphics will be rasterized
as well.
Time and image size limits
""""""""""""""""""""""""""
--------------------------
By default, OCRmyPDF permits tesseract to run for three minutes (180 seconds) per page. This is usually more than enough time to find all text on a reasonably sized page with modern hardware.
By default, OCRmyPDF permits tesseract to run for three minutes (180
seconds) per page. This is usually more than enough time to find all
text on a reasonably sized page with modern hardware.
If a page is skipped, it will be inserted without OCR. If preprocessing was requested, the preprocessed image layer will be inserted.
If a page is skipped, it will be inserted without OCR. If preprocessing
was requested, the preprocessed image layer will be inserted.
If you want to adjust the amount of time spent on OCR, change ``--tesseract-timeout``. You can also automatically skip images that exceed a certain number of megapixels with ``--skip-big``. (A 300 DPI, 8.5×11" page is 8.4 megapixels.)
If you want to adjust the amount of time spent on OCR, change
``--tesseract-timeout``. You can also automatically skip images that
exceed a certain number of megapixels with ``--skip-big``. (A 300 DPI,
8.5×11" page is 8.4 megapixels.)
.. code-block:: bash
@@ -60,21 +109,27 @@ If you want to adjust the amount of time spent on OCR, change ``--tesseract-time
ocrmypdf --tesseract-timeout 300 --skip-big 50 bigfile.pdf output.pdf
Overriding default tesseract
""""""""""""""""""""""""""""
----------------------------
OCRmyPDF checks the system ``PATH`` for the ``tesseract`` binary.
Some relevant environment variables that influence Tesseract's behavior include:
Some relevant environment variables that influence Tesseract's behavior
include:
.. envvar:: TESSDATA_PREFIX
Overrides the path to Tesseract's data files. This can allow simultaneous installation of the "best" and "fast" training data sets. OCRmyPDF does not manage this environment variable.
Overrides the path to Tesseract's data files. This can allow
simultaneous installation of the "best" and "fast" training data
sets. OCRmyPDF does not manage this environment variable.
.. envvar:: OMP_THREAD_LIMIT
Controls the number of threads Tesseract will use. OCRmyPDF will manage this environment if it is not already set. (Currently, it will set it to 1 because this gives the best results in testing.)
Controls the number of threads Tesseract will use. OCRmyPDF will
manage this environment if it is not already set. (Currently, it will
set it to 1 because this gives the best results in testing.)
For example, if you have a development build of Tesseract don't wish to use the system installation, you can launch OCRmyPDF as follows:
For example, if you have a development build of Tesseract don't wish to
use the system installation, you can launch OCRmyPDF as follows:
.. code-block:: bash
@@ -83,26 +138,33 @@ For example, if you have a development build of Tesseract don't wish to use the
TESSDATA_PREFIX=/home/user/src/tesseract \
ocrmypdf input.pdf output.pdf
In this example ``TESSDATA_PREFIX`` is required to redirect Tesseract to an alternate folder for its "tessdata" files.
In this example ``TESSDATA_PREFIX`` is required to redirect Tesseract to
an alternate folder for its "tessdata" files.
Overriding other support programs
"""""""""""""""""""""""""""""""""
---------------------------------
In addition to tesseract, OCRmyPDF uses the following external binaries:
* ``gs`` (Ghostscript)
* ``unpaper``
* ``qpdf``
In each case OCRmyPDF will search the ``PATH`` environment variable to locate the binaries.
- ``gs`` (Ghostscript)
- ``unpaper``
- ``qpdf``
In each case OCRmyPDF will search the ``PATH`` environment variable to
locate the binaries.
Changing tesseract configuration variables
""""""""""""""""""""""""""""""""""""""""""
------------------------------------------
You can override tesseract's default `control parameters <https://github.com/tesseract-ocr/tesseract/wiki/ControlParams>`_ with a configuration file.
You can override tesseract's default `control
parameters <https://github.com/tesseract-ocr/tesseract/wiki/ControlParams>`__
with a configuration file.
As an example, this configuration will disable Tesseract's dictionary for current language. Normally the dictionary is helpful for interpolating words that are unclear, but it may interfere with OCR if the document does not contain many words (for example, a list of part numbers).
As an example, this configuration will disable Tesseract's dictionary
for current language. Normally the dictionary is helpful for
interpolating words that are unclear, but it may interfere with OCR if
the document does not contain many words (for example, a list of part
numbers).
Create a file named "no-dict.cfg" with these contents:
@@ -120,11 +182,11 @@ then run ocrmypdf as follows (along with any other desired arguments):
.. warning::
Some combinations of control parameters will break Tesseract or break assumptions that OCRmyPDF makes about Tesseract's output.
Some combinations of control parameters will break Tesseract or break
assumptions that OCRmyPDF makes about Tesseract's output.
Changing the PDF renderer
-------------------------
=========================
rasterizing
Converting a PDF to an image for display.
@@ -132,42 +194,63 @@ rasterizing
rendering
Creating a new PDF from other data (such as an existing PDF).
OCRmyPDF has these PDF renderers: ``sandwich`` and ``hocr``. The renderer may be selected using ``--pdf-renderer``. The default is ``auto`` which lets OCRmyPDF select the renderer to use. Currently, ``auto`` always selects ``sandwich``.
OCRmyPDF has these PDF renderers: ``sandwich`` and ``hocr``. The
renderer may be selected using ``--pdf-renderer``. The default is
``auto`` which lets OCRmyPDF select the renderer to use. Currently,
``auto`` always selects ``sandwich``.
The ``sandwich`` renderer
"""""""""""""""""""""""""
-------------------------
The ``sandwich`` renderer uses Tesseract's new text-only PDF feature, which produces a PDF page that lays out the OCR in invisible text. This page is then "sandwiched" onto the original PDF page, allowing lossless application of OCR even to PDF pages that contain other vector objects.
The ``sandwich`` renderer uses Tesseract's new text-only PDF feature,
which produces a PDF page that lays out the OCR in invisible text. This
page is then "sandwiched" onto the original PDF page, allowing lossless
application of OCR even to PDF pages that contain other vector objects.
Currently this is the best renderer for most uses, however it is implemented in Tesseract so OCRmyPDF cannot influence it. Currently some problematic PDF viewers like Mozilla PDF.js and macOS Preview have problems with segmenting its text output, and mightrunseveralwordstogether.
Currently this is the best renderer for most uses, however it is
implemented in Tesseract so OCRmyPDF cannot influence it. Currently some
problematic PDF viewers like Mozilla PDF.js and macOS Preview have
problems with segmenting its text output, and
mightrunseveralwordstogether.
When image preprocessing features like ``--deskew`` are used, the original PDF will be rendered as a full page and the OCR layer will be placed on top.
When image preprocessing features like ``--deskew`` are used, the
original PDF will be rendered as a full page and the OCR layer will be
placed on top.
The ``hocr`` renderer
"""""""""""""""""""""
---------------------
The ``hocr`` renderer works with older versions of Tesseract. The image layer is copied from the original PDF page if possible, avoiding potentially lossy transcoding or loss of other PDF information. If preprocessing is specified, then the image layer is a new PDF.
The ``hocr`` renderer works with older versions of Tesseract. The image
layer is copied from the original PDF page if possible, avoiding
potentially lossy transcoding or loss of other PDF information. If
preprocessing is specified, then the image layer is a new PDF.
Unlike ``sandwich`` this renderer is implemented within OCRmyPDF; anyone looking to customize how OCR is presented should look here. A major disadvantage of this renderer is it not capable of correctly handling text outside the Latin alphabet. Pull requests to improve the situation are welcome.
Unlike ``sandwich`` this renderer is implemented within OCRmyPDF; anyone
looking to customize how OCR is presented should look here. A major
disadvantage of this renderer is it not capable of correctly handling
text outside the Latin alphabet. Pull requests to improve the situation
are welcome.
Currently, this renderer has the best compatibility with Mozilla's PDF.js viewer.
Currently, this renderer has the best compatibility with Mozilla's
PDF.js viewer.
This works in all versions of Tesseract.
The ``tesseract`` renderer
""""""""""""""""""""""""""
--------------------------
The ``tesseract`` renderer was removed. OCRmyPDF's new approach to text layer grafting makes it functionally equivalent to ``sandwich``.
The ``tesseract`` renderer was removed. OCRmyPDF's new approach to text
layer grafting makes it functionally equivalent to ``sandwich``.
Return code policy
------------------
==================
OCRmyPDF writes all messages to ``stderr``. ``stdout`` is reserved for piping
output files. ``stdin`` is reserved for piping input files.
OCRmyPDF writes all messages to ``stderr``. ``stdout`` is reserved for
piping output files. ``stdin`` is reserved for piping input files.
The return codes generated by the OCRmyPDF are considered part of the stable
user interface. They may be imported from ``ocrmypdf.exceptions``.
The return codes generated by the OCRmyPDF are considered part of the
stable user interface. They may be imported from
``ocrmypdf.exceptions``.
.. list-table:: Return codes
:widths: 5 35 60
@@ -218,22 +301,36 @@ user interface. They may be imported from ``ocrmypdf.exceptions``.
Debugging the intermediate files
--------------------------------
================================
OCRmyPDF normally saves its intermediate results to a temporary folder and deletes this folder when it exits, whether it succeeded or failed.
OCRmyPDF normally saves its intermediate results to a temporary folder
and deletes this folder when it exits, whether it succeeded or failed.
If the ``-k`` argument is issued on the command line, OCRmyPDF will keep the temporary folder and print the location, whether it succeeded or failed (provided the Python interpreter did not crash). An example message is:
If the ``-k`` argument is issued on the command line, OCRmyPDF will keep
the temporary folder and print the location, whether it succeeded or
failed (provided the Python interpreter did not crash). An example
message is:
.. code-block:: none
Temporary working files saved at:
Temporary working files retained at:
/tmp/com.github.ocrmypdf.u20wpz07
The organization of this folder is an implementation detail and subject to change between releases. However the general organization is that working files on a per page basis have the page number as a prefix (starting with page 1), an infix indicates the processing stage, and a suffix indicates the file type. Some important files include:
The organization of this folder is an implementation detail and subject
to change between releases. However the general organization is that
working files on a per page basis have the page number as a prefix
(starting with page 1), an infix indicates the processing stage, and a
suffix indicates the file type. Some important files include:
* ``.page.png`` - what the input page looks like
* ``.image`` - the image we will show the user if we are in a mode that changes the final appearance; may be in one of several image formats
* ``.text.pdf`` - the OCR file; this will load as a blank page but should have visible text if checked with a tool like pdftotext or pdfminder.six
* ``.ocr.png`` - the file that is sent to Tesseract for OCR; depending on arguments this may differ from the presentation image
* ``layers.rendered.pdf`` - the composite PDF, before metadata repair and optimization
* ``images/*`` - images extracted during the optimization process; here the prefix indicates a PDF object ID not a page number
- ``.page.png`` - what the input page looks like
- ``.image`` - the image we will show the user if we are in a mode that
changes the final appearance; may be in one of several image formats
- ``.text.pdf`` - the OCR file; this will load as a blank page but
should have visible text if checked with a tool like pdftotext or
pdfminder.six
- ``.ocr.png`` - the file that is sent to Tesseract for OCR; depending
on arguments this may differ from the presentation image
- ``layers.rendered.pdf`` - the composite PDF, before metadata repair
and optimization
- ``images/*`` - images extracted during the optimization process; here
the prefix indicates a PDF object ID not a page number
docs/api.rst
+101
View File
@@ -0,0 +1,101 @@
======================
Using the OCRmyPDF API
======================
OCRmyPDF originated as a command line program and continues to have this
legacy, but parts of it can be imported and used in other Python
applications.
Some applications may want to consider running ocrmypdf from a
subprocess call anyway, as this provides isolation of its activities.
Example
=======
OCRmyPDF one high-level function to run its main engine from an
application. The parameters are symmetric to the command line arguments
and largely have the same functions.
.. code-block:: python
import ocrmypdf
ocrmypdf.ocr('input.pdf', 'output.pdf', deskew=True)
With a few exceptions, all of the command line arguments are available
and may be passed as equivalent keywords.
A few differences are that ``verbose`` and ``quiet`` are not available.
Instead, output should be managed by configuring logging.
Parent process requirements
---------------------------
The :func:`ocrmypdf.run` function runs OCRmyPDF similar to command line
execution. To do this, it will: - create a monitoring thread - create
worker processes (forking itself) - manage the signal flags of worker
processes 0 execute other subprocesses (forking and executing other
programs)
The Python process that calls ``ocrmypdf.ocr()`` must be sufficiently
privileged to perform these actions. If it is not, ``ocrmypdf()`` will
fail.
There is no currently no option to manage how jobs are scheduled other
than the argument ``jobs=`` which will limit the number of worker
processes.
Forking a child process to call ``ocrmypdf.ocr()`` is suggested. That
way your application will survive and remain interactive even if
OCRmyPDF does not.
Logging
-------
OCRmyPDF will log under loggers named ``ocrmypdf``. In addition, it
imports ``pdfminer`` and ``PIL``, both of which post log messages under
those logging namespaces.
You can configure the logging as desired for your application or call
:func:`ocrmypdf.configure_logging` to configure logging the same way
OCRmyPDF itself does. The command line parameters such as ``--quiet``
and ``--verbose`` have no equivalents in the API; you must use the
provided configuration function or do configuration in a way that suits
your use case.
Progress monitoring
-------------------
OCRmyPDF uses the ``tqdm`` package to implement its progress bars.
:func:`ocrmypdf.configure_logging` will set up logging output to
``sys.stderr`` in a way that is compatible with the display of the
progress bar.
Exceptions
----------
OCRmyPDF may throw standard Python exceptions, ``ocrmypdf.exceptions.*``
exceptions, some exceptions related to multiprocessing, and
``KeyboardInterrupt``. The parent process should provide an exception
handler. OCRmyPDF will clean up its temporary files and worker processes
automatically when an exception occurs.
Programs that call OCRmyPDF should consider trapping KeyboardInterrupt
so that they allow OCR to terminate with the whole program terminating.
When OCRmyPDF succeeds conditionally, it returns an integer exit code.
Reference
---------
.. autofunction:: ocrmypdf.run
.. autoclass:: ocrmypdf.Verbosity
:members:
:undoc-members:
.. autoclass:: ocrmypdf.ExitCode
:members:
:undoc-members:
.. autofunction:: ocrmypdf.configure_logging
docs/batch.rst
+206 -168
View File
@@ -1,225 +1,263 @@
================
Batch processing
================
This article provides information about running OCRmyPDF on multiple files or configuring it as a service triggered by file system events.
This article provides information about running OCRmyPDF on multiple
files or configuring it as a service triggered by file system events.
Batch jobs
----------
==========
Consider using the excellent `GNU Parallel <https://www.gnu.org/software/parallel/>`_ to apply OCRmyPDF to multiple files at once.
Consider using the excellent `GNU
Parallel <https://www.gnu.org/software/parallel/>`__ to apply OCRmyPDF
to multiple files at once.
Both ``parallel`` and ``ocrmypdf`` will try to use all available processors. To maximize parallelism without overloading your system with processes, consider using ``parallel -j 2`` to limit parallel to running two jobs at once.
Both ``parallel`` and ``ocrmypdf`` will try to use all available
processors. To maximize parallelism without overloading your system with
processes, consider using ``parallel -j 2`` to limit parallel to running
two jobs at once.
This command will run all ocrmypdf all files named ``*.pdf`` in the current directory and write them to the previous created ``output/`` folder. It will not search subdirectories.
This command will run all ocrmypdf all files named ``*.pdf`` in the
current directory and write them to the previous created ``output/``
folder. It will not search subdirectories.
The ``--tag`` argument tells parallel to print the filename as a prefix whenever a message is printed, so that one can trace any errors to the file that produced them.
The ``--tag`` argument tells parallel to print the filename as a prefix
whenever a message is printed, so that one can trace any errors to the
file that produced them.
.. code-block:: bash
parallel --tag -j 2 ocrmypdf '{}' 'output/{}' ::: *.pdf
parallel --tag -j 2 ocrmypdf '{}' 'output/{}' ::: *.pdf
OCRmyPDF automatically repairs PDFs before parsing and gathering information from them.
OCRmyPDF automatically repairs PDFs before parsing and gathering
information from them.
Directory trees
---------------
===============
This will walk through a directory tree and run OCR on all files in place, printing the output in a way that makes
This will walk through a directory tree and run OCR on all files in
place, printing the output in a way that makes
.. code-block:: bash
find . -printf '%p' -name '*.pdf' -exec ocrmypdf '{}' '{}' \;
Alternatively, with a docker container (mounts a volume to the container where the PDFs are stored):
find . -printf '%p' -name '*.pdf' -exec ocrmypdf '{}' '{}' \;
Alternatively, with a docker container (mounts a volume to the container
where the PDFs are stored):
.. code-block:: bash
find . -printf '%p' -name '*.pdf' -exec docker run --rm -v <host dir>:<container dir> jbarlow83/ocrmypdf-alpine '<container dir>/{}' '<container dir>/{}' \;
find . -printf '%p' -name '*.pdf' -exec docker run --rm -v <host dir>:<container dir> jbarlow83/ocrmypdf-alpine '<container dir>/{}' '<container dir>/{}' \;
This only runs one ``ocrmypdf`` process at a time. This variation uses ``find`` to create a directory list and ``parallel`` to parallelize runs of ``ocrmypdf``, again updating files in place.
This only runs one ``ocrmypdf`` process at a time. This variation uses
``find`` to create a directory list and ``parallel`` to parallelize runs
of ``ocrmypdf``, again updating files in place.
.. code-block:: bash
find . -name '*.pdf' | parallel --tag -j 2 ocrmypdf '{}' '{}'
find . -name '*.pdf' | parallel --tag -j 2 ocrmypdf '{}' '{}'
Sample script
"""""""""""""
-------------
This user contributed script also provides an example of batch processing.
This user contributed script also provides an example of batch
processing.
.. code-block:: python
#!/usr/bin/env python3
# Walk through directory tree, replacing all files with OCR'd version
# Contributed by DeliciousPickle@github
#!/usr/bin/env python3
# Walk through directory tree, replacing all files with OCR'd version
# Contributed by DeliciousPickle@github
import logging
import os
import subprocess
import sys
import logging
import os
import subprocess
import sys
script_dir = os.path.dirname(os.path.realpath(__file__))
print(script_dir + '/ocr-tree.py: Start')
script_dir = os.path.dirname(os.path.realpath(__file__))
print(script_dir + '/ocr-tree.py: Start')
if len(sys.argv) > 1:
start_dir = sys.argv[1]
else:
start_dir = '.'
if len(sys.argv) > 1:
start_dir = sys.argv[1]
else:
start_dir = '.'
if len(sys.argv) > 2:
log_file = sys.argv[2]
else:
log_file = script_dir + '/ocr-tree.log'
if len(sys.argv) > 2:
log_file = sys.argv[2]
else:
log_file = script_dir + '/ocr-tree.log'
logging.basicConfig(
level=logging.INFO, format='%(asctime)s %(message)s',
filename=log_file, filemode='w')
for dir_name, subdirs, file_list in os.walk(start_dir):
logging.info('\n')
logging.info(dir_name + '\n')
os.chdir(dir_name)
for filename in file_list:
file_ext = os.path.splitext(filename)[1]
if file_ext == '.pdf':
full_path = dir_name + '/' + filename
print(full_path)
cmd = ["ocrmypdf", "--deskew", filename, filename]
logging.info(cmd)
proc = subprocess.run(
cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
result = proc.stdout
if proc.returncode == 6:
print("Skipped document because it already contained text")
elif proc.returncode == 0:
print("OCR complete")
logging.info(result)
API
"""
OCRmyPDF is currently supported as a command line interface. This means that even if you are using OCRmyPDF in a Python script, you should run it in a subprocess rather importing the ocrmypdf package.
The reason for this limitation is that the `ruffus <https://github.com/bunbun/ruffus/>`_ library that OCRmyPDF depends on is unfortunately not reentrant. OCRmyPDF works by defining each operation it does as a ruffus task that takes one or more files as input and generates one or more files as output. As such ruffus is fairly fundamental.
(If you find individual functions implemented in OCRmyPDF useful (such as ``ocrmypdf.pdfinfo``), you can use these if you wish to.)
logging.basicConfig(
level=logging.INFO, format='%(asctime)s %(message)s',
filename=log_file, filemode='w')
for dir_name, subdirs, file_list in os.walk(start_dir):
logging.info('\n')
logging.info(dir_name + '\n')
os.chdir(dir_name)
for filename in file_list:
file_ext = os.path.splitext(filename)[1]
if file_ext == '.pdf':
full_path = dir_name + '/' + filename
print(full_path)
cmd = ["ocrmypdf", "--deskew", filename, filename]
logging.info(cmd)
proc = subprocess.run(
cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
result = proc.stdout
if proc.returncode == 6:
print("Skipped document because it already contained text")
elif proc.returncode == 0:
print("OCR complete")
logging.info(result)
Synology DiskStations
"""""""""""""""""""""
---------------------
Synology DiskStations (Network Attached Storage devices) can run the Docker image of OCRmyPDF if the Synology `Docker package <https://www.synology.com/en-global/dsm/packages/Docker>`_ is installed. Attached is a script to address particular quirks of using OCRmyPDF on one of these devices.
Synology DiskStations (Network Attached Storage devices) can run the
Docker image of OCRmyPDF if the Synology `Docker
package <https://www.synology.com/en-global/dsm/packages/Docker>`__ is
installed. Attached is a script to address particular quirks of using
OCRmyPDF on one of these devices.
This is only possible for x86-based Synology products. Some Synology products use ARM or Power processors and do not support Docker. Further adjustments might be needed to deal with the Synology's relatively limited CPU and RAM.
This is only possible for x86-based Synology products. Some Synology
products use ARM or Power processors and do not support Docker. Further
adjustments might be needed to deal with the Synology's relatively
limited CPU and RAM.
.. code-block:: python
#!/bin/env python3
# Contributed by github.com/Enantiomerie
#!/bin/env python3
# Contributed by github.com/Enantiomerie
# script needs 2 arguments
# 1. source dir with *.pdf - default is location of script
# 2. move dir where *.pdf and *_OCR.pdf are moved to
# script needs 2 arguments
# 1. source dir with *.pdf - default is location of script
# 2. move dir where *.pdf and *_OCR.pdf are moved to
import logging
import os
import subprocess
import sys
import time
import shutil
import logging
import os
import subprocess
import sys
import time
import shutil
script_dir = os.path.dirname(os.path.realpath(__file__))
timestamp = time.strftime("%Y-%m-%d-%H%M_")
log_file = script_dir + '/' + timestamp + 'ocrmypdf.log'
logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s', filename=log_file, filemode='w')
script_dir = os.path.dirname(os.path.realpath(__file__))
timestamp = time.strftime("%Y-%m-%d-%H%M_")
log_file = script_dir + '/' + timestamp + 'ocrmypdf.log'
logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s', filename=log_file, filemode='w')
if len(sys.argv) > 1:
start_dir = sys.argv[1]
else:
start_dir = '.'
if len(sys.argv) > 1:
start_dir = sys.argv[1]
else:
start_dir = '.'
for dir_name, subdirs, file_list in os.walk(start_dir):
logging.info('\n')
logging.info(dir_name + '\n')
os.chdir(dir_name)
for filename in file_list:
file_ext = os.path.splitext(filename)[1]
if file_ext == '.pdf':
full_path = dir_name + '/' + filename
file_noext = os.path.splitext(filename)[0]
timestamp_OCR = time.strftime("%Y-%m-%d-%H%M_OCR_")
filename_OCR = timestamp_OCR + file_noext + '.pdf'
docker_mount = dir_name + ':/home/docker'
# create string for pdf processing
# diskstation needs a user:group docker:docker. find uid:gid of your diskstation docker:docker with id docker.
# use this uid:gid in -u flag
# rw rights for docker:docker at source dir are also necessary
# the script is processed as root user via chron
cmd = ['docker', 'run', '--rm', '-v', docker_mount, '-u=1030:65538', 'jbarlow83/ocrmypdf', , '--deskew' , filename, filename_OCR]
logging.info(cmd)
proc = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
result = proc.stdout.read()
logging.info(result)
full_path_OCR = dir_name + '/' + filename_OCR
os.chmod(full_path_OCR, 0o666)
os.chmod(full_path, 0o666)
full_path_OCR_archive = sys.argv[2]
full_path_archive = sys.argv[2] + '/no_ocr'
shutil.move(full_path_OCR,full_path_OCR_archive)
shutil.move(full_path, full_path_archive)
logging.info('Finished.\n')
for dir_name, subdirs, file_list in os.walk(start_dir):
logging.info('\n')
logging.info(dir_name + '\n')
os.chdir(dir_name)
for filename in file_list:
file_ext = os.path.splitext(filename)[1]
if file_ext == '.pdf':
full_path = dir_name + '/' + filename
file_noext = os.path.splitext(filename)[0]
timestamp_OCR = time.strftime("%Y-%m-%d-%H%M_OCR_")
filename_OCR = timestamp_OCR + file_noext + '.pdf'
docker_mount = dir_name + ':/home/docker'
# create string for pdf processing
# diskstation needs a user:group docker:docker. find uid:gid of your diskstation docker:docker with id docker.
# use this uid:gid in -u flag
# rw rights for docker:docker at source dir are also necessary
# the script is processed as root user via chron
cmd = ['docker', 'run', '--rm', '-v', docker_mount, '-u=1030:65538', 'jbarlow83/ocrmypdf', , '--deskew' , filename, filename_OCR]
logging.info(cmd)
proc = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
result = proc.stdout.read()
logging.info(result)
full_path_OCR = dir_name + '/' + filename_OCR
os.chmod(full_path_OCR, 0o666)
os.chmod(full_path, 0o666)
full_path_OCR_archive = sys.argv[2]
full_path_archive = sys.argv[2] + '/no_ocr'
shutil.move(full_path_OCR,full_path_OCR_archive)
shutil.move(full_path, full_path_archive)
logging.info('Finished.\n')
Huge batch jobs
"""""""""""""""
If you have thousands of files to work with, contact the author. Consulting work related to OCRmyPDF helps fund this open source project and all inquiries are appreciated.
Hot (watched) folders
---------------------
To set up a "hot folder" that will trigger OCR for every file inserted, use a program like Python `watchdog <https://pypi.python.org/pypi/watchdog>`_ (supports all major OS).
One could then configure a scanner to automatically place scanned files in a hot folder, so that they will be queued for OCR and copied to the destination.
.. code-block:: bash
pip install watchdog
watchdog installs the command line program ``watchmedo``, which can be told to run ``ocrmypdf`` on any .pdf added to the current directory (``.``) and place the result in the previously created ``out/`` folder.
.. code-block:: bash
cd hot-folder
mkdir out
watchmedo shell-command \
--patterns="*.pdf" \
--ignore-directories \
--command='ocrmypdf "${watch_src_path}" "out/${watch_src_path}" ' \
. # don't forget the final dot
For more complex behavior you can write a Python script around to use the watchdog API.
On file servers, you could configure watchmedo as a system service so it will run all the time.
Caveats
"""""""
* ``watchmedo`` may not work properly on a networked file system, depending on the capabilities of the file system client and server.
* This simple recipe does not filter for the type of file system event, so file copies, deletes and moves, and directory operations, will all be sent to ocrmypdf, producing errors in several cases. Disable your watched folder if you are doing anything other than copying files to it.
* If the source and destination directory are the same, watchmedo may create an infinite loop.
* On BSD, FreeBSD and older versions of macOS, you may need to increase the number of file descriptors to monitor more files, using ``ulimit -n 1024`` to watch a folder of up to 1024 files.
Alternatives
""""""""""""
* `Watchman <https://facebook.github.io/watchman/>`_ is a more powerful alternative to ``watchmedo``.
macOS Automator
---------------
You can use the Automator app with macOS, to create a Workflow or Quick Action. Use a *Run Shell Script* action in your workflow. In the context of Automator, the ``PATH`` may be set differently your Terminal's ``PATH``; you may need to explicitly set the PATH to include ``ocrmypdf``. The following example may serve as a starting point:
If you have thousands of files to work with, contact the author.
Consulting work related to OCRmyPDF helps fund this open source project
and all inquiries are appreciated.
.. image:: images/macos-workflow.png
:alt: Example macOS Automator script
Hot (watched) folders
=====================
To set up a "hot folder" that will trigger OCR for every file inserted,
use a program like Python
`watchdog <https://pypi.python.org/pypi/watchdog>`__ (supports all major
OS).
One could then configure a scanner to automatically place scanned files
in a hot folder, so that they will be queued for OCR and copied to the
destination.
.. code-block:: bash
pip install watchdog
watchdog installs the command line program ``watchmedo``, which can be
told to run ``ocrmypdf`` on any .pdf added to the current directory
(``.``) and place the result in the previously created ``out/`` folder.
.. code-block:: bash
cd hot-folder
mkdir out
watchmedo shell-command \
--patterns="*.pdf" \
--ignore-directories \
--command='ocrmypdf "${watch_src_path}" "out/${watch_src_path}" ' \
. # don't forget the final dot
For more complex behavior you can write a Python script around to use
the watchdog API.
On file servers, you could configure watchmedo as a system service so it
will run all the time.
Caveats
-------
- ``watchmedo`` may not work properly on a networked file system,
depending on the capabilities of the file system client and server.
- This simple recipe does not filter for the type of file system event,
so file copies, deletes and moves, and directory operations, will all
be sent to ocrmypdf, producing errors in several cases. Disable your
watched folder if you are doing anything other than copying files to
it.
- If the source and destination directory are the same, watchmedo may
create an infinite loop.
- On BSD, FreeBSD and older versions of macOS, you may need to increase
the number of file descriptors to monitor more files, using
``ulimit -n 1024`` to watch a folder of up to 1024 files.
Alternatives
------------
- `Watchman <https://facebook.github.io/watchman/>`__ is a more
powerful alternative to ``watchmedo``.
macOS Automator
===============
You can use the Automator app with macOS, to create a Workflow or Quick
Action. Use a *Run Shell Script* action in your workflow. In the context
of Automator, the ``PATH`` may be set differently your Terminal's
``PATH``; you may need to explicitly set the PATH to include
``ocrmypdf``. The following example may serve as a starting point:
|Example macOS Automator script|
You may customize the command sent to ocrmypdf.
.. |Example macOS Automator script| image:: images/macos-workflow.png
docs/conf.py
+1 -3
View File
@@ -30,9 +30,7 @@
# Add any Sphinx extension module names here, as strings. They can be
# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
# ones.
extensions = [
# 'sphinx.ext.mathjax',
]
extensions = ['sphinx.ext.napoleon']
# Add any paths that contain templates here, relative to this directory.
templates_path = ['_templates']
docs/cookbook.rst
+135 -65
View File
@@ -1,11 +1,12 @@
========
Cookbook
========
Basic examples
--------------
==============
Help!
^^^^^
-----
ocrmypdf has built-in help.
@@ -13,30 +14,29 @@ ocrmypdf has built-in help.
ocrmypdf --help
Add an OCR layer and convert to PDF/A
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-------------------------------------
.. code-block:: bash
ocrmypdf input.pdf output.pdf
Add an OCR layer and output a standard PDF
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
------------------------------------------
.. code-block:: bash
ocrmypdf --output-type pdf input.pdf output.pdf
Create a PDF/A with all color and grayscale images converted to JPEG
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
--------------------------------------------------------------------
.. code-block:: bash
ocrmypdf --output-type pdfa --pdfa-image-compression jpeg input.pdf output.pdf
Modify a file in place
^^^^^^^^^^^^^^^^^^^^^^
----------------------
The file will only be overwritten if OCRmyPDF is successful.
@@ -45,48 +45,58 @@ The file will only be overwritten if OCRmyPDF is successful.
ocrmypdf myfile.pdf myfile.pdf
Correct page rotation
^^^^^^^^^^^^^^^^^^^^^
---------------------
OCR will attempt to automatic correct the rotation of each page. This can help fix a scanning job that contains a mix of landscape and portrait pages.
OCR will attempt to automatic correct the rotation of each page. This
can help fix a scanning job that contains a mix of landscape and
portrait pages.
.. code-block:: bash
ocrmypdf --rotate-pages myfile.pdf myfile.pdf
You can increase (decrease) the parameter ``--rotate-pages-threshold`` to make page rotation more (less) aggressive.
You can increase (decrease) the parameter ``--rotate-pages-threshold``
to make page rotation more (less) aggressive.
If the page is "just a little off horizontal", like a crooked picture, then you want ``--deskew``. ``--rotate-pages`` is for when the cardinal angle is wrong.
If the page is "just a little off horizontal", like a crooked picture,
then you want ``--deskew``. ``--rotate-pages`` is for when the cardinal
angle is wrong.
OCR languages other than English
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
--------------------------------
OCRmyPDF assumes the document is in English unless told otherwise. OCR quality may be poor if the wrong language is used.
OCRmyPDF assumes the document is in English unless told otherwise. OCR
quality may be poor if the wrong language is used.
.. code-block:: bash
ocrmypdf -l fra LeParisien.pdf LeParisien.pdf
ocrmypdf -l eng+fra Bilingual-English-French.pdf Bilingual-English-French.pdf
Language packs must be installed for all languages specified. See :ref:`Installing additional language packs <lang-packs>`.
Language packs must be installed for all languages specified. See
:ref:`Installing additional language packs <lang-packs>`.
Unfortunately, the Tesseract OCR engine has no ability to detect the language when it is unknown.
Unfortunately, the Tesseract OCR engine has no ability to detect the
language when it is unknown.
Produce PDF and text file containing OCR text
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
---------------------------------------------
This produces a file named "output.pdf" and a companion text file named "output.txt".
This produces a file named "output.pdf" and a companion text file named
"output.txt".
.. code-block:: bash
ocrmypdf --sidecar output.txt input.pdf output.pdf
OCR images, not PDFs
^^^^^^^^^^^^^^^^^^^^
--------------------
Option: use Tesseract
"""""""""""""""""""""
~~~~~~~~~~~~~~~~~~~~~
If you are starting with images, you can just use Tesseract directly to convert images to PDFs:
If you are starting with images, you can just use Tesseract directly to
convert images to PDFs:
.. code-block:: bash
@@ -97,62 +107,88 @@ If you are starting with images, you can just use Tesseract directly to convert
# When there are multiple images
tesseract text-file-containing-list-of-image-filenames.txt output-prefix pdf
Tesseract's PDF output is quite good – OCRmyPDF uses it internally, in some cases. However, OCRmyPDF has many features not available in Tesseract like image processing, metadata control, and PDF/A generation.
Tesseract's PDF output is quite good – OCRmyPDF uses it internally, in
some cases. However, OCRmyPDF has many features not available in
Tesseract like image processing, metadata control, and PDF/A generation.
Option: use img2pdf
"""""""""""""""""""
~~~~~~~~~~~~~~~~~~~
You can also use a program like `img2pdf <https://gitlab.mister-muffin.de/josch/img2pdf>`_ to convert your images to PDFs, and then pipe the results to run ocrmypdf. The ``-`` tells ocrmypdf to read standard input.
You can also use a program like
`img2pdf <https://gitlab.mister-muffin.de/josch/img2pdf>`__ to convert
your images to PDFs, and then pipe the results to run ocrmypdf. The
``-`` tells ocrmypdf to read standard input.
.. code-block:: bash
img2pdf my-images*.jpg | ocrmypdf - myfile.pdf
``img2pdf`` is recommended because it does an excellent job at generating PDFs without transcoding images.
``img2pdf`` is recommended because it does an excellent job at
generating PDFs without transcoding images.
Option: use OCRmyPDF (single images only)
"""""""""""""""""""""""""""""""""""""""""
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For convenience, OCRmyPDF can also convert single images to PDFs on its own. If the resolution (dots per inch, DPI) of an image is not set or is incorrect, it can be overridden with ``--image-dpi``. (As 1 inch is 2.54 cm, 1 dpi = 0.39 dpcm).
For convenience, OCRmyPDF can also convert single images to PDFs on its
own. If the resolution (dots per inch, DPI) of an image is not set or is
incorrect, it can be overridden with ``--image-dpi``. (As 1 inch is 2.54
cm, 1 dpi = 0.39 dpcm).
.. code-block:: bash
ocrmypdf --image-dpi 300 image.png myfile.pdf
If you have multiple images, you must use ``img2pdf`` to convert the images to PDF.
If you have multiple images, you must use ``img2pdf`` to convert the
images to PDF.
Not recommended
"""""""""""""""
~~~~~~~~~~~~~~~
We caution against using ImageMagick or Ghostscript to convert images to PDF, since they may transcode images or produce downsampled images, sometimes without warning.
We caution against using ImageMagick or Ghostscript to convert images to
PDF, since they may transcode images or produce downsampled images,
sometimes without warning.
Image processing
----------------
================
OCRmyPDF perform some image processing on each page of a PDF, if desired. The same processing is applied to each page. It is suggested that the user review files after image processing as these commands might remove desirable content, especially from poor quality scans.
OCRmyPDF perform some image processing on each page of a PDF, if
desired. The same processing is applied to each page. It is suggested
that the user review files after image processing as these commands
might remove desirable content, especially from poor quality scans.
* ``--rotate-pages`` attempts to determine the correct orientation for each page and rotates the page if necessary.
* ``--remove-background`` attempts to detect and remove a noisy background from grayscale or color images. Monochrome images are ignored. This should not be used on documents that contain color photos as it may remove them.
* ``--deskew`` will correct pages were scanned at a skewed angle by rotating them back into place. Skew determination and correction is performed using `Postl's variance of line sums <http://www.leptonica.com/skew-measurement.html>`_ algorithm as implemented in `Leptonica <http://www.leptonica.com/index.html>`_.
* ``--clean`` uses `unpaper <https://www.flameeyes.eu/projects/unpaper>`_ to clean up pages before OCR, but does not alter the final output. This makes it less likely that OCR will try to find text in background noise.
* ``--clean-final`` uses unpaper to clean up pages before OCR and inserts the page into the final output. You will want to review each page to ensure that unpaper did not remove something important.
* ``--mask-barcodes`` will suppress any barcodes detected in a page image. Barcodes are known to confuse Tesseract OCR and interfere with the recognition of text on the same baseline as a barcode. The output file will contain the unaltered image of the barcode.
- ``--rotate-pages`` attempts to determine the correct orientation for
each page and rotates the page if necessary.
- ``--remove-background`` attempts to detect and remove a noisy
background from grayscale or color images. Monochrome images are
ignored. This should not be used on documents that contain color
photos as it may remove them.
- ``--deskew`` will correct pages were scanned at a skewed angle by
rotating them back into place. Skew determination and correction is
performed using `Postl's variance of line
sums <http://www.leptonica.com/skew-measurement.html>`__ algorithm as
implemented in `Leptonica <http://www.leptonica.com/index.html>`__.
- ``--clean`` uses
`unpaper <https://www.flameeyes.eu/projects/unpaper>`__ to clean up
pages before OCR, but does not alter the final output. This makes it
less likely that OCR will try to find text in background noise.
- ``--clean-final`` uses unpaper to clean up pages before OCR and
inserts the page into the final output. You will want to review each
page to ensure that unpaper did not remove something important.
.. note::
In many cases image processing will rasterize PDF pages as images, potentially losing quality.
In many cases image processing will rasterize PDF pages as images,
potentially losing quality.
.. warning::
``--clean-final`` and ``-remove-background`` may leave undesirable visual artifacts in some images where their algorithms have shortcomings. Files should be visually reviewed after using these options.
``--clean-final`` and ``-remove-background`` may leave undesirable
visual artifacts in some images where their algorithms have
shortcomings. Files should be visually reviewed after using these
options.
Example: OCR and correct document skew (crooked scan)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-----------------------------------------------------
Deskew:
@@ -160,55 +196,83 @@ Deskew:
ocrmypdf --deskew input.pdf output.pdf
Image processing commands can be combined. The order in which options are given does not matter. OCRmyPDF always applies the steps of the image processing pipeline in the same order (rotate, remove background, deskew, clean).
Image processing commands can be combined. The order in which options
are given does not matter. OCRmyPDF always applies the steps of the
image processing pipeline in the same order (rotate, remove background,
deskew, clean).
.. code-block:: bash
ocrmypdf --deskew --clean --rotate-pages input.pdf output.pdf
Don't actually OCR my PDF
-------------------------
=========================
If you set ``--tesseract-timeout 0`` OCRmyPDF will apply its image processing without performing OCR, if all you want to is to apply image processing or PDF/A conversion.
If you set ``--tesseract-timeout 0`` OCRmyPDF will apply its image
processing without performing OCR, if all you want to is to apply image
processing or PDF/A conversion.
.. code-block:: bash
ocrmypdf --tesseract-timeout=0 --remove-background input.pdf output.pdf
Redo existing OCR
-----------------
=================
To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--redo-ocr`` argument. (Normally, OCRmyPDF will exit with an error if asked to modify a file with OCR.)
To redo OCR on a file OCRed with other OCR software or a previous
version of OCRmyPDF and/or Tesseract, you may use the ``--redo-ocr``
argument. (Normally, OCRmyPDF will exit with an error if asked to modify
a file with OCR.)
This may be helpful for users who want to take advantage of accuracy improvements in Tesseract 4.0 for files they previously OCRed with an earlier version of Tesseract and OCRmyPDF.
This may be helpful for users who want to take advantage of accuracy
improvements in Tesseract 4.0 for files they previously OCRed with an
earlier version of Tesseract and OCRmyPDF.
.. code-block:: bash
ocrmypdf --redo-ocr input.pdf output.pdf
This method will replace OCR without rasterizing, reducing quality or removing vector content. If a file contains a mix of pure digital text and OCR, digital text will be ignored and OCR will be replaced. As such this mode is incompatible with image processing options, since they alter the appearance of the file.
This method will replace OCR without rasterizing, reducing quality or
removing vector content. If a file contains a mix of pure digital text
and OCR, digital text will be ignored and OCR will be replaced. As such
this mode is incompatible with image processing options, since they
alter the appearance of the file.
In some cases, existing OCR cannot be detected or replaced. Files produced by OCRmyPDF v2.2 or earlier, for example, are internally represented as having visible text with an opaque image drawn on top. This situation cannot be detected.
In some cases, existing OCR cannot be detected or replaced. Files
produced by OCRmyPDF v2.2 or earlier, for example, are internally
represented as having visible text with an opaque image drawn on top.
This situation cannot be detected.
If ``--redo-ocr`` does not work, you can use ``--force-ocr``, which will force rasterization of all pages, potentially reducing quality or losing vector content.
If ``--redo-ocr`` does not work, you can use ``--force-ocr``, which will
force rasterization of all pages, potentially reducing quality or losing
vector content.
Improving OCR quality
---------------------
=====================
The `Image processing`_ features can improve OCR quality.
The `Image processing <#image-processing>`__ features can improve OCR
quality.
Rotating pages and deskewing helps to ensure that the page orientation is correct before OCR begins. Removing the background and/or cleaning the page can also improve results. The ``--oversample DPI`` argument can be specified to resample images to higher resolution before attempting OCR; this can improve results as well.
Rotating pages and deskewing helps to ensure that the page orientation
is correct before OCR begins. Removing the background and/or cleaning
the page can also improve results. The ``--oversample DPI`` argument can
be specified to resample images to higher resolution before attempting
OCR; this can improve results as well.
OCR quality will suffer if the resolution of input images is not correct (since the range of pixel sizes that will be checked for possible fonts will also be incorrect).
OCR quality will suffer if the resolution of input images is not correct
(since the range of pixel sizes that will be checked for possible fonts
will also be incorrect).
PDF optimization
----------------
================
By default OCRmyPDF will attempt to perform lossless optimizations on the images inside PDFs after OCR is complete. Optimization is performed even if no OCR text is found.
By default OCRmyPDF will attempt to perform lossless optimizations on
the images inside PDFs after OCR is complete. Optimization is performed
even if no OCR text is found.
The ``--optimize N`` (short form ``-O``) argument controls optimization, where ``N`` ranges from 0 to 3 inclusive, analogous to the optimization levels in the GCC compiler.
The ``--optimize N`` (short form ``-O``) argument controls optimization,
where ``N`` ranges from 0 to 3 inclusive, analogous to the optimization
levels in the GCC compiler.
.. list-table::
:widths: auto
@@ -227,9 +291,15 @@ The ``--optimize N`` (short form ``-O``) argument controls optimization, where `
* - ``--optimize 3``
- All of the above, and enables more aggressive optimizations and targets lower image quality.
Optimization is improved when a JBIG2 encoder is available and when ``pngquant`` is installed. If either of these components are missing, then some types of images cannot be optimized.
Optimization is improved when a JBIG2 encoder is available and when
``pngquant`` is installed. If either of these components are missing,
then some types of images cannot be optimized.
The types of optimization available may expand over time. By default, OCRmyPDF compresses data streams inside PDFs, and will change inefficient compression modes to more modern versions. A program like ``qpdf`` can be used to change encodings, e.g. to inspect the internals fo a PDF.
The types of optimization available may expand over time. By default,
OCRmyPDF compresses data streams inside PDFs, and will change
inefficient compression modes to more modern versions. A program like
``qpdf`` can be used to change encodings, e.g. to inspect the internals
fo a PDF.
.. code-block:: bash
docs/docker.rst
+104 -84
View File
@@ -1,155 +1,175 @@
=====================
OCRmyPDF Docker image
=====================
OCRmyPDF is also available in a Docker image that packages recent versions of all dependencies.
OCRmyPDF is also available in a Docker image that packages recent
versions of all dependencies.
For users who already have Docker installed this may be an easy and convenient option. However, it is less performant than a system installation and may require Docker engine configuration.
For users who already have Docker installed this may be an easy and
convenient option. However, it is less performant than a system
installation and may require Docker engine configuration.
OCRmyPDF needs a generous amount of RAM, CPU cores, and temporary storage space.
OCRmyPDF needs a generous amount of RAM, CPU cores, temporary storage
space, whether running in a Docker container or on its own. It may be
necessary to ensure the container is provisioned with additional
resources.
.. _docker-install:
Installing the Docker image
---------------------------
===========================
If you have `Docker <https://docs.docker.com/>`_ installed on your system, you can install a Docker image of the latest release.
If you have `Docker <https://docs.docker.com/>`__ installed on your
system, you can install a Docker image of the latest release.
The recommended OCRmyPDF Docker image is currently named ``ocrmypdf-alpine``:
The recommended OCRmyPDF Docker image is currently named
``ocrmypdf-alpine``:
.. code-block:: bash
docker pull jbarlow83/ocrmypdf-alpine
docker pull jbarlow83/ocrmypdf-alpine
Follow the Docker installation instructions for your platform. If you can run this command successfully, your system is ready to download and execute the image:
Follow the Docker installation instructions for your platform. If you
can run this command successfully, your system is ready to download and
execute the image:
.. code-block:: bash
docker run hello-world
docker run hello-world
OCRmyPDF will use all available CPU cores. By default, the VirtualBox machine instance on Windows and macOS has only a single CPU core enabled. Use the VirtualBox Manager to determine the name of your Docker engine host, and then follow these optional steps to enable multiple CPUs:
OCRmyPDF will use all available CPU cores. By default, the VirtualBox
machine instance on Windows and macOS has only a single CPU core
enabled. Use the VirtualBox Manager to determine the name of your Docker
engine host, and then follow these optional steps to enable multiple
CPUs:
.. code-block:: bash
# Optional step for Mac OS X users
docker-machine stop "yourVM"
VBoxManage modifyvm "yourVM" --cpus 2 # or whatever number of core is desired
docker-machine start "yourVM"
eval $(docker-machine env "yourVM")
# Optional step for Mac OS X users
docker-machine stop "yourVM"
VBoxManage modifyvm "yourVM" --cpus 2 # or whatever number of core is desired
docker-machine start "yourVM"
eval $(docker-machine env "yourVM")
Using the Docker image on the command line
------------------------------------------
==========================================
**Unlike typical Docker containers**, in this mode we are using the OCRmyPDF Docker container is intended to be emphemeral – it runs for one OCR job and then terminates, just like a command line program. We are using Docker as a way of delivering an application, not a server.
**Unlike typical Docker containers**, in this mode we are using the
OCRmyPDF Docker container is intended to be emphemeral – it runs for one
OCR job and then terminates, just like a command line program. We are
using Docker as a way of delivering an application, not a server.
To start a Docker container (instance of the image):
.. code-block:: bash
docker tag jbarlow83/ocrmypdf-alpine ocrmypdf
docker run --rm ocrmypdf (... all other arguments here...)
docker tag jbarlow83/ocrmypdf-alpine ocrmypdf
docker run --rm -i ocrmypdf (... all other arguments here...)
For convenience, create a shell alias to hide the Docker command:
For convenience, create a shell alias to hide the Docker command. It is
easier to send the input file to file stdin and read the output from
stdout – this avoids the occasionally messy permission issues with
Docker entirely.
.. code-block:: bash
alias ocrmypdf='docker run --rm -v "$(pwd):/home/docker" ocrmypdf'
ocrmypdf --version # runs docker version
alias ocrmypdf='docker run --rm -i ocrmypdf'
ocrmypdf --version # runs docker version
ocrmypdf <input.pdf >output.pdf
Or in the wonderful `fish shell <https://fishshell.com/>`_:
Or in the wonderful `fish shell <https://fishshell.com/>`__:
.. code-block:: fish
alias ocrmypdf 'docker run --rm ocrmypdf'
funcsave ocrmypdf
alias ocrmypdf 'docker run --rm ocrmypdf'
funcsave ocrmypdf
Alternately, you could mount the local current working directory as a
Docker volume:
.. code-block:: bash
docker run --rm -v $(pwd):/data ocrmypdf /data/input.pdf /data/output.pdf
.. _docker-lang-packs:
Adding languages to the Docker image
------------------------------------
====================================
By default the Docker image includes English, German and Simplified Chinese, the most popular languages for OCRmyPDF users based on feedback. You may add other languages by creating a new Dockerfile based on the public one:
By default the Docker image includes English, German and Simplified
Chinese, the most popular languages for OCRmyPDF users based on
feedback. You may add other languages by creating a new Dockerfile based
on the public one:
.. code-block:: dockerfile
FROM jbarlow83/ocrmypdf-alpine
FROM jbarlow83/ocrmypdf-alpine
# Add French
RUN apk add tesseract-ocr-data-fra
# Add French
RUN apk add tesseract-ocr-data-fra
You can also copy training data to ``/usr/share/tessdata``.
Executing the test suite
------------------------
========================
The OCRmyPDF test suite is installed with image. To run it:
The OCRmyPDF test suite is installed with image. To run it:
.. code-block:: bash
docker run --entrypoint python3 jbarlow83/ocrmypdf-alpine setup.py test
docker run --entrypoint python3 jbarlow83/ocrmypdf-alpine setup.py test
Accessing the shell
===================
``bash`` is not installed in the image. To use the busybox shell in the
Docker image:
.. code-block:: bash
docker run -it --entrypoint busybox jbarlow83/ocrmypdf-alpine sh
Using the OCRmyPDF web service wrapper
--------------------------------------
======================================
The OCRmyPDF Docker image includes an example, barebones HTTP web service. The webservice may be launched as follows:
The OCRmyPDF Docker image includes an example, barebones HTTP web
service. The webservice may be launched as follows:
.. code-block:: bash
docker run --entrypoint python3 -p 5000:5000 jbarlow83/ocrmypdf-alpine webservice.py
docker run --entrypoint python3 -p 5000:5000 jbarlow83/ocrmypdf-alpine webservice.py
Unlike command line usage this program will open a socket and wait for connections.
Unlike command line usage this program will open a socket and wait for
connections.
.. warning::
The OCRmyPDF web service wrapper is intended for demonstration or development. It provides no security, no authentication, no protection against denial of service attacks, and no load balancing. The default Flask WSGI server is used, which is intended for development only. The server is single-threaded and so can respond to only one client at a time. It cannot respond to clients while busy with OCR.
The OCRmyPDF web service wrapper is intended for demonstration or
development. It provides no security, no authentication, no
protection against denial of service attacks, and no load balancing.
The default Flask WSGI server is used, which is intended for
development only. The server is single-threaded and so can respond to
only one client at a time. While running OCR, it cannot respond to
any other clients.
Clients must keep their open connection while waiting for OCR to complete. This may entail setting a long timeout; this interface is more useful for internal HTTP API calls.
Clients must keep their open connection while waiting for OCR to
complete. This may entail setting a long timeout; this interface is more
useful for internal HTTP API calls.
Unlike the rest of OCRmyPDF, this web service is licensed under the Affero GPLv3 (AGPLv3) since Ghostscript, a dependency of OCRmyPDF, is also licensed in this way.
Unlike the rest of OCRmyPDF, this web service is licensed under the
Affero GPLv3 (AGPLv3) since Ghostscript, a dependency of OCRmyPDF, is
also licensed in this way.
In addition to the above, please read our :ref:`general remarks on using OCRmyPDF as a service <ocr-service>`.
In addition to the above, please read our
:ref:`general remarks on using OCRmyPDF as a service <ocr-service>`.
Legacy Ubuntu Docker images
---------------------------
Ubuntu-based Docker image
=========================
Previously OCRmyPDF was delivered in several Docker images for different purposes, based on Ubuntu.
The Ubuntu-based images will be maintained for some time but should not be used for new deployments. They are as follows:
.. list-table::
:widths: auto
:header-rows: 1
* - Image name
- Download command
- Notes
* - ocrmypdf
- ``docker pull jbarlow83/ocrmypdf``
- Latest ocrmypdf with Tesseract 4.0.0-beta1 on Ubuntu 18.04. Includes English, French, German, Spanish, Portugeuse and Simplified Chinese.
* - ocrmypdf-polyglot
- ``docker pull jbarlow83/ocrmypdf-polyglot``
- As above, with all available language packs.
* - ocrmypdf-webservice
- ``docker pull jbarlow83/ocrmypdf-webservice``
- All language packs, and a simple HTTP wrapper allowing OCRmyPDF to be used as a web service. Note that this component is licensed under AGPLv3.
To execute the Ubuntu-based OCRmyPDF on a local file, you must `provide a writable volume to the Docker image <https://docs.docker.com/userguide/dockervolumes/>`_, and both the input and output file must be inside the writable volume. This limitation applies only to the legacy images.
This example command uses the current working directory as the writable volume:
A Ubuntu-based OCRmyPDF image is also available. The main advantage this
image offers is that it supports manylinux Python wheels (which are not
supported on Alpine Linux). This may be useful for plugins.
.. code-block:: bash
docker run --rm -v "$(pwd):/home/docker" <other docker arguments> ocrmypdf <your arguments to ocrmypdf>
In this worked example, the current working directory contains an input file called ``test.pdf`` and the output will go to ``output.pdf``:
.. code-block:: bash
docker run --rm -v "$(pwd):/home/docker" ocrmypdf --skip-text test.pdf output.pdf
.. note:: The working directory should be a writable local volume or Docker may not have permission to access it.
Note that ``ocrmypdf`` has its own separate ``-v VERBOSITYLEVEL`` argument to control debug verbosity. All Docker arguments should before the ``ocrmypdf`` image name and all arguments to ``ocrmypdf`` should be listed after.
In some environments the permissions associated with Docker can be complex to configure. The process that executes Docker may end up not having the permissions to write the specified file system. In that case one can stream the file into and out of the Docker process and avoid all permission hassles, using ``-`` as the input and output filename:
.. code-block:: bash
docker run --rm -i ocrmypdf <other arguments to ocrmypdf> - - <input.pdf >output.pdf
docker pull jbarlow83/ocrmypdf
docs/errors.rst
+29 -15
View File
@@ -1,33 +1,47 @@
=====================
Common error messages
=====================
Page already has text
---------------------
=====================
.. code::
.. code-block::
ERROR - 1: page already has text! – aborting (use --force-ocr to force OCR)
ERROR - 1: page already has text! – aborting (use --force-ocr to force OCR)
You ran ocrmypdf on a file that already contains printable text or a hidden OCR text layer (it can't quite tell the difference). You probably don't want to do this, because the file is already searchable.
You ran ocrmypdf on a file that already contains printable text or a
hidden OCR text layer (it can't quite tell the difference). You probably
don't want to do this, because the file is already searchable.
As the error message suggests, your options are:
- ``ocrmypdf --force-ocr`` to :ref:`rasterize <raster-vector>` all vector content and run OCR on the images. This is useful if a previous OCR program failed, or if the document contains a text watermark.
- ``ocrmypdf --skip-text`` to skip OCR and other processing on any pages that contain text. Text pages will be copied into the output PDF without modification.
- ``ocrmypdf --force-ocr`` to :ref:`rasterize <raster-vector>` all
vector content and run OCR on the images. This is useful if a
previous OCR program failed, or if the document contains a text
watermark.
- ``ocrmypdf --skip-text`` to skip OCR and other processing on any
pages that contain text. Text pages will be copied into the output
PDF without modification.
Input file 'filename' is not a valid PDF
----------------------------------------
========================================
OCRmyPDF passes files through qpdf, a program that fixes errors in PDFs, before it tries to work on them. In most cases this happens because the PDF is corrupt and
truncated (incomplete file copying) and not much can be done.
OCRmyPDF passes files through qpdf, a program that fixes errors in PDFs,
before it tries to work on them. In most cases this happens because the
PDF is corrupt and truncated (incomplete file copying) and not much can
be done.
You can try rewriting the file with Ghostscript or pdftk:
You can try rewriting the file with Ghostscript:
- ``gs -o output.pdf -dSAFER -sDEVICE=pdfwrite input.pdf``
.. code-block:: bash
- ``pdftk input.pdf cat output output.pdf``
gs -o output.pdf -dSAFER -sDEVICE=pdfwrite input.pdf
Sometimes Acrobat can repair PDFs with its `Preflight tool <https://helpx.adobe.com/acrobat/using/correcting-problem-areas-preflight-tool.html>`_.
``pdftk`` can also rewrite PDFs:
.. code-block:: bash
pdftk input.pdf cat output output.pdf
Sometimes Acrobat can repair PDFs with its `Preflight
tool <https://helpx.adobe.com/acrobat/using/correcting-problem-areas-preflight-tool.html>`__.
docs/images/logo-social.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 31 KiB

docs/images/logo.svg
+75
View File
@@ -0,0 +1,75 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
<svg width="100%" height="100%" viewBox="0 0 503 227" version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" xml:space="preserve" xmlns:serif="http://www.serif.com/" style="fill-rule:evenodd;clip-rule:evenodd;stroke-linecap:round;stroke-linejoin:round;stroke-miterlimit:1.5;">
<g id="svg" transform="matrix(0.965977,0,0,0.807602,0,0)">
<rect x="0" y="0" width="520" height="280" style="fill:white;"/>
<g transform="matrix(1.03522,0,0,1.23823,-69.7528,-83.422)">
<g transform="matrix(1,0,0,1,243.977,20.0703)">
<g id="Page">
<g transform="matrix(0.961773,0,0,1.05962,6.19811,-3.01071)">
<path d="M328.5,97.682C328.5,96.465 327.983,95.296 327.056,94.418C320.026,87.758 289.442,58.78 282.228,51.944C281.251,51.019 279.901,50.496 278.49,50.496C264.493,50.496 188.083,50.496 167.339,50.496C164.468,50.496 162.141,52.609 162.141,55.214C162.141,83.051 162.141,225.565 162.141,253.4C162.141,256.005 164.468,258.117 167.338,258.117C192.242,258.117 299.159,258.117 323.538,258.117C326.278,258.117 328.5,256.101 328.5,253.613C328.5,229.268 328.5,113.896 328.5,97.682Z" style="fill:rgb(253,253,253);stroke:rgb(51,51,51);stroke-width:3.95px;"/>
</g>
<g id="Dog-ear" serif:id="Dog ear" transform="matrix(1,0,0,1,-4,2)">
<path d="M277.072,48.496L277.072,93.848C277.072,95.172 277.598,96.441 278.534,97.377C279.47,98.313 280.739,98.839 282.063,98.839C294.548,98.839 326.141,98.839 326.141,98.839" style="fill:rgb(245,245,245);stroke:rgb(51,51,51);stroke-width:4px;"/>
</g>
</g>
<g transform="matrix(1,0,0,1,-29.6816,-0.395178)">
<g transform="matrix(1.00243,0,0,1.11818,-144.72,-8.80181)">
<path d="M465.73,119.654C465.73,117.605 463.874,115.941 461.588,115.941L310.259,115.941C307.973,115.941 306.117,117.605 306.117,119.654L306.117,183.108C306.117,185.157 307.973,186.821 310.259,186.821L461.588,186.821C463.874,186.821 465.73,185.157 465.73,183.108L465.73,119.654Z" style="fill:rgb(248,0,0);stroke:white;stroke-width:3.77px;"/>
</g>
<g transform="matrix(1.24571,0,0,1.35864,116.812,84.3924)">
<g transform="matrix(64,0,0,64,42.1437,77.6203)">
<path d="M0.084,0L0.084,-0.68L0.297,-0.68C0.371,-0.68 0.434,-0.663 0.487,-0.63C0.54,-0.596 0.566,-0.54 0.566,-0.462C0.566,-0.385 0.538,-0.328 0.481,-0.292C0.424,-0.255 0.36,-0.237 0.288,-0.237L0.213,-0.237L0.213,0L0.084,0ZM0.293,-0.572L0.213,-0.572L0.213,-0.344L0.295,-0.344C0.334,-0.344 0.365,-0.353 0.389,-0.371C0.413,-0.388 0.426,-0.416 0.429,-0.454C0.429,-0.498 0.417,-0.529 0.393,-0.546C0.369,-0.563 0.336,-0.572 0.293,-0.572Z" style="fill:white;fill-rule:nonzero;"/>
</g>
<g transform="matrix(64,0,0,64,79.7117,77.6203)">
<path d="M0.332,0L0.084,0L0.084,-0.68L0.336,-0.68C0.441,-0.68 0.518,-0.648 0.569,-0.585C0.62,-0.522 0.645,-0.441 0.645,-0.344C0.645,-0.239 0.618,-0.155 0.563,-0.093C0.508,-0.031 0.431,0 0.332,0ZM0.337,-0.57L0.213,-0.57L0.213,-0.109L0.33,-0.109C0.385,-0.109 0.429,-0.127 0.462,-0.163C0.495,-0.199 0.511,-0.259 0.511,-0.344C0.511,-0.415 0.497,-0.47 0.469,-0.51C0.441,-0.55 0.397,-0.57 0.337,-0.57Z" style="fill:white;fill-rule:nonzero;"/>
</g>
<g transform="matrix(64,0,0,64,123.424,77.6203)">
<path d="M0.405,-0.288L0.213,-0.288L0.213,0L0.084,0L0.084,-0.68L0.469,-0.68L0.489,-0.578L0.213,-0.578L0.213,-0.389L0.386,-0.389L0.405,-0.288Z" style="fill:white;fill-rule:nonzero;"/>
</g>
</g>
</g>
</g>
<g transform="matrix(1,0,0,1.52217,67.3796,10.7507)">
<rect x="23.501" y="81.3" width="162.305" height="61.77" style="fill:rgb(180,213,255);"/>
</g>
<g transform="matrix(0.967536,0,0,0.961535,5.90498,47.9703)">
<g transform="matrix(90.4804,0,0,90.4804,82.6698,167.705)">
<path d="M0.057,-0.337C0.057,-0.442 0.084,-0.527 0.139,-0.594C0.194,-0.66 0.271,-0.694 0.37,-0.696C0.477,-0.696 0.556,-0.662 0.607,-0.593C0.658,-0.524 0.684,-0.441 0.684,-0.344C0.684,-0.239 0.657,-0.153 0.602,-0.086C0.547,-0.019 0.469,0.014 0.37,0.014C0.264,0.014 0.185,-0.02 0.134,-0.089C0.083,-0.157 0.057,-0.24 0.057,-0.337ZM0.192,-0.338C0.192,-0.267 0.206,-0.208 0.235,-0.163C0.264,-0.118 0.308,-0.095 0.369,-0.095C0.424,-0.095 0.467,-0.115 0.5,-0.156C0.533,-0.197 0.549,-0.259 0.549,-0.344C0.549,-0.415 0.535,-0.473 0.506,-0.518C0.477,-0.563 0.433,-0.586 0.372,-0.586C0.319,-0.586 0.275,-0.564 0.242,-0.519C0.209,-0.474 0.192,-0.414 0.192,-0.338Z" style="fill:rgb(51,51,51);fill-rule:nonzero;"/>
</g>
<g transform="matrix(90.4804,0,0,90.4804,147.906,167.705)">
<path d="M0.505,-0.557C0.473,-0.567 0.448,-0.574 0.429,-0.579C0.41,-0.583 0.388,-0.585 0.361,-0.585C0.307,-0.585 0.265,-0.563 0.236,-0.519C0.207,-0.475 0.192,-0.415 0.192,-0.338C0.192,-0.272 0.204,-0.215 0.229,-0.167C0.254,-0.119 0.295,-0.095 0.353,-0.095C0.382,-0.095 0.409,-0.098 0.434,-0.104C0.459,-0.11 0.481,-0.117 0.502,-0.126L0.551,-0.03C0.525,-0.017 0.494,-0.006 0.457,0.002C0.42,0.01 0.388,0.014 0.36,0.014C0.254,0.014 0.177,-0.02 0.129,-0.088C0.081,-0.156 0.057,-0.239 0.057,-0.337C0.057,-0.442 0.084,-0.527 0.137,-0.594C0.19,-0.661 0.266,-0.694 0.365,-0.694C0.385,-0.694 0.413,-0.691 0.448,-0.684C0.483,-0.677 0.516,-0.666 0.545,-0.65L0.505,-0.557Z" style="fill:rgb(51,51,51);fill-rule:nonzero;"/>
</g>
<g transform="matrix(90.4804,0,0,90.4804,199.751,167.705)">
<path d="M0.293,-0.572L0.213,-0.572L0.213,-0.364L0.295,-0.364C0.334,-0.364 0.366,-0.372 0.391,-0.388C0.416,-0.403 0.429,-0.429 0.429,-0.465C0.429,-0.503 0.417,-0.53 0.393,-0.547C0.369,-0.564 0.336,-0.572 0.293,-0.572ZM0.479,0L0.335,-0.26C0.328,-0.259 0.32,-0.259 0.312,-0.259C0.304,-0.258 0.296,-0.258 0.288,-0.258L0.213,-0.258L0.213,0L0.084,0L0.084,-0.68L0.297,-0.68C0.371,-0.68 0.434,-0.663 0.487,-0.629C0.54,-0.595 0.566,-0.542 0.566,-0.471C0.566,-0.429 0.555,-0.393 0.534,-0.363C0.512,-0.332 0.484,-0.309 0.45,-0.292L0.617,0L0.479,0Z" style="fill:rgb(51,51,51);fill-rule:nonzero;"/>
</g>
</g>
<g transform="matrix(0.916882,0,0,1,121.475,-32.6535)">
<g transform="matrix(86.953,0,0,86.953,152.996,241.878)">
<path d="M0.479,-0.428C0.5,-0.451 0.527,-0.47 0.562,-0.484C0.596,-0.497 0.627,-0.504 0.654,-0.504C0.72,-0.504 0.767,-0.485 0.795,-0.446C0.822,-0.407 0.836,-0.36 0.836,-0.304L0.836,0L0.705,0L0.705,-0.298C0.705,-0.329 0.698,-0.352 0.683,-0.369C0.668,-0.385 0.647,-0.393 0.619,-0.393C0.6,-0.393 0.581,-0.388 0.56,-0.378C0.539,-0.368 0.521,-0.357 0.504,-0.344C0.505,-0.337 0.505,-0.331 0.506,-0.324C0.507,-0.317 0.507,-0.311 0.507,-0.304L0.507,0L0.376,0L0.376,-0.298C0.376,-0.329 0.369,-0.352 0.354,-0.369C0.339,-0.385 0.318,-0.393 0.291,-0.393C0.274,-0.393 0.258,-0.39 0.241,-0.383C0.224,-0.376 0.207,-0.367 0.192,-0.356L0.192,0L0.062,0L0.062,-0.485L0.13,-0.485L0.162,-0.441C0.184,-0.461 0.211,-0.476 0.242,-0.488C0.273,-0.499 0.3,-0.504 0.325,-0.504C0.363,-0.504 0.395,-0.497 0.42,-0.484C0.445,-0.47 0.465,-0.451 0.479,-0.428Z" style="fill:rgb(51,51,51);fill-rule:nonzero;"/>
</g>
<g transform="matrix(86.953,0,0,86.953,228.906,241.878)">
<path d="M0.156,0.023L0.179,-0.034L0.006,-0.467L0.14,-0.485L0.252,-0.191L0.358,-0.485L0.495,-0.485L0.278,0.064C0.263,0.103 0.236,0.137 0.197,0.165C0.158,0.193 0.118,0.212 0.075,0.222L0.029,0.115C0.052,0.105 0.077,0.093 0.104,0.079C0.13,0.064 0.147,0.046 0.156,0.023Z" style="fill:rgb(51,51,51);fill-rule:nonzero;"/>
</g>
</g>
<g id="Selectors" transform="matrix(0.965977,0,0,0.807602,67.3796,67.3718)">
<g id="Right-selector" serif:id="Right selector">
<g transform="matrix(1.03522,0,0,1.23823,2.07044,0)">
<path d="M185.806,161.156L185.806,67.132" style="fill:none;stroke:rgb(76,159,255);stroke-width:4px;stroke-linecap:butt;"/>
</g>
<g transform="matrix(1.03522,0,0,1.23823,161.788,169.469)">
<circle cx="31.523" cy="34.314" r="10.021" style="fill:rgb(76,159,255);stroke:rgb(76,159,255);stroke-width:4px;stroke-linecap:butt;"/>
</g>
</g>
<g id="Left-selector" serif:id="Left selector">
<g transform="matrix(1.03522,0,0,1.23823,-170.092,0)">
<path d="M185.806,161.156L185.806,67.132" style="fill:none;stroke:rgb(76,159,255);stroke-width:4px;stroke-linecap:butt;"/>
</g>
<g transform="matrix(1.03522,0,0,1.23823,-10.3742,28.2274)">
<circle cx="31.523" cy="34.314" r="10.021" style="fill:rgb(76,159,255);stroke:rgb(76,159,255);stroke-width:4px;stroke-linecap:butt;"/>
</g>
</g>
</g>
</g>
</g>
</svg>
Side by Side

After

Width:  |  Height:  |  Size: 9.3 KiB

docs/index.rst
+1 -5
View File
@@ -1,8 +1,3 @@
.. ocrmypdf documentation master file, created by
sphinx-quickstart on Sun Sep 4 14:29:43 2016.
You can adapt this file completely to your liking, but it should at least
contain the root `toctree` directive.
OCRmyPDF documentation
======================
@@ -27,6 +22,7 @@ PDF is the best format for storing and exchanging scanned documents. Unfortunat
cookbook
docker
advanced
api
batch
security
errors
docs/installation.rst
+186 -112
View File
@@ -1,3 +1,4 @@
===================
Installing OCRmyPDF
===================
@@ -18,10 +19,10 @@ installing the Python binary wheels.
:local:
Installing on Linux
-------------------
===================
Debian and Ubuntu 16.10 or newer
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
--------------------------------
.. |deb-stable| image:: https://repology.org/badge/version-for-repo/debian_stable/ocrmypdf.svg
:alt: Debian 9 stable ("stretch")
@@ -52,22 +53,33 @@ Debian and Ubuntu 16.10 or newer
| |ubu-1710| |ubu-1804| |ubu-1810| |
+-------------------------------------------+
Users of Debian 9 ("stretch") or later or Ubuntu 16.10 or later may simply
Users of Debian 9 ("stretch") or later or Ubuntu 16.10 or later may
simply
.. code-block:: bash
apt-get install ocrmypdf
As indicated in the table above, Debian and Ubuntu releases may lag behind the latest version. If the version available for your platform is out of date, you could opt to install the latest version from source. See `Installing HEAD revision from sources`_.
As indicated in the table above, Debian and Ubuntu releases may lag
behind the latest version. If the version available for your platform is
out of date, you could opt to install the latest version from source.
See `Installing HEAD revision from
sources <#installing-head-revision-from-sources>`__.
For full details on version availability for your platform, check the `Debian Package Tracker <https://tracker.debian.org/pkg/ocrmypdf>`_ or `Ubuntu launchpad.net <https://launchpad.net/ocrmypdf>`_.
For full details on version availability for your platform, check the
`Debian Package Tracker <https://tracker.debian.org/pkg/ocrmypdf>`__ or
`Ubuntu launchpad.net <https://launchpad.net/ocrmypdf>`__.
.. note::
OCRmyPDF for Debian and Ubuntu currently omit the JBIG2 encoder. OCRmyPDF works fine without it but will produce larger output files. If you build jbig2enc from source, ocrmypdf 7.0.0 and later will automatically detect it (specifically the ``jbig2`` binary) on the ``PATH``. To add JBIG2 encoding, see :ref:`jbig2`.
OCRmyPDF for Debian and Ubuntu currently omit the JBIG2 encoder.
OCRmyPDF works fine without it but will produce larger output files.
If you build jbig2enc from source, ocrmypdf 7.0.0 and later will
automatically detect it (specifically the ``jbig2`` binary) on the
``PATH``. To add JBIG2 encoding, see :ref:`jbig2`.
Fedora 29 or newer
^^^^^^^^^^^^^^^^^^
------------------
.. |fedora-29| image:: https://repology.org/badge/version-for-repo/fedora29/ocrmypdf.svg
:alt: Fedora 29
@@ -90,28 +102,28 @@ Users of Fedora 29 later may simply
dnf install ocrmypdf
For full details on version availability, check the `Fedora Package Tracker
<https://apps.fedoraproject.org/packages/ocrmypdf>`_.
For full details on version availability, check the `Fedora Package
Tracker <https://apps.fedoraproject.org/packages/ocrmypdf>`__.
If the version available for your platform is out of date, you could opt to
install the latest version from source. See `Installing HEAD revision from
sources`_.
If the version available for your platform is out of date, you could opt
to install the latest version from source. See `Installing HEAD revision
from sources <#installing-head-revision-from-sources>`__.
.. note::
OCRmyPDF for Fedora currently omits the JBIG2 encoder due to patent issues.
OCRmyPDF works fine without it but will produce larger output files. If you
build jbig2enc from source, ocrmypdf 7.0.0 and later will automatically
detect it on the ``PATH``. To add JBIG2 encoding, see `Installing the JBIG2
encoder <jbig2>`_.
OCRmyPDF for Fedora currently omits the JBIG2 encoder due to patent
issues. OCRmyPDF works fine without it but will produce larger output
files. If you build jbig2enc from source, ocrmypdf 7.0.0 and later
will automatically detect it on the ``PATH``. To add JBIG2 encoding,
see `Installing the JBIG2 encoder <jbig2>`__.
.. _ubuntu-lts-latest:
Installing the latest version on Ubuntu 18.04 LTS
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-------------------------------------------------
Ubuntu 18.04 includes ocrmypdf 6.1.2. To install a more recent version, first
install the system version to get most of the dependencies:
Ubuntu 18.04 includes ocrmypdf 6.1.2. To install a more recent version,
first install the system version to get most of the dependencies:
.. code-block:: bash
@@ -120,16 +132,17 @@ install the system version to get most of the dependencies:
ocrmypdf \
python3-pip
There are a few dependency changes between ocrmypdf 6.1.2 and 7.x. Let's get
these, too.
There are a few system dependency changes since ocrmypdf 6.1.2. Let's
get these, too.
.. code-block:: bash
sudo apt-get install \
libexempi3 \
libxml2 \
pngquant
Then install the most recent ocrmypdf for the local user and set the user's ``PATH`` to check for the user's Python packages.
Then install the most recent ocrmypdf for the local user and set the
user's ``PATH`` to check for the user's Python packages.
.. code-block:: bash
@@ -139,13 +152,13 @@ Then install the most recent ocrmypdf for the local user and set the user's ``PA
To add JBIG2 encoding, see :ref:`jbig2`.
Ubuntu 16.04 LTS
^^^^^^^^^^^^^^^^
----------------
No package is available for Ubuntu 16.04. OCRmyPDF 8.0 and newer require Python
3.6. Ubuntu 16.04 ships Python 3.5, but you can install Python 3.6 on it. Or,
you can skip Python 3.6 and install OCRmyPDF 7.x or older - for that procedure,
please see the installation documentation for the version of OCRmyPDF you plan
to use.
No package is available for Ubuntu 16.04. OCRmyPDF 8.0 and newer require
Python 3.6. Ubuntu 16.04 ships Python 3.5, but you can install Python
3.6 on it. Or, you can skip Python 3.6 and install OCRmyPDF 7.x or older
- for that procedure, please see the installation documentation for the
version of OCRmyPDF you plan to use.
**Install system packages for OCRmyPDF**
@@ -167,13 +180,13 @@ to use.
tesseract-ocr \
unpaper
This will install a Python 3.6 binary at ``/usr/bin/python3.6`` alongside the
system's Python 3.5. Do not remove the system Python. This will also install
Tesseract 4.0 from a PPA, since the version available in Ubuntu 16.04 is too old
for OCRmyPDF.
This will install a Python 3.6 binary at ``/usr/bin/python3.6``
alongside the system's Python 3.5. Do not remove the system Python. This
will also install Tesseract 4.0 from a PPA, since the version available
in Ubuntu 16.04 is too old for OCRmyPDF.
Now install pip for Python 3.6. This will install the Python 3.6 version of
``pip`` at ``/usr/local/bin/pip``.
Now install pip for Python 3.6. This will install the Python 3.6 version
of ``pip`` at ``/usr/local/bin/pip``.
.. code-block:: bash
@@ -181,8 +194,8 @@ Now install pip for Python 3.6. This will install the Python 3.6 version of
**Install OCRmyPDF**
OCRmyPDF requires the locale to be set for UTF-8. **On some minimal Ubuntu
installations systems**, it may be necessary to set the locale.
OCRmyPDF requires the locale to be set for UTF-8. **On some minimal
Ubuntu installations systems**, it may be necessary to set the locale.
.. code-block:: bash
@@ -201,11 +214,12 @@ environment variable contains ``$HOME/.local/bin``.
To add JBIG2 encoding, see :ref:`jbig2`.
Ubuntu 14.04 LTS
^^^^^^^^^^^^^^^^
----------------
Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other
options, because of its age. Several backports are required. For explanations of
some steps of this procedure, see the similar steps for Ubuntu 16.04.
Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some
other options, because of its age. Several backports are required. For
explanations of some steps of this procedure, see the similar steps for
Ubuntu 16.04.
Install system dependencies:
@@ -222,12 +236,12 @@ Install system dependencies:
qpdf
We will need backports of Ghostscript 9.16, libav-11 (for unpaper 6.1),
Tesseract 4.00 (alpha), and Python 3.6. This will replace Ghostscript and
Tesseract 3.x on your system. Python 3.6 will be installed alongside the system
Python 3.4.
Tesseract 4.00 (alpha), and Python 3.6. This will replace Ghostscript
and Tesseract 3.x on your system. Python 3.6 will be installed alongside
the system Python 3.4.
If you prefer to not modify your system in this matter, consider using a Docker
container.
If you prefer to not modify your system in this matter, consider using a
Docker container.
.. code-block:: bash
@@ -253,7 +267,10 @@ Now we need to install ``pip`` and let it install ocrmypdf:
curl https://bootstrap.pypa.io/ez_setup.py -o - | python3.6 && python3.6 -m easy_install pip
pip3.6 install ocrmypdf
These installation instructions omit the optional dependency ``unpaper``, which is only available at version 0.4.2 in Ubuntu 14.04. The author could not find a backport of ``unpaper``, and created a .deb package to do the job of installing unpaper 6.1 (for x86 64-bit only):
These installation instructions omit the optional dependency
``unpaper``, which is only available at version 0.4.2 in Ubuntu 14.04.
The author could not find a backport of ``unpaper``, and created a .deb
package to do the job of installing unpaper 6.1 (for x86 64-bit only):
.. code-block:: bash
@@ -263,44 +280,52 @@ These installation instructions omit the optional dependency ``unpaper``, which
To add JBIG2 encoding, see :ref:`jbig2`.
ArchLinux (AUR)
^^^^^^^^^^^^^^^
---------------
.. image:: https://repology.org/badge/version-for-repo/aur/ocrmypdf.svg
:alt: ArchLinux
:target: https://repology.org/metapackage/ocrmypdf
There is an `ArchLinux User Repository package for ocrmypdf <https://aur.archlinux.org/packages/ocrmypdf/>`_. You can use the following command.
There is an `ArchLinux User Repository package for
ocrmypdf <https://aur.archlinux.org/packages/ocrmypdf/>`__. You can use
the following command.
.. code-block:: bash
yaourt -S ocrmypdf
If you have any difficulties with installation, check the repository package page.
If you have any difficulties with installation, check the repository
package page.
Other Linux packages
^^^^^^^^^^^^^^^^^^^^
--------------------
See the `Repology <https://repology.org/metapackage/ocrmypdf/versions>`_ page.
See the
`Repology <https://repology.org/metapackage/ocrmypdf/versions>`__ page.
In general, first install the OCRmyPDF package for your system, then optionally use the procedure `Installing with Python pip`_ to install a more recent version.
In general, first install the OCRmyPDF package for your system, then
optionally use the procedure `Installing with Python
pip <#installing-with-python-pip>`__ to install a more recent version.
Installing on macOS
-------------------
===================
Homebrew
^^^^^^^^
--------
.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg
:alt: homebrew
:target: http://brewformulas.org/Ocrmypdf
OCRmyPDF is now a standard `Homebrew <https://brew.sh>`_ formula. To install on macOS:
OCRmyPDF is now a standard `Homebrew <https://brew.sh>`__ formula. To
install on macOS:
.. code-block:: bash
brew install ocrmypdf
This will include only the English language pack. If you need other languages you can optionally install them all:
This will include only the English language pack. If you need other
languages you can optionally install them all:
.. code-block:: bash
@@ -308,18 +333,23 @@ This will include only the English language pack. If you need other languages yo
.. note::
Users who previously installed OCRmyPDF on macOS using ``pip install ocrmypdf`` should remove the pip version (``pip3 uninstall ocrmypdf``) before switching to the Homebrew version.
Users who previously installed OCRmyPDF on macOS using
``pip install ocrmypdf`` should remove the pip version
(``pip3 uninstall ocrmypdf``) before switching to the Homebrew
version.
.. note::
Users who previously installed OCRmyPDF from the private tap should switch to the mainline version (``brew untap jbarlow83/ocrmypdf``) and install from there.
Users who previously installed OCRmyPDF from the private tap should
switch to the mainline version (``brew untap jbarlow83/ocrmypdf``)
and install from there.
Manual installation on macOS
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
----------------------------
These instructions probably work on all macOS supported by Homebrew.
If it's not already present, `install Homebrew <http://brew.sh/>`_.
If it's not already present, `install Homebrew <http://brew.sh/>`__.
Update Homebrew:
@@ -327,18 +357,22 @@ Update Homebrew:
brew update
Install or upgrade the required Homebrew packages, if any are missing. To do this, download the ``Brewfile`` that lists all of the dependencies to the current directory, and run ``brew bundle`` to process them (installing or upgrading as needed). ``Brewfile`` is a plain text file.
Install or upgrade the required Homebrew packages, if any are missing.
To do this, download the ``Brewfile`` that lists all of the dependencies
to the current directory, and run ``brew bundle`` to process them
(installing or upgrading as needed). ``Brewfile`` is a plain text file.
.. code-block:: bash
wget https://github.com/jbarlow83/OCRmyPDF/raw/master/.travis/Brewfile
brew bundle
This will include the English, French, German and Spanish language packs. If you need other languages you can optionally install them all:
This will include the English, French, German and Spanish language
packs. If you need other languages you can optionally install them all:
.. _macos-all-languages:
.. code-block:: bash
.. code-block:: bash
brew install tesseract --with-all-languages # Option 2: for all language packs
@@ -367,24 +401,28 @@ The command line program should now be available:
ocrmypdf --help
Installing on FreeBSD
---------------------
=====================
FreeBSD 11.2 is known to work. Other versions likely work but have not been tested.
FreeBSD 11.2 is known to work. Other versions likely work but have not
been tested.
In general it should work to:
#. `Install and build pikepdf <https://pikepdf.readthedocs.io/en/latest/installation.html#installing-on-freebsd-11-2>`_.
#. `Install and build
pikepdf <https://pikepdf.readthedocs.io/en/latest/installation.html#installing-on-freebsd-11-2>`__.
#. Install the equivalent list of dependencies for Linux.
Installing the Docker image
---------------------------
===========================
For some users, installing the Docker image will be easier than installing all of OCRmyPDF's dependencies. For Windows, it is the only option.
For some users, installing the Docker image will be easier than
installing all of OCRmyPDF's dependencies. For Windows, it is the only
option.
See `OCRmyPDF Docker Image <docker>`_ for more information.
See `OCRmyPDF Docker Image <docker>`__ for more information.
Installing on Windows
---------------------
=====================
Direct installation on Windows is not currently possible, but it works well in
Windows Subsystem for Linux:
@@ -422,79 +460,113 @@ You can also :ref:`Install the Docker <docker-install>` container on Windows. En
your command prompt can run the docker "hello world" container.
Installing with Python pip
--------------------------
==========================
OCRmyPDF is delivered by PyPI because it is a convenient way to install the latest version. However, PyPI and ``pip`` cannot address the fact that ``ocrmypdf`` depends on certain non-Python system libraries and programs being instsalled.
OCRmyPDF is delivered by PyPI because it is a convenient way to install
the latest version. However, PyPI and ``pip`` cannot address the fact
that ``ocrmypdf`` depends on certain non-Python system libraries and
programs being instsalled.
For best results, first install `your platform's version <https://repology.org/metapackage/ocrmypdf/versions>`_ of ``ocrmypdf``, using the instructions elsewhere in this document. Then you can use ``pip`` to get the latest version if your platform version is out of date. Chances are that this will satisfy most dependencies.
For best results, first install `your platform's
version <https://repology.org/metapackage/ocrmypdf/versions>`__ of
``ocrmypdf``, using the instructions elsewhere in this document. Then
you can use ``pip`` to get the latest version if your platform version
is out of date. Chances are that this will satisfy most dependencies.
Use ``ocrmypdf --version`` to confirm what version was installed.
Then you can install the latest OCRmyPDF from the Python wheels. First try:
Then you can install the latest OCRmyPDF from the Python wheels. First
try:
.. code-block:: bash
pip3 install --user ocrmypdf
You should then be able to run ``ocrmypdf --version`` and see that the latest version was located.
You should then be able to run ``ocrmypdf --version`` and see that the
latest version was located.
Since ``pip3 install --user`` does not work correctly on some platforms, notably Ubuntu 16.04 and older, and the Homebrew version of Python, instead use this for a system wide installation:
Since ``pip3 install --user`` does not work correctly on some platforms,
notably Ubuntu 16.04 and older, and the Homebrew version of Python,
instead use this for a system wide installation:
.. code-block:: bash
pip3 install ocrmypdf
Requirements for pip and HEAD install
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-------------------------------------
OCRmyPDF currently requires these external programs and libraries to be installed, and must be satisfied using the operating system package manager. ``pip`` cannot provide them.
OCRmyPDF currently requires these external programs and libraries to be
installed, and must be satisfied using the operating system package
manager. ``pip`` cannot provide them.
- Python 3.6 or newer
- Ghostscript 9.15 or newer
- qpdf 8.1.0 or newer
- Tesseract 4.0.0-alpha or newer
- Python 3.6 or newer
- Ghostscript 9.15 or newer
- qpdf 8.1.0 or newer
- Tesseract 4.0.0-alpha or newer
As of ocrmypdf 7.2.1, the following versions are recommended:
- Python 3.7
- Ghostscript 9.23 or newer
- qpdf 8.2.1
- Tesseract 4.0.0 or newer
- jbig2enc 0.29 or newer
- pngquant 2.5 or newer
- unpaper 6.1
- Python 3.7
- Ghostscript 9.23 or newer
- qpdf 8.2.1
- Tesseract 4.0.0 or newer
- jbig2enc 0.29 or newer
- pngquant 2.5 or newer
- unpaper 6.1
jbig2enc, pngquant, and unpaper are optional. If missing certain features are disabled. OCRmyPDF will discover them as soon as they are available.
jbig2enc, pngquant, and unpaper are optional. If missing certain
features are disabled. OCRmyPDF will discover them as soon as they are
available.
**jbig2enc**, if present, will be used to optimize the encoding of monochrome images. This can significantly reduce the file size of the output file. It is not required. `jbig2enc <https://github.com/agl/jbig2enc>`_ is not generally available for Ubuntu or Debian due to lingering concerns about patent issues, but can easily be built from source. To add JBIG2 encoding, see :ref:`jbig2`.
**jbig2enc**, if present, will be used to optimize the encoding of
monochrome images. This can significantly reduce the file size of the
output file. It is not required.
`jbig2enc <https://github.com/agl/jbig2enc>`__ is not generally
available for Ubuntu or Debian due to lingering concerns about patent
issues, but can easily be built from source. To add JBIG2 encoding, see
:ref:`jbig2`.
**pngquant**, if present, is optionally used to optimize the encoding of PNG-style images in PDFs (actually, any that are that losslessly encoded) by lossily quantizing to a smaller color palette. It is only activated then the ``--optimize`` argument is ``2`` or ``3``.
**pngquant**, if present, is optionally used to optimize the encoding of
PNG-style images in PDFs (actually, any that are that losslessly
encoded) by lossily quantizing to a smaller color palette. It is only
activated then the ``--optimize`` argument is ``2`` or ``3``.
**unpaper**, if present, enables the ``--clean`` and ``--clean-final`` command line options.
These are in addition to the Python packaging dependencies, meaning that unfortunately, the ``pip install`` command cannot satisfy all of them.
**unpaper**, if present, enables the ``--clean`` and ``--clean-final``
command line options.
These are in addition to the Python packaging dependencies, meaning that
unfortunately, the ``pip install`` command cannot satisfy all of them.
Installing HEAD revision from sources
-------------------------------------
=====================================
If you have ``git`` and Python 3.6 or newer installed, you can install from source. When the ``pip`` installer runs, it will alert you if dependencies are missing.
If you have ``git`` and Python 3.6 or newer installed, you can install
from source. When the ``pip`` installer runs, it will alert you if
dependencies are missing.
If you prefer to build every from source, you will need to `build pikepdf from source <https://pikepdf.readthedocs.io/en/latest/installation.html#building-from-source>`_. First ensure you can build and install pikepdf.
If you prefer to build every from source, you will need to `build
pikepdf from
source <https://pikepdf.readthedocs.io/en/latest/installation.html#building-from-source>`__.
First ensure you can build and install pikepdf.
To install the HEAD revision from sources in the current Python 3 environment:
To install the HEAD revision from sources in the current Python 3
environment:
.. code-block:: bash
pip3 install git+https://github.com/jbarlow83/OCRmyPDF.git
Or, to install in `development mode <https://pythonhosted.org/setuptools/setuptools.html#development-mode>`_, allowing customization of OCRmyPDF, use the ``-e`` flag:
Or, to install in `development
mode <https://pythonhosted.org/setuptools/setuptools.html#development-mode>`__,
allowing customization of OCRmyPDF, use the ``-e`` flag:
.. code-block:: bash
pip3 install -e git+https://github.com/jbarlow83/OCRmyPDF.git
You may find it easiest to install in a virtual environment, rather than system-wide:
You may find it easiest to install in a virtual environment, rather than
system-wide:
.. code-block:: bash
@@ -504,8 +576,8 @@ You may find it easiest to install in a virtual environment, rather than system-
cd OCRmyPDF
pip3 install .
However, ``ocrmypdf`` will only be accessible on the system PATH
when you activate the virtual environment.
However, ``ocrmypdf`` will only be accessible on the system PATH when
you activate the virtual environment.
To run the program:
@@ -519,7 +591,7 @@ dependencies. Older version than the ones mentioned in the release notes
are likely not to be compatible to OCRmyPDF.
For development
^^^^^^^^^^^^^^^
---------------
To install all of the development and test requirements:
@@ -535,15 +607,17 @@ To install all of the development and test requirements:
To add JBIG2 encoding, see :ref:`jbig2`.
Shell completions
-----------------
=================
Completions for ``bash`` and ``fish`` are available in the project's
``misc/completion`` folder. The ``bash`` completions are likely ``zsh``
compatible but this has not been confirmed. Package maintainers, please install
these at the appropriate locations for your system.
compatible but this has not been confirmed. Package maintainers, please
install these at the appropriate locations for your system.
To manually install the ``bash`` completion, copy ``misc/completion/ocrmypdf.bash`` to
``/etc/bash_completion.d/ocrmypdf`` (rename the file).
To manually install the ``bash`` completion, copy
``misc/completion/ocrmypdf.bash`` to ``/etc/bash_completion.d/ocrmypdf``
(rename the file).
To manually install the ``fish`` completion, copy ``misc/completion/ocrmypdf.fish`` to
To manually install the ``fish`` completion, copy
``misc/completion/ocrmypdf.fish`` to
``~/.config/fish/completions/ocrmypdf.fish``.
docs/introduction.rst
+166 -57
View File
@@ -1,119 +1,228 @@
============
Introduction
============
OCRmyPDF is a Python 3 package that adds OCR layers to PDFs.
About OCR
---------
=========
`Optical character recognition <https://en.wikipedia.org/wiki/Optical_character_recognition>`_ is technology that converts images of typed or handwritten text, such as in a scanned document, to computer text that can be searched and copied.
`Optical character
recognition <https://en.wikipedia.org/wiki/Optical_character_recognition>`__
is technology that converts images of typed or handwritten text, such as
in a scanned document, to computer text that can be searched and copied.
OCRmyPDF uses `Tesseract <https://github.com/tesseract-ocr/tesseract>`_, the best available open source OCR engine, to perform OCR.
OCRmyPDF uses
`Tesseract <https://github.com/tesseract-ocr/tesseract>`__, the best
available open source OCR engine, to perform OCR.
.. _raster-vector:
About PDFs
----------
==========
PDFs are page description files that attempts to preserve a layout exactly. They contain `vector graphics <http://vector-conversions.com/vectorizing/raster_vs_vector.html>`_ that can contain raster objects such as scanned images. Because PDFs can contain multiple pages (unlike many image formats) and can contain fonts and text, it is a good formats for exchanging scanned documents.
PDFs are page description files that attempts to preserve a layout
exactly. They contain `vector
graphics <http://vector-conversions.com/vectorizing/raster_vs_vector.html>`__
that can contain raster objects such as scanned images. Because PDFs can
contain multiple pages (unlike many image formats) and can contain fonts
and text, it is a good formats for exchanging scanned documents.
.. image:: images/bitmap_vs_svg.svg
|image|
A PDF page might contain multiple images, even if it only appears to have one image. Some scanners or scanning software will segment pages into monochromatic text and color regions for example, to improve the compression ratio and appearance of the page.
Rasterizing a PDF is the process of generating an image suitable for display or analyzing with an OCR engine. OCR engines like Tesseract work with images, not vector objects.
A PDF page might contain multiple images, even if it only appears to
have one image. Some scanners or scanning software will segment pages
into monochromatic text and color regions for example, to improve the
compression ratio and appearance of the page.
Rasterizing a PDF is the process of generating an image suitable for
display or analyzing with an OCR engine. OCR engines like Tesseract work
with images, not vector objects.
About PDF/A
-----------
===========
`PDF/A <https://en.wikipedia.org/wiki/PDF/A>`_ is an ISO-standardized subset of the full PDF specification that is designed for archiving (the 'A' stands for Archive). PDF/A differs from PDF primarily by omitting features that would make it difficult to read the file in the future, such as embedded Javascript, video, audio and references to external fonts. All fonts and resources needed to interpret the PDF must be contained within it. Because PDF/A disables Javascript and other types of embedded content, it is probably more secure.
`PDF/A <https://en.wikipedia.org/wiki/PDF/A>`__ is an ISO-standardized
subset of the full PDF specification that is designed for archiving (the
'A' stands for Archive). PDF/A differs from PDF primarily by omitting
features that would make it difficult to read the file in the future,
such as embedded Javascript, video, audio and references to external
fonts. All fonts and resources needed to interpret the PDF must be
contained within it. Because PDF/A disables Javascript and other types
of embedded content, it is probably more secure.
There are various conformance levels and versions, such as "PDF/A-2b".
Generally speaking, the best format for scanned documents is PDF/A. Some governments and jurisdictions, US Courts in particular, `mandate the use of PDF/A <https://pdfblog.com/2012/02/13/what-is-pdfa/>`_ for scanned documents.
Generally speaking, the best format for scanned documents is PDF/A. Some
governments and jurisdictions, US Courts in particular, `mandate the use
of PDF/A <https://pdfblog.com/2012/02/13/what-is-pdfa/>`__ for scanned
documents.
Since most people who scan documents are interested in reading them indefinitely into the future, OCRmyPDF generates PDF/A-2b by default.
PDF/A has a few drawbacks. Some PDF viewers include an alert that the file is a PDF/A, which may confuse some users. It also tends to produce larger files than PDF, because it embeds certain resources even if they are commonly available. PDF/A files can be digitally signed, but may not be encrypted, to ensure they can be read in the future. Fortunately, converting from PDF/A to a regular PDF is trivial, and any PDF viewer can view PDF/A.
Since most people who scan documents are interested in reading them
indefinitely into the future, OCRmyPDF generates PDF/A-2b by default.
PDF/A has a few drawbacks. Some PDF viewers include an alert that the
file is a PDF/A, which may confuse some users. It also tends to produce
larger files than PDF, because it embeds certain resources even if they
are commonly available. PDF/A files can be digitally signed, but may not
be encrypted, to ensure they can be read in the future. Fortunately,
converting from PDF/A to a regular PDF is trivial, and any PDF viewer
can view PDF/A.
What OCRmyPDF does
------------------
==================
OCRmyPDF analyzes each page of a PDF to determine the colorspace and resolution (DPI) needed to capture all of the information on that page without losing content. It uses `Ghostscript <http://ghostscript.com/>`_ to rasterize the page, and then performs on OCR on the rasterized image to create an OCR "layer". The layer is then grafted back onto the original PDF.
OCRmyPDF analyzes each page of a PDF to determine the colorspace and
resolution (DPI) needed to capture all of the information on that page
without losing content. It uses
`Ghostscript <http://ghostscript.com/>`__ to rasterize the page, and
then performs on OCR on the rasterized image to create an OCR "layer".
The layer is then grafted back onto the original PDF.
While one can use a program like Ghostscript or ImageMagick to get an image and put the image through Tesseract, that actually creates a new PDF and many details may be lost. OCRmyPDF can produce a minimally changed PDF as output.
While one can use a program like Ghostscript or ImageMagick to get an
image and put the image through Tesseract, that actually creates a new
PDF and many details may be lost. OCRmyPDF can produce a minimally
changed PDF as output.
OCRmyPDF also some image processing options like deskew which improve the appearance of files and quality of OCR. When these are used, the OCR layer is grafted onto the processed image instead.
By default, OCRmyPDF produces archival PDFs – PDF/A, which are a stricter subset of PDF features designed for long term archives. If regular PDFs are desired, this can be disabled with ``--output-type pdf``.
OCRmyPDF also some image processing options like deskew which improve
the appearance of files and quality of OCR. When these are used, the OCR
layer is grafted onto the processed image instead.
By default, OCRmyPDF produces archival PDFs – PDF/A, which are a
stricter subset of PDF features designed for long term archives. If
regular PDFs are desired, this can be disabled with
``--output-type pdf``.
Why you shouldn't do this manually
----------------------------------
==================================
A PDF is similar to an HTML file, in that it contains document structure along with images. Sometimes a PDF does nothing more than present a full page image, but often there is additional content that would be lost.
A PDF is similar to an HTML file, in that it contains document structure
along with images. Sometimes a PDF does nothing more than present a full
page image, but often there is additional content that would be lost.
A manual process could work like either of these:
1. Rasterize each page as an image, OCR the images, and combine the output into a PDF. This preserves the layout of each page, but resamples all images (possibly losing quality, increasing file size, introducing compression artifacts, etc.).
1. Rasterize each page as an image, OCR the images, and combine the
output into a PDF. This preserves the layout of each page, but
resamples all images (possibly losing quality, increasing file size,
introducing compression artifacts, etc.).
2. Extract each image, OCR, and combine the output into a PDF. This
loses the context in which images are used in the PDF, meaning that
cropping, rotation and scaling of pages may be lost. Some scanned
PDFs use multiple images segmented into black and white, grayscale
and color regions, with stencil masks to prevent overlap, as this can
enhance the appearance of a file while reducing file size. Clearly,
reassembling these images will be easy. This also loses and text or
vector art on any pages in a PDF with both scanned and pure digital
content.
2. Extract each image, OCR, and combine the output into a PDF. This loses the context in which images are used in the PDF, meaning that cropping, rotation and scaling of pages may be lost. Some scanned PDFs use multiple images segmented into black and white, grayscale and color regions, with stencil masks to prevent overlap, as this can enhance the appearance of a file while reducing file size. Clearly, reassembling these images will be easy. This also loses and text or vector art on any pages in a PDF with both scanned and pure digital content.
In the case of a PDF that is nothing other than a container of images
(no rotation, scaling, cropping, one image per page), the second
approach can be lossless.
In the case of a PDF that is nothing other than a container of images (no rotation, scaling, cropping, one image per page), the second approach can be lossless.
OCRmyPDF uses several strategies depending on input options and the input PDF itself, but generally speaking it rasterizes a page for OCR and then grafts the OCR back onto the original. As such it can handle complex PDFs and still preserve their contents as much as possible.
OCRmyPDF also supports a many, many edge cases that have cropped over several years of development. We support PDF features like images inside of Form XObjects, and pages with UserUnit scaling. We support rare image formats like non-monochrome 1-bit images. We warn about files you may not to OCR. Thanks to pikepdf and QPDF, we auto-repair PDFs that are damaged. (Not that you need to know what any of these are! You should be able to throw any PDF at it.)
OCRmyPDF uses several strategies depending on input options and the
input PDF itself, but generally speaking it rasterizes a page for OCR
and then grafts the OCR back onto the original. As such it can handle
complex PDFs and still preserve their contents as much as possible.
OCRmyPDF also supports a many, many edge cases that have cropped over
several years of development. We support PDF features like images inside
of Form XObjects, and pages with UserUnit scaling. We support rare image
formats like non-monochrome 1-bit images. We warn about files you may
not to OCR. Thanks to pikepdf and QPDF, we auto-repair PDFs that are
damaged. (Not that you need to know what any of these are! You should be
able to throw any PDF at it.)
Limitations
-----------
===========
OCRmyPDF is limited by the Tesseract OCR engine. As such it experiences these limitations, as do any other programs that rely on Tesseract:
OCRmyPDF is limited by the Tesseract OCR engine. As such it experiences
these limitations, as do any other programs that rely on Tesseract:
* The OCR is not as accurate as commercial solutions such as Abbyy.
* It is not capable of recognizing handwriting.
* It may find gibberish and report this as OCR output.
* If a document contains languages outside of those given in the ``-l LANG`` arguments, results may be poor.
* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns, and may try to join text across columns.
* Poor quality scans may produce poor quality OCR. Garbage in, garbage out.
* It does not expose information about what font family text belongs to.
- The OCR is not as accurate as commercial solutions such as Abbyy.
- It is not capable of recognizing handwriting.
- It may find gibberish and report this as OCR output.
- If a document contains languages outside of those given in the
``-l LANG`` arguments, results may be poor.
- It is not always good at analyzing the natural reading order of
documents. For example, it may fail to recognize that a document
contains two columns, and may try to join text across columns.
- Poor quality scans may produce poor quality OCR. Garbage in, garbage
out.
- It does not expose information about what font family text belongs
to.
OCRmyPDF is also limited by the PDF specification:
* PDF encodes the position of text glyphs but does not encode document structure. There is no markup that divides a document in sections, paragraphs, sentences, or even words (since blank spaces are not represented). As such all elements of document structure including the spaces between words must be derived heuristically. Some PDF viewers do a better job of this than others.
* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround (when using ``--pdf-renderer hocr``). While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones.
- PDF encodes the position of text glyphs but does not encode document
structure. There is no markup that divides a document in sections,
paragraphs, sentences, or even words (since blank spaces are not
represented). As such all elements of document structure including
the spaces between words must be derived heuristically. Some PDF
viewers do a better job of this than others.
- Because some popular open source PDF viewers have a particularly hard
time with spaces betweem words, OCRmyPDF appends a space to each text
element as a workaround (when using ``--pdf-renderer hocr``). While
this mixes document structure with graphical information that ideally
should be left to the PDF viewer to interpret, it improves
compatibility with some viewers and does not cause problems for
better ones.
Ghostscript also imposes some limitations:
* PDFs containing JBIG2-encoded content will be converted to CCITT Group4 encoding, which has lower compression ratios, if Ghostscript PDF/A is enabled.
* PDFs containing JPEG 2000-encoded content will be converted to JPEG encoding, which may introduce compression artifacts, if Ghostscript PDF/A is enabled.
* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format. (Ghostscript 9.25+ can copy JPEG images without transcoding them; earlier versions will transcode.)
* Ghostscript's PDF/A conversion removes any XMP metadata that is not one of the standard XMP metadata namespaces for PDFs. In particular, PRISM Metdata is removed.
- PDFs containing JBIG2-encoded content will be converted to CCITT
Group4 encoding, which has lower compression ratios, if Ghostscript
PDF/A is enabled.
- PDFs containing JPEG 2000-encoded content will be converted to JPEG
encoding, which may introduce compression artifacts, if Ghostscript
PDF/A is enabled.
- Ghostscript may transcode grayscale and color images, either lossy to
lossless or lossless to lossy, based on an internal algorithm. This
behavior can be suppressed by setting ``--pdfa-image-compression`` to
``jpeg`` or ``lossless`` to set all images to one type or the other.
Ghostscript has no option to maintain the input image's format.
(Ghostscript 9.25+ can copy JPEG images without transcoding them;
earlier versions will transcode.)
- Ghostscript's PDF/A conversion removes any XMP metadata that is not
one of the standard XMP metadata namespaces for PDFs. In particular,
PRISM Metdata is removed.
Regarding OCRmyPDF itself:
* PDFs that use transparency are not currently represented in the test suite
* The Python API exported by ``import ocrmypdf`` is design to help scripts that use OCRmyPDF but is not currently capable of running OCRmyPDF jobs due to limitations in an underlying library.
- PDFs that use transparency are not currently represented in the test
suite
Similar programs
----------------
================
To the author's knowledge, OCRmyPDF is the most feature-rich and thoroughly tested command line OCR PDF conversion tool. If it does not meet your needs, contributions and suggestions are welcome. If not, consider one of these similar open source programs:
To the author's knowledge, OCRmyPDF is the most feature-rich and
thoroughly tested command line OCR PDF conversion tool. If it does not
meet your needs, contributions and suggestions are welcome. If not,
consider one of these similar open source programs:
* pdf2pdfocr
* pdfsandwich
* pypdfocr
* pdfbeads
- pdf2pdfocr
- pdfsandwich
- pypdfocr
- pdfbeads
Web front-ends
--------------
==============
The Docker image ``ocrmypdf-alpine`` provides a web service front-end that allows files to submitted over HTTP and the results "downloaded". This is an HTTP server intended to simplify web services deployments; it is not intended to be deployed on the public internet and no real security measures to speak of.
The Docker image ``ocrmypdf-alpine`` provides a web service front-end
that allows files to submitted over HTTP and the results "downloaded".
This is an HTTP server intended to simplify web services deployments; it
is not intended to be deployed on the public internet and no real
security measures to speak of.
In addition, the following third-party integrations are available:
* `Nextcloud OCR <https://github.com/janis91/ocr>`_ is a free software plugin for the Nextcloud private cloud software
- `Nextcloud OCR <https://github.com/janis91/ocr>`__ is a free software
plugin for the Nextcloud private cloud software
OCRmyPDF is not designed to be secure against malware-bearing PDFs (see `Using OCRmyPDF online <ocr-service>`_). Users should ensure they comply with OCRmyPDF's licenses and the licenses of all dependencies. In particular, OCRmyPDF requires Ghostscript, which is licensed under AGPLv3.
OCRmyPDF is not designed to be secure against malware-bearing PDFs (see
`Using OCRmyPDF online <ocr-service>`__). Users should ensure they
comply with OCRmyPDF's licenses and the licenses of all dependencies. In
particular, OCRmyPDF requires Ghostscript, which is licensed under
AGPLv3.
.. |image| image:: images/bitmap_vs_svg.svg
docs/jbig2.rst
+34 -14
View File
@@ -1,35 +1,55 @@
.. _jbig2:
============================
Installing the JBIG2 encoder
============================
Most Linux distributions do not include a JBIG2 encoder since JBIG2 encoding was patented for a long time. All known JBIG2 US patents have expired as of 2017, but it is possible that unknown patents exist.
Most Linux distributions do not include a JBIG2 encoder since JBIG2
encoding was patented for a long time. All known JBIG2 US patents have
expired as of 2017, but it is possible that unknown patents exist.
JBIG2 encoding is recommended for OCRmyPDF and is used to losslessly create smaller PDFs. If JBIG2 encoding not available, lower quality encodings will be used.
JBIG2 encoding is recommended for OCRmyPDF and is used to losslessly
create smaller PDFs. If JBIG2 encoding not available, lower quality
encodings will be used.
JBIG2 decoding is not patented and is performed automatically by most PDF viewers. It is widely supported has been part of the PDF specification since 2001.
JBIG2 decoding is not patented and is performed automatically by most
PDF viewers. It is widely supported has been part of the PDF
specification since 2001.
On macOS, Homebrew packages jbig2enc and OCRmyPDF includes it by default. The Docker image for OCRmyPDF also builds its own JBIG2 encoder from source.
On macOS, Homebrew packages jbig2enc and OCRmyPDF includes it by
default. The Docker image for OCRmyPDF also builds its own JBIG2 encoder
from source.
For all other Linux, you must build a JBIG2 encoder from source:
.. code-block:: bash
git clone https://github.com/agl/jbig2enc
cd jbig2enc
./autogen.sh
./configure && make
[sudo] make install
git clone https://github.com/agl/jbig2enc
cd jbig2enc
./autogen.sh
./configure && make
[sudo] make install
.. _jbig2-lossy:
Lossy mode JBIG2
----------------
================
OCRmyPDF provides lossy mode JBIG2 as an advanced feature. Users should `review the technical concerns with JBIG2 in lossy mode <https://abbyy.technology/en:kb:tip:jbig2_compression_and_ocr>`_ and decide if this feature is acceptable for their use case.
OCRmyPDF provides lossy mode JBIG2 as an advanced feature. Users should
`review the technical concerns with JBIG2 in lossy
mode <https://abbyy.technology/en:kb:tip:jbig2_compression_and_ocr>`__
and decide if this feature is acceptable for their use case.
JBIG2 lossy mode does achieve higher compression ratios than any other monochrome (bitonal) compression technology; for large text documents the savings are considerable. JBIG2 lossless still gives great compression ratios and is a major improvement over the older CCITT G4 standard. As explained above, there is some risk of substitution errors.
JBIG2 lossy mode does achieve higher compression ratios than any other
monochrome (bitonal) compression technology; for large text documents
the savings are considerable. JBIG2 lossless still gives great
compression ratios and is a major improvement over the older CCITT G4
standard. As explained above, there is some risk of substitution errors.
To turn on JBIG2 lossy mode, add the argument ``--jbig2-lossy``. ``--optimize {1,2,3}`` are necessary for the argument to take effect also required. Also, a JBIG2 encoder must be installed as described in the previous section.
To turn on JBIG2 lossy mode, add the argument ``--jbig2-lossy``.
``--optimize {1,2,3}`` are necessary for the argument to take effect
also required. Also, a JBIG2 encoder must be installed as described in
the previous section.
*Due to an oversight, ocrmypdf v7.0 and v7.1 used lossy mode by default.*
*Due to an oversight, ocrmypdf v7.0 and v7.1 used lossy mode by
default.*
docs/languages.rst
+24 -14
View File
@@ -1,16 +1,20 @@
.. _lang-packs:
====================================
Installing additional language packs
====================================
OCRmyPDF uses Tesseract for OCR, and relies on its language packs for languages other than English.
OCRmyPDF uses Tesseract for OCR, and relies on its language packs for
languages other than English.
Tesseract supports `most languages <https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#languages>`_.
Tesseract supports `most
languages <https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#languages>`__.
For Linux users, you can often find packages that provide language packs:
For Linux users, you can often find packages that provide language
packs:
Debian and Ubuntu users
-----------------------
=======================
.. code-block:: bash
@@ -20,11 +24,13 @@ Debian and Ubuntu users
# Install Chinese Simplified language pack
apt-get install tesseract-ocr-chi-sim
You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple
languages can be requested using either ``-l eng+fre`` (English and French) or ``-l eng -l fre``.
You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as
to what languages it should search for. Multiple languages can be
requested using either ``-l eng+fre`` (English and French) or
``-l eng -l fre``.
Fedora users
------------
============
.. code-block:: bash
@@ -34,16 +40,20 @@ Fedora users
# Install Chinese Simplified language pack
dnf install tesseract-langpack-chi_sim
You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to
what languages it should search for. Multiple languages can be requested using
either ``-l eng+fre`` (English and French) or ``-l eng -l fre``.
You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as
to what languages it should search for. Multiple languages can be
requested using either ``-l eng+fre`` (English and French) or
``-l eng -l fre``.
macOS users
-----------
===========
You can install additional language packs by :ref:`installing Tesseract using Homebrew with all language packs <macos-all-languages>`.
You can install additional language packs by
:ref:`installing Tesseract using Homebrew with all language packs <macos-all-languages>`.
Docker users
------------
============
Users of the OCRmyPDF Docker image should install language packs into a derived Docker image as :ref:`described in that section <docker-lang-packs>`.
Users of the OCRmyPDF Docker image should install language packs into a
derived Docker image as
:ref:`described in that section <docker-lang-packs>`.
docs/plugins.rst
+24
View File
@@ -0,0 +1,24 @@
=======
Plugins
=======
You can use plugins to customize the behavior of OCRmyPDF at certain
points of interest.
Currently, it is possible to: - override the decision for whether or not
to perform OCR on a particular file - modify the image is about to be
sent for OCR
How plugins are imported
========================
Plugins are imported on demand, by the OCRmyPDF worker process that
needs to use them. As such, plugins cannot share state with each other,
and will be imported many times, once for each worker process.
Plugins currently cannot override the same hook.
How plugins are invoked
=======================
Plugins may be called from the command line:
docs/release_notes.rst
+1176 -765
View File
File diff suppressed because it is too large Load Diff
docs/security.rst
+116 -33
View File
@@ -1,79 +1,162 @@
===================
PDF security issues
===================
OCRmyPDF should only be used on PDFs you trust. It is not designed to protect you against malware.
OCRmyPDF should only be used on PDFs you trust. It is not designed to
protect you against malware.
Recognizing that many users have an interest in handling PDFs and applying OCR to PDFs they did not generate themselves, this article discusses the security implications of PDFs and how users can protect themselves.
Recognizing that many users have an interest in handling PDFs and
applying OCR to PDFs they did not generate themselves, this article
discusses the security implications of PDFs and how users can protect
themselves.
The disclaimer applies: this software has no warranties of any kind.
PDFs may contain malware
------------------------
========================
PDF is a rich, complex file format. The official PDF 1.7 specification, ISO 32000:2008, is hundreds of pages long and references several annexes each of which are similar in length. PDFs can contain video, audio, XML, JavaScript and other programming, and forms. In some cases, they can open internet connections to pre-selected URLs. All of these possible attack vectors.
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>`_.
In short, PDFs `may contain
viruses <https://security.stackexchange.com/questions/64052/can-a-pdf-file-contain-a-virus>`__.
This `article <https://theinvisiblethings.blogspot.ca/2013/02/converting-untrusted-pdfs-into-trusted.html>`_ describes a high-paranoia method which allows potentially hostile PDFs to be viewed and rasterized safely in a disposable virtual machine. A trusted PDF created in this manner is converted to images and loses all information making it searchable and losing all compression. OCRmyPDF could be used restore searchability.
This
`article <https://theinvisiblethings.blogspot.ca/2013/02/converting-untrusted-pdfs-into-trusted.html>`__
describes a high-paranoia method which allows potentially hostile PDFs
to be viewed and rasterized safely in a disposable virtual machine. A
trusted PDF created in this manner is converted to images and loses all
information making it searchable and losing all compression. OCRmyPDF
could be used restore searchability.
How OCRmyPDF processes PDFs
---------------------------
===========================
OCRmyPDF must open and interpret your PDF in order to insert an OCR layer. First, it runs all PDFs through `pikepdf <https://github.com/pikepdf/pikepdf>`_, a library based on `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.
OCRmyPDF must open and interpret your PDF in order to insert an OCR
layer. First, it runs all PDFs through
`pikepdf <https://github.com/pikepdf/pikepdf>`__, a library based on
`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.
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.
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.
.. _ocr-service:
Using OCRmyPDF online or as a service
-------------------------------------
=====================================
OCRmyPDF is not designed for use as a public web service where a malicious user could upload a chosen PDF. In particular, it is not necessarily secure against PDF malware or PDFs that cause denial of service. OCRmyPDF relies on Ghostscript, and therefore, if deployed online one should be prepared to comply with Ghostscript's Affero GPL license, OCRmyPDF's GPL license, and any other licenses.
OCRmyPDF is not designed for use as a public web service where a
malicious user could upload a chosen PDF. In particular, it is not
necessarily secure against PDF malware or PDFs that cause denial of
service. OCRmyPDF relies on Ghostscript, and therefore, if deployed
online one should be prepared to comply with Ghostscript's Affero GPL
license, OCRmyPDF's GPL license, and any other licenses.
Setting aside these concerns, a side effect of OCRmyPDF is it may incidentally sanitize PDFs that contain certain types of malware. It runs ``qpdf`` to repair the PDF, which could correct malformed PDF structures that are part of an attack. When PDF/A output is selected (the default), the input PDF is partially reconstructed by Ghostscript. When ``--force-ocr`` is used, all pages are rasterized and reconverted to PDF, which could remove malware in embedded images.
Setting aside these concerns, a side effect of OCRmyPDF is it may
incidentally sanitize PDFs that contain certain types of malware. It
runs ``qpdf`` to repair the PDF, which could correct malformed PDF
structures that are part of an attack. When PDF/A output is selected
(the default), the input PDF is partially reconstructed by Ghostscript.
When ``--force-ocr`` is used, all pages are rasterized and reconverted
to PDF, which could remove malware in embedded images.
OCRmyPDF should be relatively safe to use in a trusted intranet, with some considerations:
OCRmyPDF should be relatively safe to use in a trusted intranet, with
some considerations:
Limiting CPU usage
^^^^^^^^^^^^^^^^^^
------------------
OCRmyPDF will attempt to use all available CPUs and storage, so executing ``nice ocrmypdf`` or limiting the number of jobs with the ``-j`` argument may ensure the server remains available. Another option would be run OCRmyPDF jobs inside a Docker container, a virtual machine, or a cloud instance, which can impose its own limits on CPU usage and be terminated "from orbit" if it fails to complete.
OCRmyPDF will attempt to use all available CPUs and storage, so
executing ``nice ocrmypdf`` or limiting the number of jobs with the
``-j`` argument may ensure the server remains available. Another option
would be run OCRmyPDF jobs inside a Docker container, a virtual machine,
or a cloud instance, which can impose its own limits on CPU usage and be
terminated "from orbit" if it fails to complete.
Temporary storage requirements
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
------------------------------
OCRmyPDF will use a large amount of temporary storage for its work, proportional to the total number of pixels needed to rasterize the PDF. The raster image of a 8.5×11" color page at 300 DPI takes 25 MB uncompressed; OCRmyPDF saves its intermediates as PNG, but that still means it requires about 9 MB per intermediate based on average compression ratios. Multiple intermediates per page are also required, depending on the command line given. A rule of thumb would be to allow 100 MB of temporary storage per page in a file – meaning that a small cloud servers or small VM partitions should be provisioned with plenty of extra space, if say, a 500 page file might be sent.
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 check temporary storage usage on actual files, run
``ocrmypdf -k ...`` which will preserve and print the path to temporary
storage when the job is done.
To change where temporary files are stored, change the ``TMPDIR`` environment variable for ocrmypdf's environment. (Python's ``tempfile.gettempdir()`` returns the root directory in which temporary files will be stored.) For example, one could redirect ``TMPDIR`` to a large RAM disk to avoid wear on HDD/SSD and potentially improve performance. On Amazon Web Services, ``TMPDIR`` can be set to `empheral storage <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/InstanceStorage.html>`_.
To change where temporary files are stored, change the ``TMPDIR``
environment variable for ocrmypdf's environment. (Python's
``tempfile.gettempdir()`` returns the root directory in which temporary
files will be stored.) For example, one could redirect ``TMPDIR`` to a
large RAM disk to avoid wear on HDD/SSD and potentially improve
performance. On Amazon Web Services, ``TMPDIR`` can be set to `empheral
storage <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/InstanceStorage.html>`__.
Timeouts
^^^^^^^^
--------
To prevent excessively long OCR jobs consider setting ``--tesseract-timeout`` and/or ``--skip-big`` arguments. ``--skip-big`` is particularly helpful if your PDFs include documents such as reports on standard page sizes with large images attached - often large images are not worth OCR'ing anyway.
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.
The author also provides professional services that include OCR and
building databases around PDFs, and is happy to provide consultation.
Abbyy Cloud OCR is a viable commercial alternative with a web services
API.
Password protection, digital signatures and certification
---------------------------------------------------------
=========================================================
Password protected PDFs usually have two passwords, and owner and user password. When the user password is set to empty, PDF readers will open the file automatically and marked it as "(SECURED)". While not as reliable as a digital signature, this indicates that whoever set the password approved of the file at that time. When the user password is set, the document cannot be viewed without the password.
Password protected PDFs usually have two passwords, and owner and user
password. When the user password is set to empty, PDF readers will open
the file automatically and marked it as "(SECURED)". While not as
reliable as a digital signature, this indicates that whoever set the
password approved of the file at that time. When the user password is
set, the document cannot be viewed without the password.
Either way, OCRmyPDF does not remove passwords from PDFs and exits with an error on encountering them.
Either way, OCRmyPDF does not remove passwords from PDFs and exits with
an error on encountering them.
``qpdf``, one of OCRmyPDF's dependencies, can remove passwords. If the owner and user password are set, a password is required for ``qpdf``. If only the owner password is set, then the password can be stripped, even if one does not have the owner password.
``qpdf``, one of OCRmyPDF's dependencies, can remove passwords. If the
owner and user password are set, a password is required for ``qpdf``. If
only the owner password is set, then the password can be stripped, even
if one does not have the owner password.
After OCR is applied, password protection is not permitted on PDF/A documents but the file can be converted to regular PDF.
After OCR is applied, password protection is not permitted on PDF/A
documents but the file can be converted to regular PDF.
Many programs exist which are capable of inserting an image of someone's signature. On its own, this offers no security guarantees. It is trivial to remove the signature image and apply it to other files. This practice offers no real security.
Many programs exist which are capable of inserting an image of someone's
signature. On its own, this offers no security guarantees. It is trivial
to remove the signature image and apply it to other files. This practice
offers no real security.
Important documents can be digitally signed and certified to attest to their authorship. OCRmyPDF cannot do this. Open source tools such as pdfbox (Java) have this capability as does Adobe Acrobat.
Important documents can be digitally signed and certified to attest to
their authorship. OCRmyPDF cannot do this. Open source tools such as
pdfbox (Java) have this capability as does Adobe Acrobat.
misc/completion/ocrmypdf.bash
+6 -2
View File
@@ -1,5 +1,7 @@
# ocrmypdf completion -*- shell-script -*-
set -o errexit
_ocrmypdf()
{
local cur prev cword words split
@@ -49,7 +51,7 @@ _ocrmypdf()
return
;;
-v|--verbose)
COMPREPLY=( $( compgen -W '{1..9}' -- "$cur" ) ) # max level ?
COMPREPLY=( $( compgen -W '{0..2}' -- "$cur" ) ) # max level ?
return
;;
--tesseract-pagesegmode)
@@ -69,7 +71,7 @@ _ocrmypdf()
--sidecar --version --jobs --quiet --verbose --title --author
--subject --keywords --rotate-pages --remove-background --deskew
--clean --clean-final --unpaper-args --oversample --remove-vectors
--mask-barcodes --threshold --force-ocr --skip-text --redo-ocr
--threshold --force-ocr --skip-text --redo-ocr
--skip-big --jpeg-quality --png-quality --jbig2-lossy
--max-image-mpixels --tesseract-config --tesseract-pagesegmode
--help --tesseract-oem --pdf-renderer --tesseract-timeout
@@ -84,4 +86,6 @@ _ocrmypdf()
} &&
complete -F _ocrmypdf ocrmypdf
set +o errexit
# ex: filetype=sh
misc/completion/ocrmypdf.fish
+7 -2
View File
@@ -9,7 +9,6 @@ complete -c ocrmypdf -s d -l deskew -d "fix small horizontal alignment skew"
complete -c ocrmypdf -s c -l clean -d "clean document images before OCR"
complete -c ocrmypdf -s i -l clean-final -d "clean document images and keep result"
complete -c ocrmypdf -l remove-vectors -d "don't send vector objects to OCR"
complete -c ocrmypdf -l mask-barcodes -d "mask barcodes from OCR"
complete -c ocrmypdf -l threshold -d "threshold images before OCR"
complete -c ocrmypdf -s f -l force-ocr -d "OCR documents that already have printable text"
@@ -47,8 +46,14 @@ function __fish_ocrmypdf_optimize
end
complete -c ocrmypdf -x -s O -l optimize -a '(__fish_ocrmypdf_optimize)' -d "select optimization level"
function __fish_ocrmypdf_verbose
echo -e "0\t"(_ "standard output messages")
echo -e "1\t"(_ "troubleshooting output messages")
echo -e "2\t"(_ "debugging output messages")
end
complete -c ocrmypdf -x -s v -l verbose -a '(__fish_ocrmypdf_verbose)' -d "set verbosity level"
complete -c ocrmypdf -x -s j -l jobs -d "how many worker processes to use"
complete -c ocrmypdf -x -s v -a '(seq 1 9)'
complete -c ocrmypdf -x -l title -d "set metadata"
complete -c ocrmypdf -x -l author -d "set metadata"
complete -c ocrmypdf -x -l subject -d "set metadata"
misc/media/logo.afdesign
BIN
View File
Binary file not shown.
.docker/webservice.py → misc/webservice.py
View File
requirements/main.txt
+1 -1
View File
@@ -10,4 +10,4 @@ Pillow >= 5.0.0, != 5.1.0 ; sys_platform == "darwin"
pycparser == 2.19
python-xmp-toolkit == 2.0.1
reportlab == 3.5.13
ruffus == 2.8.1
tqdm == 4.32.1
setup.cfg
+2
View File
@@ -13,6 +13,8 @@ norecursedirs = lib .pc .git output cache resources
testpaths = tests
filterwarnings =
ignore:.*XMLParser.*:DeprecationWarning
markers =
slow
[isort]
multi_line_output=3
setup.py
+2 -5
View File
@@ -26,9 +26,6 @@ if sys.version_info < (3, 6):
sys.exit(1)
from setuptools import setup, find_packages
from subprocess import STDOUT, check_output, CalledProcessError
from collections.abc import Mapping
import re
# pylint: disable=w0613
@@ -104,10 +101,10 @@ setup(
# Pillow < 4 has BytesIO/TIFF bug w/img2pdf 0.2.3
# block 5.1.0, broken wheels
'reportlab >= 3.3.0', # oldest released version with sane image handling
'ruffus >= 2.7.0',
'tqdm >= 4',
],
tests_require=tests_require,
entry_points={'console_scripts': ['ocrmypdf = ocrmypdf.__main__:run_pipeline']},
entry_points={'console_scripts': ['ocrmypdf = ocrmypdf.__main__:run']},
package_data={'ocrmypdf': ['data/sRGB.icc']},
include_package_data=True,
zip_safe=False,
src/ocrmypdf/__init__.py
+1
View File
@@ -44,3 +44,4 @@ from . import hocrtransform
from . import leptonica
from . import pdfa
from . import pdfinfo
from .api import ocr, configure_logging, Verbosity
src/ocrmypdf/__main__.py
+33 -1113
View File
File diff suppressed because it is too large Load Diff
src/ocrmypdf/_weave.py → src/ocrmypdf/_graft.py
+82 -130
View File
@@ -15,17 +15,12 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
from contextlib import suppress
from itertools import groupby
from pathlib import Path
import os
from contextlib import suppress
from pathlib import Path
import pikepdf
from .exec import tesseract
from .helpers import flatten_groups, page_number
MAX_REPLACE_PAGES = int(os.environ.get('_OCRMYPDF_MAX_REPLACE_PAGES', 100))
@@ -48,7 +43,7 @@ def _update_page_resources(*, page, font, font_key, procset):
resources['/ProcSet'] = procset
def strip_invisible_text(pdf, page, log):
def strip_invisible_text(pdf, page):
stream = []
in_text_obj = False
render_mode = 0
@@ -93,7 +88,7 @@ def strip_invisible_text(pdf, page, log):
page.Contents = pikepdf.Stream(pdf, content_stream)
def _weave_layers_graft(
def _graft_text_layer(
*, pdf_base, page_num, text, font, font_key, procset, rotation, strip_old_text, log
):
"""Insert the text layer from text page 0 on to pdf_base at page_num"""
@@ -106,16 +101,6 @@ def _weave_layers_graft(
pdf_text = pikepdf.open(text)
pdf_text_contents = pdf_text.pages[0].Contents.read_bytes()
if not tesseract.has_textonly_pdf():
# If we don't have textonly_pdf, edit the stream to delete the
# instruction to draw the image Tesseract generated, which we do not
# use.
stream = bytearray(pdf_text_contents)
pattern = b'/Im1 Do'
idx = stream.find(pattern)
stream[idx : (idx + len(pattern))] = b' ' * len(pattern)
pdf_text_contents = bytes(stream)
base_page = pdf_base.pages.p(page_num)
# The text page always will be oriented up by this stage but the original
@@ -130,6 +115,7 @@ def _weave_layers_graft(
translate = pikepdf.PdfMatrix().translated(-wt / 2, -ht / 2)
untranslate = pikepdf.PdfMatrix().translated(wp / 2, hp / 2)
corner = pikepdf.PdfMatrix().translated(mediabox[0], mediabox[1])
# -rotation because the input is a clockwise angle and this formula
# uses CCW
rotation = -rotation % 360
@@ -143,19 +129,20 @@ def _weave_layers_graft(
scale_x = wp / wt
scale_y = hp / ht
log.debug('%r', (scale_x, scale_y))
# log.debug('%r', scale_x, scale_y)
scale = pikepdf.PdfMatrix().scaled(scale_x, scale_y)
# Translate the text so it is centered at (0, 0), rotate it there, adjust
# for a size different between initial and text PDF, then untranslate
ctm = translate @ rotate @ scale @ untranslate
# for a size different between initial and text PDF, then untranslate, and
# finally move the lower left corner to match the mediabox
ctm = translate @ rotate @ scale @ untranslate @ corner
pdf_text_contents = b'q %s cm\n' % ctm.encode() + pdf_text_contents + b'\nQ\n'
new_text_layer = pikepdf.Stream(pdf_base, pdf_text_contents)
if strip_old_text:
strip_invisible_text(pdf_base, base_page, log)
strip_invisible_text(pdf_base, base_page)
base_page.page_contents_add(new_text_layer, prepend=True)
@@ -189,144 +176,109 @@ def _find_font(text, pdf_base):
return None, None
def weave_layers(infiles, output_file, log, context):
"""Apply text layer and/or image layer changes to baseline file
class OcrGrafter:
def __init__(self, context):
self.context = context
self.log = context.log
self.path_base = Path(context.origin).resolve()
This is where the magic happens. infiles will be the main PDF to modify,
and optional .text.pdf and .image-layer.pdf files, organized however ruffus
organizes them.
self.pdf_base = pikepdf.open(self.path_base)
self.font, self.font_key = None, None
From .text.pdf, we copy the content stream (which contains the Tesseract
OCR results), and rotate it into place. The first time we do this, we also
copy the GlyphlessFont, and then reference that font again.
self.pdfinfo = context.pdfinfo
self.output_file = context.get_path('graft_layers.pdf')
For .image-layer.pdf, we check if this is a "pointer" to the original file,
or a new file. If a new file, we replace the page and remember that we
replaced this page.
self.procset = self.pdf_base.make_indirect(
pikepdf.Object.parse(b'[ /PDF /Text /ImageB /ImageC /ImageI ]')
)
Every 100 open files, we save intermediate results, to avoid any resource
limits, since pikepdf/qpdf need to keep a lot of open file handles in the
background. When objects are copied from one file to another qpdf, qpdf
doesn't actually copy the data until asked to write, so all the resources
it may need to remain available.
self.emplacements = 1
self.interim_count = 0
For completeness, we set up a /ProcSet on every page, although it's
unlikely any PDF viewer cares about this anymore.
"""
def input_sorter(key):
try:
return page_number(key)
except ValueError:
return -1
flat_inputs = sorted(flatten_groups(infiles), key=input_sorter)
groups = groupby(flat_inputs, key=input_sorter)
# Extract first item
_, basegroup = next(groups)
base = list(basegroup)[0]
path_base = Path(base).resolve()
pdf_base = pikepdf.open(path_base)
font, font_key, procset = None, None, None
pdfinfo = context.get_pdfinfo()
procset = pdf_base.make_indirect(
pikepdf.Object.parse(b'[ /PDF /Text /ImageB /ImageC /ImageI ]')
)
emplacements = 1
interim_count = 0
# Iterate rest
for page_num, layers in groups:
layers = list(layers)
log.debug(page_num)
log.debug(layers)
text = next((ii for ii in layers if ii.endswith('.text.pdf')), None)
image = next((ii for ii in layers if ii.endswith('.image-layer.pdf')), None)
if text and not font:
font, font_key = _find_font(text, pdf_base)
def graft_page(self, page_result):
pageno, image, text, _sidecar, autorotate_correction = page_result
if text and not self.font:
self.font, self.font_key = _find_font(text, self.pdf_base)
emplaced_page = False
content_rotation = pdfinfo[page_num - 1].rotation
content_rotation = self.pdfinfo[pageno].rotation
path_image = Path(image).resolve() if image else None
if path_image is not None and path_image != path_base:
if path_image is not None and path_image != self.path_base:
# We are updating the old page with a rasterized PDF of the new
# page (without changing objgen, to preserve references)
log.debug("Emplacement update")
self.log.debug("Emplacement update")
with pikepdf.open(image) as pdf_image:
emplacements += 1
self.emplacements += 1
foreign_image_page = pdf_image.pages[0]
pdf_base.pages.append(foreign_image_page)
local_image_page = pdf_base.pages[-1]
pdf_base.pages[page_num - 1].emplace(local_image_page)
del pdf_base.pages[-1]
self.pdf_base.pages.append(foreign_image_page)
local_image_page = self.pdf_base.pages[-1]
self.pdf_base.pages[pageno].emplace(local_image_page)
del self.pdf_base.pages[-1]
emplaced_page = True
autorotate_correction = context.get_rotation(page_num - 1)
if emplaced_page:
content_rotation = autorotate_correction
text_rotation = autorotate_correction
text_misaligned = (text_rotation - content_rotation) % 360
log.debug(
'%r',
[text_rotation, autorotate_correction, text_misaligned, content_rotation],
self.log.debug(
f"Rotations for page {pageno}: [text, auto, misalign, content] = "
f"{text_rotation}, {autorotate_correction}, "
f"{text_misaligned}, {content_rotation}"
)
if text and font:
if text and self.font:
# Graft the text layer onto this page, whether new or old
strip_old = context.get_options().redo_ocr
_weave_layers_graft(
pdf_base=pdf_base,
page_num=page_num,
strip_old = self.context.options.redo_ocr
_graft_text_layer(
pdf_base=self.pdf_base,
page_num=pageno + 1,
text=text,
font=font,
font_key=font_key,
font=self.font,
font_key=self.font_key,
rotation=text_misaligned,
procset=procset,
procset=self.procset,
strip_old_text=strip_old,
log=log,
log=self.log,
)
# Correct the rotation if applicable
pdf_base.pages[page_num - 1].Rotate = (
self.pdf_base.pages[pageno].Rotate = (
content_rotation - autorotate_correction
) % 360
if emplacements % MAX_REPLACE_PAGES == 0:
# Periodically save and reload the Pdf object. This will keep a
# lid on our memory usage for very large files. Attach the font to
# page 1 even if page 1 doesn't use it, so we have a way to get it
# back.
# TODO refactor this to outside the loop
page0 = pdf_base.pages[0]
_update_page_resources(
page=page0, font=font, font_key=font_key, procset=procset
)
if self.emplacements % MAX_REPLACE_PAGES == 0:
self.save_and_reload()
# We cannot read and write the same file, that will corrupt it
# but we don't to keep more copies than we need to. Delete intermediates.
# {interim_count} is the opened file we were updateing
# {interim_count - 1} can be deleted
# {interim_count + 1} is the new file will produce and open
old_file = output_file + f'_working{interim_count - 1}.pdf'
if not context.get_options().keep_temporary_files:
with suppress(FileNotFoundError):
os.unlink(old_file)
def save_and_reload(self):
# Periodically save and reload the Pdf object. This will keep a
# lid on our memory usage for very large files. Attach the font to
# page 1 even if page 1 doesn't use it, so we have a way to get it
# back.
page0 = self.pdf_base.pages[0]
_update_page_resources(
page=page0, font=self.font, font_key=self.font_key, procset=self.procset
)
next_file = output_file + f'_working{interim_count + 1}.pdf'
pdf_base.save(next_file)
pdf_base.close()
# We cannot read and write the same file, that will corrupt it
# but we don't to keep more copies than we need to. Delete intermediates.
# {interim_count} is the opened file we were updateing
# {interim_count - 1} can be deleted
# {interim_count + 1} is the new file will produce and open
old_file = self.output_file + f'_working{self.interim_count - 1}.pdf'
if not self.context.options.keep_temporary_files:
with suppress(FileNotFoundError):
os.unlink(old_file)
pdf_base = pikepdf.open(next_file)
procset = pdf_base.pages[0].Resources.ProcSet
font, font_key = None, None # Ensure we reacquire this information
interim_count += 1
next_file = self.output_file + f'_working{self.interim_count + 1}.pdf'
self.pdf_base.save(next_file)
self.pdf_base.close()
pdf_base.save(output_file)
pdf_base.close()
self.pdf_base = pikepdf.open(next_file)
self.procset = self.pdf_base.pages[0].Resources.ProcSet
self.font, self.font_key = None, None # Ensure we reacquire this information
self.interim_count += 1
def finalize(self):
self.pdf_base.save(self.output_file)
self.pdf_base.close()
return self.output_file
src/ocrmypdf/_jobcontext.py
+91 -50
View File
@@ -15,69 +15,110 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import shutil
import sys
from contextlib import suppress
from multiprocessing.managers import SyncManager
from .pdfinfo import PdfInfo
import os
class JobContext:
"""Holds our context for a particular run of the pipeline
class PicklableLoggerMixin:
def __init__(self):
self._log = None
A multiprocessing manager effectively creates a separate process
that keeps the master job context object. Other threads access
job context via multiprocessing proxy objects.
@property
def log(self):
if not self._log:
self._log = self.get_logger()
return self._log
While this would naturally lend itself @property's it seems to make
a little more sense to use functions to make it explicitly that the
invocation requires marshalling data across a process boundary.
def __getstate__(self):
# Python 3.6 is incapable of pickling a logger and marshalling it to another
# process (threading._RLock error), so we disconnect it before pickling,
# and create a new logger in the worker process.
state = self.__dict__.copy()
state['_log'] = None
return state
class PDFContext(PicklableLoggerMixin):
"""Holds our context for a particular run of the pipeline"""
def __init__(self, options, work_folder, origin, pdfinfo):
PicklableLoggerMixin.__init__(self)
self.options = options
self.work_folder = work_folder
self.origin = origin
self.pdfinfo = pdfinfo
if options:
self.name = os.path.basename(options.input_file)
else:
self.name = 'origin.pdf'
if self.name == '-':
self.name = 'stdin'
def get_logger(self):
return make_logger(self.options, filename=self.name)
def get_path(self, name):
return os.path.join(self.work_folder, name)
def get_page_contexts(self):
npages = len(self.pdfinfo)
for n in range(npages):
yield PageContext(self, n)
class PageContext(PicklableLoggerMixin):
"""Holds our context for a page
Must be pickable, so only store intrinsic/simple data elements
"""
def __init__(self):
self.pdfinfo = None
self.options = None
self.work_folder = None
self.rotations = {}
def __init__(self, pdf_context, pageno):
PicklableLoggerMixin.__init__(self)
self.work_folder = pdf_context.work_folder
self.origin = pdf_context.origin
self.options = pdf_context.options
self.name = pdf_context.name
self.pageno = pageno
self.pageinfo = pdf_context.pdfinfo[pageno]
self._log = None
def generate_pdfinfo(self, infile):
self.pdfinfo = PdfInfo(infile)
def get_logger(self):
return make_logger(self.options, filename=self.name, page=self.pageno + 1)
def get_pdfinfo(self):
"What we know about the input PDF"
return self.pdfinfo
def set_pdfinfo(self, pdfinfo):
self.pdfinfo = pdfinfo
def get_options(self):
return self.options
def set_options(self, options):
self.options = options
def get_work_folder(self):
return self.work_folder
def set_work_folder(self, work_folder):
self.work_folder = work_folder
def get_rotation(self, pageno):
return self.rotations.get(pageno, 0)
def set_rotation(self, pageno, value):
self.rotations[pageno] = value
class JobContextManager(SyncManager):
pass
def get_path(self, name):
return os.path.join(self.work_folder, "%06d_%s" % (self.pageno + 1, name))
def cleanup_working_files(work_folder, options):
if options.keep_temporary_files:
print(f"Temporary working files saved at:\n{work_folder}", file=sys.stderr)
print(f"Temporary working files retained at:\n{work_folder}", file=sys.stderr)
else:
with suppress(FileNotFoundError):
shutil.rmtree(work_folder)
shutil.rmtree(work_folder, ignore_errors=True)
class LogNameAdapter(logging.LoggerAdapter):
def process(self, msg, kwargs):
# return '[%s] %s' % (self.extra['filename'], msg), kwargs
return '%s' % (msg,), kwargs
class LogNamePageAdapter(logging.LoggerAdapter):
def process(self, msg, kwargs):
return (
#'[%s:%05u] %s' % (self.extra['filename'], self.extra['page'], msg),
'%4u: %s' % (self.extra['page'], msg),
kwargs,
)
def make_logger(options=None, prefix='ocrmypdf', filename=None, page=None):
log = logging.getLogger(prefix)
if filename and page:
adapter = LogNamePageAdapter(log, dict(filename=filename, page=page))
elif filename:
adapter = LogNameAdapter(log, dict(filename=filename))
else:
adapter = log
return adapter
src/ocrmypdf/_pipeline.py
+223 -591
View File
File diff suppressed because it is too large Load Diff
src/ocrmypdf/_sync.py
+381
View File
@@ -0,0 +1,381 @@
# © 2016 James R. Barlow: github.com/jbarlow83
#
# This file is part of OCRmyPDF.
#
# OCRmyPDF is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OCRmyPDF is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import logging.handlers
import multiprocessing
import os
import signal
import sys
import threading
from collections import namedtuple
from tempfile import mkdtemp
from tqdm import tqdm
from . import __version__
from ._jobcontext import PDFContext, cleanup_working_files, make_logger
from ._pipeline import (
convert_to_pdfa,
copy_final,
create_ocr_image,
create_pdf_page_from_image,
create_visible_page_jpg,
generate_postscript_stub,
get_orientation_correction,
get_pdfinfo,
is_ocr_required,
merge_sidecars,
metadata_fixup,
ocr_tesseract_hocr,
ocr_tesseract_textonly_pdf,
optimize_pdf,
preprocess_clean,
preprocess_deskew,
preprocess_remove_background,
rasterize,
rasterize_preview,
render_hocr_page,
should_visible_page_image_use_jpg,
triage,
validate_pdfinfo_options,
)
from ._validation import (
check_requested_output_file,
create_input_file,
report_output_file_size,
)
from ._graft import OcrGrafter
from .exceptions import ExitCode, ExitCodeException
from .exec import qpdf
from .helpers import available_cpu_count
from .pdfa import file_claims_pdfa
PageResult = namedtuple(
'PageResult', 'pageno, pdf_page_from_image, ocr, text, orientation_correction'
)
def preprocess(page_context, image, remove_background, deskew, clean):
if remove_background:
image = preprocess_remove_background(image, page_context)
if deskew:
image = preprocess_deskew(image, page_context)
if clean:
image = preprocess_clean(image, page_context)
return image
def exec_page_sync(page_context):
options = page_context.options
orientation_correction = 0
pdf_page_from_image_out = None
ocr_out = None
text_out = None
if is_ocr_required(page_context):
if options.rotate_pages:
# Rasterize
rasterize_preview_out = rasterize_preview(page_context.origin, page_context)
orientation_correction = get_orientation_correction(
rasterize_preview_out, page_context
)
rasterize_out = rasterize(
page_context.origin,
page_context,
correction=orientation_correction,
remove_vectors=False,
)
if not any([options.clean, options.clean_final, options.remove_vectors]):
ocr_image = preprocess_out = preprocess(
page_context,
rasterize_out,
options.remove_background,
options.deskew,
clean=False,
)
else:
if not options.lossless_reconstruction:
preprocess_out = preprocess(
page_context,
rasterize_out,
options.remove_background,
options.deskew,
clean=options.clean_final,
)
if options.remove_vectors:
rasterize_ocr_out = rasterize(
page_context.origin,
page_context,
correction=orientation_correction,
remove_vectors=True,
output_tag='_ocr',
)
else:
rasterize_ocr_out = rasterize_out
ocr_image = preprocess(
page_context,
rasterize_ocr_out,
options.remove_background,
options.deskew,
clean=options.clean,
)
ocr_image_out = create_ocr_image(ocr_image, page_context)
pdf_page_from_image_out = None
if not options.lossless_reconstruction:
visible_image_out = preprocess_out
if should_visible_page_image_use_jpg(page_context.pageinfo):
visible_image_out = create_visible_page_jpg(
visible_image_out, page_context
)
pdf_page_from_image_out = create_pdf_page_from_image(
visible_image_out, page_context
)
if options.pdf_renderer == 'hocr':
(hocr_out, text_out) = ocr_tesseract_hocr(ocr_image_out, page_context)
ocr_out = render_hocr_page(hocr_out, page_context)
if options.pdf_renderer == 'sandwich':
(ocr_out, text_out) = ocr_tesseract_textonly_pdf(
ocr_image_out, page_context
)
return PageResult(
pageno=page_context.pageno,
pdf_page_from_image=pdf_page_from_image_out,
ocr=ocr_out,
text=text_out,
orientation_correction=orientation_correction,
)
def post_process(pdf_file, context):
pdf_out = pdf_file
if context.options.output_type.startswith('pdfa'):
ps_stub_out = generate_postscript_stub(context)
pdf_out = convert_to_pdfa(pdf_out, ps_stub_out, context)
pdf_out = metadata_fixup(pdf_out, context)
return optimize_pdf(pdf_out, context)
def worker_init(queue):
"""Initialize a process pool worker"""
# Ignore SIGINT (our parent process will kill us gracefully)
signal.signal(signal.SIGINT, signal.SIG_IGN)
# Reconfigure the root logger for this process to send all messages to a queue
h = logging.handlers.QueueHandler(queue)
root = logging.getLogger()
root.handlers = []
root.addHandler(h)
def worker_thread_init(queue):
pass
def log_listener(queue):
"""Listen to the worker processes and forward the messages to logging
For simplicity this is a thread rather than a process. Only one process
should actually write to sys.stderr or whatever we're using, so if this is
made into a process the main application needs to be directed to it.
See https://docs.python.org/3/howto/logging-cookbook.html#logging-to-a-single-file-from-multiple-processes
"""
while True:
try:
record = queue.get()
if record is None:
break
logger = logging.getLogger(record.name)
logger.handle(record)
except Exception:
import traceback
print("Logging problem", file=sys.stderr)
traceback.print_exc(file=sys.stderr)
def exec_concurrent(context):
"""Execute the pipeline concurrently"""
# Run exec_page_sync on every page context
max_workers = min(len(context.pdfinfo), context.options.jobs)
if max_workers > 1:
context.log.info("Start processing %d pages concurrent", max_workers)
# Tesseract 4.0 is multithreaded, and we also run multiple workers. We want to
# avoid the situation where we end up trying to run NxN jobs on N CPU cores,
# as that gives poor performance. Performance testing shows we're better off
# parallelizing ocrmypdf and forcing Tesseract to be single threaded, which we
# get by setting the envvar OMP_THREAD_LIMIT to 1. But if the page count of the
# input file is small, then we allow Tesseract to use threads, subject to the
# constraint: (ocrmypdf workers) * (tesseract threads) <= max_workers and limiting
# Tesseract to 4 threads.
tess_threads = min(4, context.options.jobs // max_workers)
if context.options.tesseract_env is None:
context.options.tesseract_env = os.environ.copy()
context.options.tesseract_env.setdefault('OMP_THREAD_LIMIT', str(tess_threads))
if tess_threads > 1:
context.log.info("Using Tesseract OpenMP thread limit %d", tess_threads)
if context.options.use_threads:
from multiprocessing.dummy import Pool
initializer = worker_thread_init
else:
Pool = multiprocessing.Pool
initializer = worker_init
sidecars = [None] * len(context.pdfinfo)
ocrgraft = OcrGrafter(context)
log_queue = multiprocessing.Queue(-1)
listener = threading.Thread(target=log_listener, args=(log_queue,))
listener.start()
with tqdm(
total=(2 * len(context.pdfinfo)),
desc='OCR',
unit='page',
unit_scale=0.5,
disable=not context.options.progress_bar,
) as pbar, Pool(
processes=max_workers, initializer=initializer, initargs=(log_queue,)
) as pool:
results = pool.imap_unordered(exec_page_sync, context.get_page_contexts())
while True:
try:
page_result = results.next()
sidecars[page_result.pageno] = page_result.text
pbar.update()
ocrgraft.graft_page(page_result)
pbar.update()
except StopIteration:
break
except (Exception, KeyboardInterrupt):
pool.terminate()
log_queue.put_nowait(None) # Terminate log listener
# Don't try listener.join() here, will deadlock
raise
log_queue.put_nowait(None)
listener.join()
# Output sidecar text
if context.options.sidecar:
text = merge_sidecars(sidecars, context)
# Copy text file to destination
copy_final(text, context.options.sidecar, context)
# Merge layers to one single pdf
pdf = ocrgraft.finalize()
# PDF/A and metadata
pdf = post_process(pdf, context)
# Copy PDF file to destination
copy_final(pdf, context.options.output_file, context)
class NeverRaise(Exception):
"""An exception that is never raised"""
pass
def run_pipeline(options, api=False):
log = make_logger(options, __name__)
# Any changes to options will not take effect for options that are already
# bound to function parameters in the pipeline. (For example
# options.input_file, options.pdf_renderer are already bound.)
if not options.jobs:
options.jobs = available_cpu_count()
work_folder = mkdtemp(prefix="com.github.ocrmypdf.")
try:
check_requested_output_file(options)
start_input_file = create_input_file(options, work_folder)
# Triage image or pdf
origin_pdf = triage(
start_input_file, os.path.join(work_folder, 'origin.pdf'), options, log
)
# Gather pdfinfo and create context
pdfinfo = get_pdfinfo(
origin_pdf,
detailed_page_analysis=options.redo_ocr,
progbar=options.progress_bar,
)
context = PDFContext(options, work_folder, origin_pdf, pdfinfo)
# Validate options are okay for this pdf
validate_pdfinfo_options(context)
# Execute the pipeline
exec_concurrent(context)
if options.output_file == '-':
log.info("Output sent to stdout")
elif os.path.samefile(options.output_file, os.devnull):
pass # Say nothing when sending to dev null
else:
if options.output_type.startswith('pdfa'):
pdfa_info = file_claims_pdfa(options.output_file)
if pdfa_info['pass']:
log.info(
"Output file is a %s (as expected)", pdfa_info['conformance']
)
else:
log.warning(
"Output file is okay but is not PDF/A (seems to be %s)",
pdfa_info['conformance'],
)
return ExitCode.pdfa_conversion_failed
if not qpdf.check(options.output_file, log):
log.warning('Output file: The generated PDF is INVALID')
return ExitCode.invalid_output_pdf
report_output_file_size(options, start_input_file, options.output_file)
except (KeyboardInterrupt if not api else NeverRaise) as e:
if options.verbose >= 1:
log.exception("KeyboardInterrupt")
else:
log.error("KeyboardInterrupt")
return ExitCode.ctrl_c
except (ExitCodeException if not api else NeverRaise) as e:
if str(e):
log.error("%s: %s", type(e).__name__, str(e))
else:
log.error(type(e).__name__)
return e.exit_code
except (Exception if not api else NeverRaise) as e:
log.exception("An exception occurred while executing the pipeline")
return ExitCode.other_error
finally:
cleanup_working_files(work_folder, options)
return ExitCode.ok
src/ocrmypdf/_validation.py
+445
View File
@@ -0,0 +1,445 @@
#!/usr/bin/env python3
# © 2015-17 James R. Barlow: github.com/jbarlow83
#
# This file is part of OCRmyPDF.
#
# OCRmyPDF is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OCRmyPDF is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import os
import sys
from pathlib import Path
from shutil import copyfileobj
import PIL
from ._unicodefun import verify_python3_env
from .exceptions import (
BadArgsError,
InputFileError,
MissingDependencyError,
OutputFileAccessError,
)
from .exec import (
check_external_program,
ghostscript,
jbig2enc,
pngquant,
qpdf,
tesseract,
unpaper,
)
from .helpers import is_file_writable, re_symlink, is_iterable_notstr, monotonic
# -------------
# External dependencies
HOCR_OK_LANGS = frozenset(['eng', 'deu', 'spa', 'ita', 'por'])
log = logging.getLogger(__name__)
# --------
# Critical environment tests
verify_python3_env()
def check_options_languages(options):
if not options.language:
options.language = ['eng'] # Enforce English hegemony
# Support v2.x "eng+deu" language syntax
if '+' in options.language[0]:
options.language = options.language[0].split('+')
languages = set(options.language)
if not languages.issubset(tesseract.languages()):
msg = (
"The installed version of tesseract does not have language "
"data for the following requested languages: \n"
)
for lang in languages - tesseract.languages():
msg += lang + '\n'
raise MissingDependencyError(msg)
def check_options_output(options):
# We have these constraints to check for.
# 1. Ghostscript < 9.20 mangles multibyte Unicode
# 2. hocr doesn't work on non-Latin languages (so don't select it)
languages = set(options.language)
is_latin = languages.issubset(HOCR_OK_LANGS)
if options.pdf_renderer == 'hocr' and not is_latin:
msg = (
"The 'hocr' PDF renderer is known to cause problems with one "
"or more of the languages in your document. Use "
"--pdf-renderer auto (the default) to avoid this issue."
)
log.warning(msg)
if ghostscript.version() < '9.20' and options.output_type != 'pdf' and not is_latin:
# https://bugs.ghostscript.com/show_bug.cgi?id=696874
# Ghostscript < 9.20 fails to encode multibyte characters properly
msg = (
"The installed version of Ghostscript does not work correctly "
"with the OCR languages you specified. Use --output-type pdf or "
"upgrade to Ghostscript 9.20 or later to avoid this issue."
)
msg += f"Found Ghostscript {ghostscript.version()}"
log.warning(msg)
# Decide on what renderer to use
if options.pdf_renderer == 'auto':
options.pdf_renderer = 'sandwich'
if options.pdf_renderer == 'sandwich' and not tesseract.has_textonly_pdf(
options.tesseract_env
):
raise MissingDependencyError(
"You are using an alpha version of Tesseract 4.0 that does not support "
"the textonly_pdf parameter. We don't support versions this old."
)
if options.output_type == 'pdfa':
options.output_type = 'pdfa-2'
if options.output_type == 'pdfa-3' and ghostscript.version() < '9.19':
raise MissingDependencyError(
"--output-type pdfa-3 requires Ghostscript 9.19 or later"
)
lossless_reconstruction = False
if not any(
(
options.deskew,
options.clean_final,
options.force_ocr,
options.remove_background,
)
):
lossless_reconstruction = True
options.lossless_reconstruction = lossless_reconstruction
if not options.lossless_reconstruction and options.redo_ocr:
raise BadArgsError(
"--redo-ocr is not currently compatible with --deskew, "
"--clean-final, and --remove-background"
)
def check_options_sidecar(options):
if options.sidecar == '\0':
if options.output_file == '-':
raise BadArgsError(
"--sidecar filename must be specified when output file is stdout."
)
options.sidecar = options.output_file + '.txt'
def check_options_preprocessing(options):
if options.clean_final:
options.clean = True
if options.unpaper_args and not options.clean:
raise BadArgsError("--clean is required for --unpaper-args")
if options.clean:
check_external_program(
program='unpaper',
package='unpaper',
version_checker=unpaper.version,
need_version='6.1',
required_for=['--clean, --clean-final'],
)
try:
if options.unpaper_args:
options.unpaper_args = unpaper.validate_custom_args(
options.unpaper_args
)
except Exception as e:
raise BadArgsError(str(e))
def _pages_from_ranges(ranges):
if is_iterable_notstr(ranges):
return set(ranges)
pages = []
page_groups = ranges.replace(' ', '').split(',')
for g in page_groups:
if not g:
continue
try:
start, end = g.split('-')
except ValueError:
pages.append(int(g) - 1)
else:
pages.extend(range(int(start) - 1, int(end)))
if not monotonic(pages):
log.warning(
"List of pages to process contains duplicate pages, or pages that are "
"out of order"
)
if any(page < 0 for page in pages):
raise BadArgsError("pages refers to a page number less than 1")
log.debug("OCRing only these pages: %s", pages)
return set(pages)
def check_options_ocr_behavior(options):
exclusive_options = sum(
[
(1 if opt else 0)
for opt in (options.force_ocr, options.skip_text, options.redo_ocr)
]
)
if exclusive_options >= 2:
raise BadArgsError("Choose only one of --force-ocr, --skip-text, --redo-ocr.")
if options.pages and options.sidecar:
raise BadArgsError("--pages and --sidecar are mutually exclusive")
if options.pages:
options.pages = _pages_from_ranges(options.pages)
def check_options_optimizing(options):
if options.optimize >= 2:
check_external_program(
program='pngquant',
package='pngquant',
version_checker=pngquant.version,
need_version='2.0.1',
required_for='--optimize {2,3}',
)
if options.optimize >= 2:
# Although we use JBIG2 for optimize=1, don't nag about it unless the
# user is asking for more optimization
check_external_program(
program='jbig2',
package='jbig2enc',
version_checker=jbig2enc.version,
need_version='0.28',
required_for='--optimize {2,3} | --jbig2-lossy',
recommended=True if not options.jbig2_lossy else False,
)
if options.optimize == 0 and any(
[options.jbig2_lossy, options.png_quality, options.jpeg_quality]
):
log.warning(
"The arguments --jbig2-lossy, --png-quality, and --jpeg-quality "
"will be ignored because --optimize=0."
)
def check_options_advanced(options):
if options.pdfa_image_compression != 'auto' and options.output_type.startswith(
'pdfa'
):
log.warning(
"--pdfa-image-compression argument has no effect when "
"--output-type is not 'pdfa', 'pdfa-1', or 'pdfa-2'"
)
if tesseract.v4(options.tesseract_env) and (
options.user_words or options.user_patterns
):
log.warning('Tesseract 4.x ignores --user-words, so this has no effect')
def check_options_metadata(options):
import unicodedata
docinfo = [options.title, options.author, options.keywords, options.subject]
for s in (m for m in docinfo if m):
for c in s:
if unicodedata.category(c) == 'Co' or ord(c) >= 0x10000:
raise ValueError(
"One of the metadata strings contains "
"an unsupported Unicode character: '{}' (U+{})".format(
c, hex(ord(c))[2:].upper()
)
)
def check_options_pillow(options):
PIL.Image.MAX_IMAGE_PIXELS = int(options.max_image_mpixels * 1_000_000)
if PIL.Image.MAX_IMAGE_PIXELS == 0:
PIL.Image.MAX_IMAGE_PIXELS = None
def check_options(options):
check_options_languages(options)
check_options_metadata(options)
check_options_output(options)
check_options_sidecar(options)
check_options_preprocessing(options)
check_options_ocr_behavior(options)
check_options_optimizing(options)
check_options_advanced(options)
check_options_pillow(options)
check_dependency_versions(options)
def check_closed_streams(options):
"""Work around Python issue with multiprocessing forking on closed streams
https://bugs.python.org/issue28326
Attempting to a fork/exec a new Python process when any of std{in,out,err}
are closed or not flushable for some reason may raise an exception.
Fix this by opening devnull if the handle seems to be closed. Do this
globally to avoid tracking places all places that fork.
Seems to be specific to multiprocessing.Process not all Python process
forkers.
The error actually occurs when the stream object is not flushable,
but replacing an open stream object that is not flushable with
/dev/null is a bad idea since it will create a silent failure. Replacing
a closed handle with /dev/null seems safe.
"""
if sys.version_info[0:3] >= (3, 6, 4):
return True # Issued fixed in Python 3.6.4+
if sys.stderr is None:
sys.stderr = open(os.devnull, 'w')
if sys.stdin is None:
if options.input_file == '-':
log.error("Trying to read from stdin but stdin seems closed")
return False
sys.stdin = open(os.devnull, 'r')
if sys.stdout is None:
if options.output_file == '-':
# Can't replace stdout if the user is piping
# If this case can even happen, it must be some kind of weird
# stream.
log.error(
"Output was set to stdout '-' but the stream attached to "
"stdout does not support the flush() system call. This "
"will fail."
)
return False
sys.stdout = open(os.devnull, 'w')
return True
def log_page_orientations(pdfinfo):
direction = {0: 'n', 90: 'e', 180: 's', 270: 'w'}
orientations = []
for n, page in enumerate(pdfinfo):
angle = page.rotation or 0
if angle != 0:
orientations.append('{0}{1}'.format(n + 1, direction.get(angle, '')))
if orientations:
log.info('Page orientations detected: %s', ' '.join(orientations))
def create_input_file(options, work_folder):
if options.input_file == '-':
# stdin
log.info('reading file from standard input')
target = os.path.join(work_folder, 'stdin')
with open(target, 'wb') as stream_buffer:
copyfileobj(sys.stdin.buffer, stream_buffer)
return target
else:
try:
target = os.path.join(work_folder, 'origin')
re_symlink(options.input_file, target)
return target
except FileNotFoundError:
raise InputFileError(f"File not found - {options.input_file}")
def check_requested_output_file(options):
if options.output_file == '-':
if sys.stdout.isatty():
raise BadArgsError(
"Output was set to stdout '-' but it looks like stdout "
"is connected to a terminal. Please redirect stdout to a "
"file."
)
elif not is_file_writable(options.output_file):
raise OutputFileAccessError(
f"Output file location ({options.output_file}) is not a writable file."
)
def report_output_file_size(options, input_file, output_file):
try:
output_size = Path(output_file).stat().st_size
input_size = Path(input_file).stat().st_size
except FileNotFoundError:
return # Outputting to stream or something
ratio = output_size / input_size
if ratio < 1.35 or input_size < 25000:
return # Seems fine
reasons = []
image_preproc = {
'deskew',
'clean_final',
'remove_background',
'oversample',
'force_ocr',
}
for arg in image_preproc:
if getattr(options, arg, False):
reasons.append(
f"The argument --{arg.replace('_', '-')} was issued, causing transcoding."
)
if reasons:
explanation = "Possible reasons for this include:\n" + '\n'.join(reasons) + "\n"
else:
explanation = "No reason for this increase is known. Please report this issue."
log.warning(
f"The output file size is {ratio:.2f}× larger than the input file.\n"
f"{explanation}"
)
def check_dependency_versions(options):
check_external_program(
program='tesseract',
package={'darwin': 'tesseract', 'linux': 'tesseract-ocr'},
version_checker=tesseract.version,
need_version='4.0.0', # using backport for Travis CI
)
check_external_program(
program='gs',
package='ghostscript',
version_checker=ghostscript.version,
need_version='9.15', # limited by Travis CI / Ubuntu 14.04 backports
)
if ghostscript.version() == '9.24':
raise MissingDependencyError(
"Ghostscript 9.24 contains serious regressions and is not "
"supported. Please upgrade to Ghostscript 9.25 or use an older "
"version."
)
check_external_program(
program='qpdf',
package='qpdf',
version_checker=qpdf.version,
need_version='8.0.2',
)
src/ocrmypdf/api.py
+239
View File
@@ -0,0 +1,239 @@
# © 2019 James R. Barlow: github.com/jbarlow83
#
# This file is part of OCRmyPDF.
#
# OCRmyPDF is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OCRmyPDF is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import os
import sys
from enum import IntEnum
from pathlib import Path
from tqdm import tqdm
from .cli import parser
from ._sync import run_pipeline
from ._validation import check_options
class TqdmConsole:
"""Wrapper to log messages in a way that is compatible with tqdm progress bar"""
def __init__(self, file):
self.file = file
self.py36 = sys.version_info >= (3, 6)
def write(self, msg):
# When no progress bar is active, tqdm.write() routes to print()
if self.py36:
if msg.strip() != '':
tqdm.write(msg.rstrip(), end='\n', file=self.file)
else:
tqdm.write(msg.rstrip(), end='\n', file=self.file)
def flush(self):
if hasattr(self.file, "flush"):
self.file.flush()
class Verbosity(IntEnum):
"""Verbosity level for configure_logging."""
quiet = -1 #: Suppress most messages
default = 0 #: Default level of logging
debug = 1 #: Output ocrmypdf debug messages
debug_all = 2 #: More detailed debugging from ocrmypdf and dependent modules
def configure_logging(verbosity, progress_bar_friendly=True, manage_root_logger=False):
"""Set up logging.
Library users may wish to use this function if they want their log output to be
similar to ocrmypdf command line interface. If not used, the external application
should configure logging on its own.
ocrmypdf will perform all of its logging under the `"ocrmypdf"` logging namespace.
In addition, ocrmypdf imports pdfminer, which logs under `"pdfminer"`. A library
user may wish to configure both; note that pdfminer is extremely chatty at the log
level logging.INFO.
Library users may perform additional configuration afterwards.
Args:
verbosity (Verbosity): Verbosity level.
progress_bar_friendly (bool): Install the TqdmConsole log handler, which is
compatible with the tqdm progress bar; without this log messages will
overwrite the progress bar
manage_root_logger (bool): Configure the process's root logger, to ensure
all log output is sent through
"""
prefix = '' if manage_root_logger else 'ocrmypdf'
log = logging.getLogger(prefix)
log.setLevel(logging.INFO)
if progress_bar_friendly:
console = logging.StreamHandler(stream=TqdmConsole(sys.stderr))
else:
console = logging.StreamHandler(stream=sys.stderr)
if verbosity < 0:
console.setLevel(logging.ERROR)
elif verbosity >= 1:
console.setLevel(logging.DEBUG)
else:
console.setLevel(logging.INFO)
formatter = logging.Formatter('%(levelname)7s - %(message)s')
if verbosity >= 1:
log.setLevel(logging.DEBUG)
if verbosity >= 2:
formatter = logging.Formatter('%(name)s - %(levelname)7s - %(message)s')
console.setFormatter(formatter)
log.addHandler(console)
if verbosity <= 1:
pdfminer_log = logging.getLogger('pdfminer')
pdfminer_log.setLevel(logging.ERROR)
pil_log = logging.getLogger('PIL')
pil_log.setLevel(logging.INFO)
if manage_root_logger:
logging.captureWarnings(True)
def create_options(*, input_file, output_file, **kwargs):
cmdline = []
deferred = []
for arg, val in kwargs.items():
if val is None:
continue
if arg == 'tesseract_env':
deferred.append((arg, val))
continue
cmd_style_arg = arg.replace('_', '-')
cmdline.append(f"--{cmd_style_arg}")
if isinstance(val, bool):
continue
if isinstance(val, (int, float)):
cmdline.append(str(val))
elif isinstance(val, str):
cmdline.append(val)
elif isinstance(val, Path):
cmdline.append(str(val))
else:
raise TypeError(f"{arg}: {val} ({type(val)})")
cmdline.append(str(input_file))
cmdline.append(str(output_file))
parser.api_mode = True
options = parser.parse_args(cmdline)
for keyword, val in deferred:
setattr(options, keyword, val)
# If we are running a Tesseract spoof, ensure it knows what the input file is
if os.environ.get('PYTEST_CURRENT_TEST') and options.tesseract_env:
options.tesseract_env['_OCRMYPDF_TEST_INFILE'] = input_file
return options
def ocr( # pylint: disable=unused-argument
input_file,
output_file,
*,
language=None,
image_dpi=None,
output_type=None,
sidecar=None,
jobs=None,
use_threads=None,
title=None,
author=None,
subject=None,
keywords=None,
rotate_pages=None,
remove_background=None,
deskew=None,
clean=None,
clean_final=None,
unpaper_args=None,
oversample=None,
remove_vectors=None,
threshold=None,
force_ocr=None,
skip_text=None,
redo_ocr=None,
skip_big=None,
optimize=None,
jpg_quality=None,
png_quality=None,
jbig2_lossy=None,
jbig2_page_group_size=None,
pages=None,
max_image_mpixels=None,
tesseract_config=None,
tesseract_pagesegmode=None,
tesseract_oem=None,
pdf_renderer=None,
tesseract_timeout=None,
rotate_pages_threshold=None,
pdfa_image_compression=None,
user_words=None,
user_patterns=None,
keep_temporary_files=None,
progress_bar=None,
tesseract_env=None,
):
"""Run OCRmyPDF on one PDF or image.
For most arguments, see documentation for the equivalent command line parameter.
A few specific arguments are discussed here:
Args:
use_threads (bool): Use worker threads instead of processes. This reduces
performance but may make debugging easier since it is easier to set
breakpoints.
tesseract_env (dict): Override environment variables for Tesseract
Raises:
ocrmypdf.PdfMergeFailedError: If the input PDF is malformed, preventing merging
with the OCR layer.
ocrmypdf.MissingDependencyError: If a required dependency program is missing or
was not found on PATH.
ocrmypdf.UnsupportedImageFormatError: If the input file type was an image that
could not be read, or some other file type that is not a PDF.
ocrmypdf.DpiError: If the input file is an image, but the resolution of the
image is not credible (allowing it to proceed would cause poor OCR).
ocrmypdf.OutputFileAccessError: If an attempt to write to the intended output
file failed.
ocrmypdf.PriorOcrFoundError: If the input PDF seems to have OCR or digital
text already, and settings did not tell us to proceed.
ocrmypdf.InputFileError: Any other problem with the input file.
ocrmypdf.SubprocessOutputError: Any error related to executing a subprocess.
ocrmypdf.EncryptedPdfERror: If the input PDF is encrypted (password protected).
OCRmyPDF does not remove passwords.
ocrmypdf.TesseractConfigError: If Tesseract reported its configuration was not
valid.
Returns:
:class:`ocrmypdf.ExitCode`
"""
options = create_options(**locals())
check_options(options)
return run_pipeline(options, api=True)
src/ocrmypdf/cli.py
+478
View File
@@ -0,0 +1,478 @@
# © 2015-19 James R. Barlow: github.com/jbarlow83
#
# This file is part of OCRmyPDF.
#
# OCRmyPDF is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OCRmyPDF is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import argparse
from . import PROGRAM_NAME, VERSION
def numeric(basetype, min_=None, max_=None):
"""Validator for numeric params"""
min_ = basetype(min_) if min_ is not None else None
max_ = basetype(max_) if max_ is not None else None
def _numeric(string):
value = basetype(string)
if (min_ is not None and value < min_) or (max_ is not None and value > max_):
msg = "%r not in valid range %r" % (string, (min_, max_))
raise argparse.ArgumentTypeError(msg)
return value
_numeric.__name__ = basetype.__name__
return _numeric
class ArgumentParser(argparse.ArgumentParser):
"""Override parser's default behavior of calling sys.exit()
https://stackoverflow.com/questions/5943249/python-argparse-and-controlling-overriding-the-exit-status-code
"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.api_mode = False
def error(self, message):
if not self.api_mode:
super().error(message)
return
raise ValueError(message)
parser = ArgumentParser(
prog=PROGRAM_NAME,
fromfile_prefix_chars='@',
formatter_class=argparse.RawDescriptionHelpFormatter,
description="""\
Generates a searchable PDF or PDF/A from a regular PDF.
OCRmyPDF rasterizes each page of the input PDF, optionally corrects page
rotation and performs image processing, runs the Tesseract OCR engine on the
image, and then creates a PDF from the OCR information.
""",
epilog="""\
OCRmyPDF attempts to keep the output file at about the same size. If a file
contains losslessly compressed images, and output file will be losslessly
compressed as well.
PDF is a page description file that attempts to preserve a layout exactly.
A PDF can contain vector objects (such as text or lines) and raster objects
(images). A page might have multiple images. OCRmyPDF is prepared to deal
with the wide variety of PDFs that exist in the wild.
When a PDF page contains text, OCRmyPDF assumes that the page has already
been OCRed or is a "born digital" page that should not be OCRed. The default
behavior is to exit in this case without producing a file. You can use the
option --skip-text to ignore pages with text, or --force-ocr to rasterize
all objects on the page and produce an image-only PDF as output.
ocrmypdf --skip-text file_with_some_text_pages.pdf output.pdf
ocrmypdf --force-ocr word_document.pdf output.pdf
If you are concerned about long-term archiving of PDFs, use the default option
--output-type pdfa which converts the PDF to a standardized PDF/A-2b. This
converts images to sRGB colorspace, removes some features from the PDF such
as Javascript or forms. If you want to minimize the number of changes made to
your PDF, use --output-type pdf.
If OCRmyPDF is given an image file as input, it will attempt to convert the
image to a PDF before processing. For more control over the conversion of
images to PDF, use the Python package img2pdf or other image to PDF software.
For example, this command uses img2pdf to convert all .png files beginning
with the 'page' prefix to a PDF, fitting each image on A4-sized paper, and
sending the result to OCRmyPDF through a pipe. img2pdf is a dependency of
ocrmypdf so it is already installed.
img2pdf --pagesize A4 page*.png | ocrmypdf - myfile.pdf
Online documentation is located at:
https://ocrmypdf.readthedocs.io/en/latest/introduction.html
""",
)
parser.add_argument(
'input_file',
metavar="input_pdf_or_image",
help="PDF file containing the images to be OCRed (or '-' to read from "
"standard input)",
)
parser.add_argument(
'output_file',
metavar="output_pdf",
help="Output searchable PDF file (or '-' to write to standard output). "
"Existing files will be ovewritten. If same as input file, the "
"input file will be updated only if processing is successful.",
)
parser.add_argument(
'-l',
'--language',
action='append',
help="Language(s) of the file to be OCRed (see tesseract --list-langs for "
"all language packs installed in your system). Use -l eng+deu for "
"multiple languages.",
)
parser.add_argument(
'--image-dpi',
metavar='DPI',
type=int,
help="For input image instead of PDF, use this DPI instead of file's.",
)
parser.add_argument(
'--output-type',
choices=['pdfa', 'pdf', 'pdfa-1', 'pdfa-2', 'pdfa-3'],
default='pdfa',
help="Choose output type. 'pdfa' creates a PDF/A-2b compliant file for "
"long term archiving (default, recommended) but may not suitable "
"for users who want their file altered as little as possible. 'pdfa' "
"also has problems with full Unicode text. 'pdf' attempts to "
"preserve file contents as much as possible. 'pdf-a1' creates a "
"PDF/A1-b file. 'pdf-a2' is equivalent to 'pdfa'. 'pdf-a3' creates a "
"PDF/A3-b file.",
)
# Use null string '\0' as sentinel to indicate the user supplied no argument,
# since that is the only invalid character for filepaths on all platforms
# bool('\0') is True in Python
parser.add_argument(
'--sidecar',
nargs='?',
const='\0',
default=None,
metavar='FILE',
help="Generate sidecar text files that contain the same text recognized "
"by Tesseract. This may be useful for building a OCR text database. "
"If FILE is omitted, the sidecar file be named {output_file}.txt "
"If FILE is set to '-', the sidecar is written to stdout (a "
"convenient way to preview OCR quality). The output file and sidecar "
"may not both use stdout at the same time.",
)
parser.add_argument(
'--version',
action='version',
version=VERSION,
help="Print program version and exit",
)
jobcontrol = parser.add_argument_group("Job control options")
jobcontrol.add_argument(
'-j',
'--jobs',
metavar='N',
type=numeric(int, 0, 256),
help="Use up to N CPU cores simultaneously (default: use all).",
)
jobcontrol.add_argument(
'-q', '--quiet', action='store_true', help="Suppress INFO messages"
)
jobcontrol.add_argument(
'-v',
'--verbose',
type=numeric(int, 0, 2),
default=0,
const=1,
nargs='?',
help="Print more verbose messages for each additional verbose level. Use "
"`-v 1` typically for much more detailed logging. Higher numbers "
"are probably only useful in debugging.",
)
jobcontrol.add_argument(
'--no-progress-bar',
action='store_false',
dest='progress_bar',
help=argparse.SUPPRESS,
)
jobcontrol.add_argument('--use-threads', action='store_true', help=argparse.SUPPRESS)
metadata = parser.add_argument_group(
"Metadata options",
"Set output PDF/A metadata (default: copy input document's metadata)",
)
metadata.add_argument(
'--title', type=str, help="Set document title (place multiple words in quotes)"
)
metadata.add_argument('--author', type=str, help="Set document author")
metadata.add_argument('--subject', type=str, help="Set document subject description")
metadata.add_argument('--keywords', type=str, help="Set document keywords")
preprocessing = parser.add_argument_group(
"Image preprocessing options",
"Options to improve the quality of the final PDF and OCR",
)
preprocessing.add_argument(
'-r',
'--rotate-pages',
action='store_true',
help="Automatically rotate pages based on detected text orientation",
)
preprocessing.add_argument(
'--remove-background',
action='store_true',
help="Attempt to remove background from gray or color pages, setting it "
"to white ",
)
preprocessing.add_argument(
'-d', '--deskew', action='store_true', help="Deskew each page before performing OCR"
)
preprocessing.add_argument(
'-c',
'--clean',
action='store_true',
help="Clean pages from scanning artifacts before performing OCR, and send "
"the cleaned page to OCR, but do not include the cleaned page in "
"the output",
)
preprocessing.add_argument(
'-i',
'--clean-final',
action='store_true',
help="Clean page as above, and incorporate the cleaned image in the final "
"PDF. Might remove desired content.",
)
preprocessing.add_argument(
'--unpaper-args',
type=str,
default=None,
help="A quoted string of arguments to pass to unpaper. Requires --clean. "
"Example: --unpaper-args '--layout double'.",
)
preprocessing.add_argument(
'--oversample',
metavar='DPI',
type=numeric(int, 0, 5000),
default=0,
help="Oversample images to at least the specified DPI, to improve OCR "
"results slightly",
)
preprocessing.add_argument(
'--remove-vectors',
action='store_true',
help="EXPERIMENTAL. Mask out any vector objects in the PDF so that they "
"will not be included in OCR. This can eliminate false characters.",
)
preprocessing.add_argument(
'--threshold',
action='store_true',
help="EXPERIMENTAL. Threshold image to 1bpp before sending it to Tesseract for OCR. Can "
"improve OCR quality compared to Tesseract's thresholder.",
)
ocrsettings = parser.add_argument_group("OCR options", "Control how OCR is applied")
ocrsettings.add_argument(
'-f',
'--force-ocr',
action='store_true',
help="Rasterize any text or vector objects on each page, apply OCR, and "
"save the rastered output (this rewrites the PDF)",
)
ocrsettings.add_argument(
'-s',
'--skip-text',
action='store_true',
help="Skip OCR on any pages that already contain text, but include the "
"page in final output; useful for PDFs that contain a mix of "
"images, text pages, and/or previously OCRed pages",
)
ocrsettings.add_argument(
'--redo-ocr',
action='store_true',
help="Attempt to detect and remove the hidden OCR layer from files that "
"were previously OCRed with OCRmyPDF or another program. Apply OCR "
"to text found in raster images. Existing visible text objects will "
"not be changed. If there is no existing OCR, OCR will be added.",
)
ocrsettings.add_argument(
'--skip-big',
type=numeric(float, 0, 5000),
metavar='MPixels',
help="Skip OCR on pages larger than the specified amount of megapixels, "
"but include skipped pages in final output",
)
optimizing = parser.add_argument_group(
"Optimization options", "Control how the PDF is optimized after OCR"
)
optimizing.add_argument(
'-O',
'--optimize',
type=int,
choices=range(0, 4),
default=1,
help=(
"Control how PDF is optimized after processing:"
"0 - do not optimize; "
"1 - do safe, lossless optimizations (default); "
"2 - do some lossy optimizations; "
"3 - do aggressive lossy optimizations (including lossy JBIG2)"
),
)
optimizing.add_argument(
'--jpeg-quality',
type=numeric(int, 0, 100),
default=0,
metavar='Q',
help=(
"Adjust JPEG quality level for JPEG optimization. "
"100 is best quality and largest output size; "
"1 is lowest quality and smallest output; "
"0 uses the default."
),
)
optimizing.add_argument(
'--jpg-quality',
type=numeric(int, 0, 100),
default=0,
metavar='Q',
dest='jpeg_quality',
help=argparse.SUPPRESS, # Alias for --jpeg-quality
)
optimizing.add_argument(
'--png-quality',
type=numeric(int, 0, 100),
default=0,
metavar='Q',
help=(
"Adjust PNG quality level to use when quantizing PNGs. "
"Values have same meaning as with --jpeg-quality"
),
)
optimizing.add_argument(
'--jbig2-lossy',
action='store_true',
help=(
"Enable JBIG2 lossy mode (better compression, not suitable for some "
"use cases - see documentation)."
),
)
optimizing.add_argument(
'--jbig2-page-group-size',
type=numeric(int, 1, 10000),
default=0,
metavar='N',
# Adjust number of pages to consider at once for JBIG2 compression
help=argparse.SUPPRESS,
)
advanced = parser.add_argument_group(
"Advanced", "Advanced options to control Tesseract's OCR behavior"
)
advanced.add_argument(
'--pages',
type=str,
help="Limit OCR to the specified pages (ranges or comma separated), skipping others",
)
advanced.add_argument(
'--max-image-mpixels',
action='store',
type=numeric(float, 0),
metavar='MPixels',
help="Set maximum number of pixels to unpack before treating an image as a "
"decompression bomb",
default=128.0,
)
advanced.add_argument(
'--tesseract-config',
action='append',
metavar='CFG',
default=[],
help="Additional Tesseract configuration files -- see documentation",
)
advanced.add_argument(
'--tesseract-pagesegmode',
action='store',
type=int,
metavar='PSM',
choices=range(0, 14),
help="Set Tesseract page segmentation mode (see tesseract --help)",
)
advanced.add_argument(
'--tesseract-oem',
action='store',
type=int,
metavar='MODE',
choices=range(0, 4),
help=(
"Set Tesseract 4.0 OCR engine mode: "
"0 - original Tesseract only; "
"1 - neural nets LSTM only; "
"2 - Tesseract + LSTM; "
"3 - default."
),
)
advanced.add_argument(
'--pdf-renderer',
choices=['auto', 'hocr', 'sandwich'],
default='auto',
help="Choose OCR PDF renderer - the default option is to let OCRmyPDF "
"choose. See documentation for discussion.",
)
advanced.add_argument(
'--tesseract-timeout',
default=180.0,
type=numeric(float, 0),
metavar='SECONDS',
help='Give up on OCR after the timeout, but copy the preprocessed page '
'into the final output',
)
advanced.add_argument(
'--rotate-pages-threshold',
default=14.0,
type=numeric(float, 0, 1000),
metavar='CONFIDENCE',
help="Only rotate pages when confidence is above this value (arbitrary "
"units reported by tesseract)",
)
advanced.add_argument(
'--pdfa-image-compression',
choices=['auto', 'jpeg', 'lossless'],
default='auto',
help="Specify how to compress images in the output PDF/A. 'auto' lets "
"OCRmyPDF decide. 'jpeg' changes all grayscale and color images to "
"JPEG compression. 'lossless' uses PNG-style lossless compression "
"for all images. Monochrome images are always compressed using a "
"lossless codec. Compression settings "
"are applied to all pages, including those for which OCR was "
"skipped. Not supported for --output-type=pdf ; that setting "
"preserves the original compression of all images.",
)
advanced.add_argument(
'--user-words',
metavar='FILE',
help="Specify the location of the Tesseract user words file. This is a "
"list of words Tesseract should consider while performing OCR in "
"addition to its standard language dictionaries. This can improve "
"OCR quality especially for specialized and technical documents.",
)
advanced.add_argument(
'--user-patterns',
metavar='FILE',
help="Specify the location of the Tesseract user patterns file.",
)
debugging = parser.add_argument_group(
"Debugging", "Arguments to help with troubleshooting and debugging"
)
debugging.add_argument(
'-k',
'--keep-temporary-files',
action='store_true',
help="Keep temporary files (helpful for debugging)",
)
debugging.add_argument('--tesseract-env', type=str, help=argparse.SUPPRESS)
src/ocrmypdf/exec/__init__.py
+17 -14
View File
@@ -17,6 +17,7 @@
"""Wrappers to manage subprocess calls"""
import logging
import os
import re
import sys
@@ -24,8 +25,10 @@ from subprocess import run, STDOUT, PIPE, CalledProcessError
from ..exceptions import MissingDependencyError, ExitCode
from collections.abc import Mapping
log = logging.Logger(__name__)
def get_version(program, *, version_arg='--version', regex=r'(\d+(\.\d+)*)'):
def get_version(program, *, version_arg='--version', regex=r'(\d+(\.\d+)*)', env=None):
"Get the version of the specified program"
args_prog = [program, version_arg]
try:
@@ -36,6 +39,7 @@ def get_version(program, *, version_arg='--version', regex=r'(\d+(\.\d+)*)'):
stdout=PIPE,
stderr=STDOUT,
check=True,
env=env,
)
output = proc.stdout
except FileNotFoundError as e:
@@ -115,7 +119,7 @@ def _get_platform():
return sys.platform
def _error_trailer(log, program, package, **kwargs):
def _error_trailer(program, package, **kwargs):
if isinstance(package, Mapping):
package = package[_get_platform()]
@@ -125,7 +129,7 @@ def _error_trailer(log, program, package, **kwargs):
log.info(linux_install_advice.format(**locals()))
def _error_missing_program(log, program, package, required_for, recommended):
def _error_missing_program(program, package, required_for, recommended):
if required_for:
log.error(missing_optional_program.format(**locals()))
elif recommended:
@@ -135,9 +139,7 @@ def _error_missing_program(log, program, package, required_for, recommended):
_error_trailer(**locals())
def _error_old_version(
log, program, package, need_version, found_version, required_for
):
def _error_old_version(program, package, need_version, found_version, required_for):
if required_for:
log.error(old_version_required_for.format(**locals()))
else:
@@ -147,27 +149,28 @@ def _error_old_version(
def check_external_program(
*,
log,
program,
package,
version_checker,
need_version,
required_for=None,
recommended=False,
**kwargs, # To consume log parameter
):
if kwargs:
if not 'log' in kwargs:
log.warning('check_external_program(log=...) is deprecated')
try:
found_version = version_checker()
except (CalledProcessError, FileNotFoundError, MissingDependencyError):
_error_missing_program(log, program, package, required_for, recommended)
_error_missing_program(program, package, required_for, recommended)
if not recommended:
sys.exit(ExitCode.missing_dependency)
raise MissingDependencyError()
return
if found_version < need_version:
_error_old_version(
log, program, package, need_version, found_version, required_for
)
_error_old_version(program, package, need_version, found_version, required_for)
if not recommended:
sys.exit(ExitCode.missing_dependency)
raise MissingDependencyError()
log.debug(f'Found {program} {found_version}')
log.debug('Found %s %s', program, found_version)
src/ocrmypdf/exec/ghostscript.py
+13 -6
View File
@@ -15,6 +15,7 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import re
from functools import lru_cache
from os import fspath
@@ -24,8 +25,11 @@ from tempfile import NamedTemporaryFile
from PIL import Image
from . import get_version
from ..exceptions import SubprocessOutputError
from . import get_version
gslog = logging.getLogger()
@lru_cache(maxsize=1)
@@ -132,6 +136,8 @@ def rasterize_pdf(
res = round(xres, 6), round(yres, 6)
if not page_dpi:
page_dpi = res
if not log:
log = gslog
with NamedTemporaryFile(delete=True) as tmp:
args_gs = (
@@ -160,12 +166,11 @@ def rasterize_pdf(
p = run(args_gs, stdout=PIPE, stderr=STDOUT, universal_newlines=True)
if _gs_error_reported(p.stdout):
log.error(p.stdout)
else:
elif p.stdout:
log.debug(p.stdout)
if p.returncode != 0:
log.error('Ghostscript rasterizing failed')
raise SubprocessOutputError()
raise SubprocessOutputError('Ghostscript rasterizing failed')
tmp.seek(0)
with Image.open(tmp) as im:
@@ -210,6 +215,9 @@ def generate_pdfa(
images entirely. (The feature was added in 9.23 but broken, and the 9.24
release of Ghostscript had regressions, so we don't support it until 9.25.)
"""
if not log:
log = gslog
compression_args = []
if compression == 'jpeg':
compression_args = [
@@ -287,5 +295,4 @@ def generate_pdfa(
# PDF/A - check PDF/A status elsewhere
copy(gs_pdf.name, fspath(output_file))
else:
log.error('Ghostscript PDF/A rendering failed')
raise SubprocessOutputError()
raise SubprocessOutputError('Ghostscript PDF/A rendering failed')
src/ocrmypdf/exec/tesseract.py
+74 -45
View File
@@ -22,18 +22,14 @@ from collections import namedtuple
from contextlib import suppress
from functools import lru_cache
from os import fspath
from subprocess import (
PIPE,
STDOUT,
CalledProcessError,
TimeoutExpired,
check_output,
run,
)
from textwrap import dedent
from subprocess import PIPE, STDOUT, CalledProcessError, TimeoutExpired, run
from . import get_version
from ..exceptions import MissingDependencyError, TesseractConfigError
from ..exceptions import (
MissingDependencyError,
TesseractConfigError,
SubprocessOutputError,
)
from ..helpers import page_number
OrientationConfidence = namedtuple('OrientationConfidence', ('angle', 'confidence'))
@@ -56,18 +52,16 @@ HOCR_TEMPLATE = """<?xml version="1.0" encoding="UTF-8"?>
"""
@lru_cache(maxsize=1)
def version():
return get_version('tesseract', regex=r'tesseract\s(.+)')
def version(tesseract_env=None):
return get_version('tesseract', regex=r'tesseract\s(.+)', env=tesseract_env)
def v4():
def v4(tesseract_env=None):
"Is this Tesseract v4.0?"
return version() >= '4'
return version(tesseract_env) >= '4'
@lru_cache(maxsize=1)
def has_textonly_pdf():
def has_textonly_pdf(tesseract_env=None):
"""Does Tesseract have textonly_pdf capability?
Available in v4.00.00alpha since January 2017. Best to
@@ -76,41 +70,51 @@ def has_textonly_pdf():
args_tess = ['tesseract', '--print-parameters', 'pdf']
params = ''
try:
params = check_output(args_tess, universal_newlines=True, stderr=STDOUT)
proc = run(
args_tess,
check=True,
universal_newlines=True,
stdout=PIPE,
stderr=STDOUT,
env=tesseract_env,
)
params = proc.stdout
except CalledProcessError as e:
print("Could not --print-parameters from tesseract", file=sys.stderr)
raise MissingDependencyError from e
raise MissingDependencyError(
"Could not --print-parameters from tesseract"
) from e
if 'textonly_pdf' in params:
return True
return False
@lru_cache(maxsize=1)
def languages():
def languages(tesseract_env=None):
def lang_error(output):
msg = dedent(
"""Tesseract failed to report available languages.
Output from Tesseract:
-----------
"""
msg = (
"Tesseract failed to report available languages.\n"
"Output from Tesseract:\n"
"-----------\n"
)
msg += output
print(msg, file=sys.stderr)
return msg
args_tess = ['tesseract', '--list-langs']
try:
proc = run(
args_tess, universal_newlines=True, stdout=PIPE, stderr=STDOUT, check=True
args_tess,
universal_newlines=True,
stdout=PIPE,
stderr=STDOUT,
check=True,
env=tesseract_env,
)
output = proc.stdout
except CalledProcessError as e:
lang_error(e.output)
raise MissingDependencyError from e
raise MissingDependencyError(lang_error(e.output)) from e
header, *rest = output.splitlines()
if not header.startswith('List of available languages'):
lang_error(output)
raise MissingDependencyError
raise MissingDependencyError(lang_error(output))
return set(lang.strip() for lang in rest)
@@ -123,7 +127,7 @@ def tess_base_args(langs, engine_mode):
return args
def get_orientation(input_file, engine_mode, timeout: float, log):
def get_orientation(input_file, engine_mode, timeout: float, log, tesseract_env=None):
args_tesseract = tess_base_args(['osd'], engine_mode) + [
'--psm',
'0',
@@ -132,7 +136,15 @@ def get_orientation(input_file, engine_mode, timeout: float, log):
]
try:
stdout = check_output(args_tesseract, stderr=STDOUT, timeout=timeout)
p = run(
args_tesseract,
stdout=PIPE,
stderr=STDOUT,
timeout=timeout,
check=True,
env=tesseract_env,
)
stdout = p.stdout
except TimeoutExpired:
return OrientationConfidence(angle=0, confidence=0.0)
except CalledProcessError as e:
@@ -142,7 +154,7 @@ def get_orientation(input_file, engine_mode, timeout: float, log):
or b'Image too large' in e.output
):
return OrientationConfidence(0, 0)
raise e from e
raise SubprocessOutputError() from e
else:
osd = {}
for line in stdout.decode().splitlines():
@@ -159,7 +171,7 @@ def get_orientation(input_file, engine_mode, timeout: float, log):
def tesseract_log_output(log, stdout, input_file):
prefix = f"{(page_number(input_file)):4d}: [tesseract] "
prefix = "[tesseract] "
try:
text = stdout.decode()
@@ -231,6 +243,7 @@ def generate_hocr(
pagesegmode: int,
user_words,
user_patterns,
tesseract_env,
log,
):
@@ -254,7 +267,15 @@ def generate_hocr(
args_tesseract.extend([input_file, prefix, 'hocr', 'txt'] + tessconfig)
try:
log.debug(args_tesseract)
stdout = check_output(args_tesseract, stderr=STDOUT, timeout=timeout)
p = run(
args_tesseract,
stdout=PIPE,
stderr=STDOUT,
timeout=timeout,
check=True,
env=tesseract_env,
)
stdout = p.stdout
except TimeoutExpired:
# Generate a HOCR file with no recognized text if tesseract times out
# Temporary workaround to hocrTransform not being able to function if
@@ -267,7 +288,7 @@ def generate_hocr(
_generate_null_hocr(output_hocr, output_sidecar, input_file)
return
raise e from e
raise SubprocessOutputError() from e
else:
tesseract_log_output(log, stdout, input_file)
# The sidecar text file will get the suffix .txt; rename it to
@@ -306,9 +327,10 @@ def generate_pdf(
pagesegmode: int,
user_words,
user_patterns,
tesseract_env,
log,
):
'''Use Tesseract to render a PDF.
"""Use Tesseract to render a PDF.
input_image -- image to analyze
skip_pdf -- if we time out, use this file as output
@@ -320,14 +342,14 @@ def generate_pdf(
tessconfig -- tesseract configuration
timeout -- timeout (seconds)
log -- logger object
'''
"""
args_tesseract = tess_base_args(language, engine_mode)
if pagesegmode is not None:
args_tesseract.extend(['--psm', str(pagesegmode)])
if text_only and has_textonly_pdf():
if text_only and has_textonly_pdf(tesseract_env):
args_tesseract.extend(['-c', 'textonly_pdf=1'])
if user_words:
@@ -342,10 +364,17 @@ def generate_pdf(
# to the number of order parameters here
args_tesseract.extend([input_image, prefix, 'pdf', 'txt'] + tessconfig)
try:
log.debug(args_tesseract)
stdout = check_output(args_tesseract, stderr=STDOUT, timeout=timeout)
p = run(
args_tesseract,
stdout=PIPE,
stderr=STDOUT,
timeout=timeout,
check=True,
env=tesseract_env,
)
stdout = p.stdout
if os.path.exists(prefix + '.txt'):
shutil.move(prefix + '.txt', output_text)
except TimeoutExpired:
@@ -356,6 +385,6 @@ def generate_pdf(
if b'Image too large' in e.output:
use_skip_page(text_only, skip_pdf, output_pdf, output_text)
return
raise e from e
raise SubprocessOutputError() from e
else:
tesseract_log_output(log, stdout, input_image)
src/ocrmypdf/exec/unpaper.py
+7 -9
View File
@@ -29,11 +29,7 @@ from tempfile import TemporaryDirectory
from . import get_version
from ..exceptions import MissingDependencyError, SubprocessOutputError
try:
from PIL import Image
except ImportError:
print("Could not find Python3 imaging library", file=sys.stderr)
raise
from PIL import Image
@lru_cache(maxsize=1)
@@ -55,16 +51,18 @@ def run(input_file, output_file, dpi, log, mode_args):
else:
im = im.convert(mode='RGB')
except IOError as e:
log.error("Could not convert image with type " + im.mode)
im.close()
raise MissingDependencyError() from e
raise MissingDependencyError(
"Could not convert image with type " + im.mode
) from e
try:
suffix = SUFFIXES[im.mode]
except KeyError:
log.error("Failed to convert image to a supported format.")
im.close()
raise MissingDependencyError() from e
raise MissingDependencyError(
"Failed to convert image to a supported format."
) from e
with TemporaryDirectory() as tmpdir:
input_pnm = os.path.join(tmpdir, f'input{suffix}')
src/ocrmypdf/filters.py
+10
View File
@@ -0,0 +1,10 @@
from PIL import Image
import PIL.ImageOps
def invert(im):
return PIL.ImageOps.invert(im.convert('L'))
def whiteout(im):
return Image.new(im.mode, im.size)
src/ocrmypdf/helpers.py
+22 -30
View File
@@ -15,32 +15,35 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import multiprocessing
import os
import sys
import warnings
from collections.abc import Iterable
from contextlib import suppress
from functools import partial, wraps
from functools import wraps
from pathlib import Path
log = logging.getLogger(__name__)
def re_symlink(input_file, soft_link_name, log=None):
def re_symlink(input_file, soft_link_name, *args, **kwargs):
"""
Helper function: relinks soft symbolic link if necessary
"""
if len(args) == 1 and isinstance(args[0], logging.Logger):
log.warning("Deprecated: re_symlink(,log)")
if 'log' in kwargs:
log.warning('Deprecated: re_symlink(...log=)')
input_file = os.fspath(input_file)
soft_link_name = os.fspath(soft_link_name)
if log is None:
prdebug = partial(print, file=sys.stderr)
else:
prdebug = log.debug
# Guard against soft linking to oneself
if input_file == soft_link_name:
prdebug(
"Warning: No symbolic link made. You are using "
+ "the original data directory as the working directory."
log.warning(
"No symbolic link made. You are using "
"the original data directory as the working directory."
)
return
@@ -48,16 +51,16 @@ def re_symlink(input_file, soft_link_name, log=None):
if os.path.lexists(soft_link_name):
# do not delete or overwrite real (non-soft link) file
if not os.path.islink(soft_link_name):
raise FileExistsError("%s exists and is not a link" % soft_link_name)
raise FileExistsError(f"{soft_link_name} exists and is not a link")
try:
os.unlink(soft_link_name)
except OSError:
prdebug("Can't unlink %s" % (soft_link_name))
log.debug("Can't unlink %s", soft_link_name)
if not os.path.exists(input_file):
raise FileNotFoundError("trying to create a broken symlink to %s" % input_file)
raise FileNotFoundError(f"trying to create a broken symlink to {input_file}")
prdebug("os.symlink(%s, %s)" % (input_file, soft_link_name))
log.debug("os.symlink(%s, %s)", input_file, soft_link_name)
# Create symbolic link using absolute path
os.symlink(os.path.abspath(input_file), soft_link_name)
@@ -67,6 +70,11 @@ def is_iterable_notstr(thing):
return isinstance(thing, Iterable) and not isinstance(thing, str)
def monotonic(L):
"""Does list increase monotonically?"""
return all(b > a for a, b in zip(L, L[1:]))
def page_number(input_file):
"""Get one-based page number implied by filename (000002.pdf -> 2)"""
return int(os.path.basename(os.fspath(input_file))[0:6])
@@ -77,14 +85,6 @@ def available_cpu_count():
return multiprocessing.cpu_count()
except NotImplementedError:
pass
try:
import psutil
return psutil.cpu_count()
except (ImportError, AttributeError):
pass
warnings.warn(
"Could not get CPU count. Assuming one (1) CPU." "Use -j N to set manually."
)
@@ -122,14 +122,6 @@ def is_file_writable(test_file):
return True
def flatten_groups(groups):
for obj in groups:
if is_iterable_notstr(obj):
yield from obj
else:
yield obj
def deprecated(func):
"""Warn that function is deprecated"""
src/ocrmypdf/optimize.py
+74 -46
View File
@@ -16,20 +16,21 @@
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import concurrent.futures
import logging
import sys
import tempfile
from collections import defaultdict
from os import fspath
from pathlib import Path
from PIL import Image
from tqdm import tqdm
import pikepdf
from pikepdf import Name, Dictionary, Array
from pikepdf import Name, Dictionary
from . import leptonica
from ._jobcontext import JobContext
from ._jobcontext import PDFContext
from .exec import jbig2enc, pngquant
from .exceptions import OutputFileAccessError
from .helpers import re_symlink
DEFAULT_JPEG_QUALITY = 75
@@ -267,9 +268,17 @@ def _produce_jbig2_images(jbig2_groups, root, log, options):
with concurrent.futures.ThreadPoolExecutor(max_workers=options.jobs) as executor:
futures = jbig2_futures(executor, root, jbig2_groups)
for future in concurrent.futures.as_completed(futures):
proc = future.result()
log.debug(proc.stderr.decode())
with tqdm(
total=len(jbig2_groups),
desc="JBIG2",
unit='item',
disable=not options.progress_bar,
) as pbar:
for future in concurrent.futures.as_completed(futures):
proc = future.result()
if proc.stderr:
log.debug(proc.stderr.decode())
pbar.update()
def convert_to_jbig2(pike, jbig2_groups, root, log, options):
@@ -311,7 +320,9 @@ def convert_to_jbig2(pike, jbig2_groups, root, log, options):
def transcode_jpegs(pike, jpegs, root, log, options):
for xref in jpegs:
for xref in tqdm(
jpegs, desc="JPEGs", unit='image', disable=not options.progress_bar
):
in_jpg = Path(jpg_name(root, xref))
opt_jpg = in_jpg.with_suffix('.opt.jpg')
@@ -340,15 +351,26 @@ def transcode_pngs(pike, images, image_name_fn, root, log, options):
with concurrent.futures.ThreadPoolExecutor(
max_workers=options.jobs
) as executor:
futures = []
for xref in images:
log.debug(image_name_fn(root, xref))
executor.submit(
pngquant.quantize,
image_name_fn(root, xref),
png_name(root, xref),
png_quality[0],
png_quality[1],
futures.append(
executor.submit(
pngquant.quantize,
image_name_fn(root, xref),
png_name(root, xref),
png_quality[0],
png_quality[1],
)
)
with tqdm(
desc="PNGs",
total=len(futures),
unit='image',
disable=not options.progress_bar,
) as pbar:
for _future in concurrent.futures.as_completed(futures):
pbar.update()
for xref in images:
im_obj = pike.get_object(xref, 0)
@@ -427,11 +449,11 @@ def transcode_pngs(pike, images, image_name_fn, root, log, options):
im_obj.write(compdata.read(), filter=Name.FlateDecode, decode_parms=dparms)
def optimize(input_file, output_file, log, context):
options = context.get_options()
def optimize(input_file, output_file, context):
log = context.log
options = context.options
if options.optimize == 0:
re_symlink(input_file, output_file, log)
re_symlink(input_file, output_file)
return
if options.jpeg_quality == 0:
@@ -441,40 +463,44 @@ def optimize(input_file, output_file, log, context):
if options.jbig2_page_group_size == 0:
options.jbig2_page_group_size = 10 if options.jbig2_lossy else 1
pike = pikepdf.Pdf.open(input_file)
with pikepdf.Pdf.open(input_file) as pike:
root = Path(output_file).parent / 'images'
root.mkdir(exist_ok=True)
root = Path(output_file).parent / 'images'
root.mkdir(exist_ok=True)
jpegs, pngs = extract_images_generic(pike, root, log, options)
transcode_jpegs(pike, jpegs, root, log, options)
# if options.optimize >= 2:
# Try pngifying the jpegs
# transcode_pngs(pike, jpegs, jpg_name, root, log, options)
transcode_pngs(pike, pngs, png_name, root, log, options)
jpegs, pngs = extract_images_generic(pike, root, log, options)
transcode_jpegs(pike, jpegs, root, log, options)
# if options.optimize >= 2:
# Try pngifying the jpegs
# transcode_pngs(pike, jpegs, jpg_name, root, log, options)
transcode_pngs(pike, pngs, png_name, root, log, options)
jbig2_groups = extract_images_jbig2(pike, root, log, options)
convert_to_jbig2(pike, jbig2_groups, root, log, options)
jbig2_groups = extract_images_jbig2(pike, root, log, options)
convert_to_jbig2(pike, jbig2_groups, root, log, options)
target_file = Path(output_file).with_suffix('.opt.pdf')
pike.remove_unreferenced_resources()
pike.save(
target_file,
preserve_pdfa=True,
object_stream_mode=pikepdf.ObjectStreamMode.generate,
)
target_file = Path(output_file).with_suffix('.opt.pdf')
pike.remove_unreferenced_resources()
pike.save(
target_file,
preserve_pdfa=True,
object_stream_mode=pikepdf.ObjectStreamMode.generate,
)
input_size = Path(input_file).stat().st_size
output_size = Path(target_file).stat().st_size
if output_size == 0:
raise OutputFileAccessError(
f"Output file not created after optimizing. We probably ran "
f"out of disk space in the temporary folder: {tempfile.gettempdir()}."
)
ratio = input_size / output_size
savings = 1 - output_size / input_size
log.info(f"Optimize ratio: {ratio:.2f} savings: {(100 * savings):.1f}%")
if savings < 0:
log.info("Optimize did not improve the file - discarded")
re_symlink(input_file, output_file, log)
re_symlink(input_file, output_file)
else:
re_symlink(target_file, output_file, log)
re_symlink(target_file, output_file)
def main(infile, outfile, level, jobs=1):
@@ -484,30 +510,32 @@ def main(infile, outfile, level, jobs=1):
class OptimizeOptions:
"""Emulate ocrmypdf's options"""
def __init__(self, jobs, optimize, jpeg_quality, png_quality, jb2lossy):
def __init__(
self, input_file, jobs, optimize, jpeg_quality, png_quality, jb2lossy
):
self.input_file = input_file
self.jobs = jobs
self.optimize = optimize
self.jpeg_quality = jpeg_quality
self.png_quality = png_quality
self.jbig2_page_group_size = 0
self.jbig2_lossy = jb2lossy
self.quiet = True
self.progress_bar = False
logging.basicConfig(level=logging.DEBUG)
log = logging.getLogger()
ctx = JobContext()
options = OptimizeOptions(
input_file=infile,
jobs=jobs,
optimize=int(level),
jpeg_quality=0, # Use default
png_quality=0,
jb2lossy=False,
)
ctx.set_options(options)
with TemporaryDirectory() as td:
context = PDFContext(options, td, infile, None)
tmpout = Path(td) / 'out.pdf'
optimize(infile, tmpout, log, ctx)
optimize(infile, tmpout, context)
copy(fspath(tmpout), fspath(outfile))
src/ocrmypdf/pdfinfo/__init__.py
+1 -792
View File
@@ -16,795 +16,4 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
from collections import namedtuple
from decimal import Decimal
from enum import Enum
from math import hypot, isclose
from os import fspath
from pathlib import Path
from unittest.mock import Mock
from warnings import warn
import re
from pikepdf import PdfMatrix
import pikepdf
from . import ghosttext
from .layout import get_page_analysis, get_text_boxes
from ..exceptions import EncryptedPdfError, MissingDependencyError
Colorspace = Enum('Colorspace', 'gray rgb cmyk lab icc index sep devn pattern jpeg2000')
Encoding = Enum(
'Encoding', 'ccitt jpeg jpeg2000 jbig2 asciihex ascii85 lzw flate ' + 'runlength'
)
FRIENDLY_COLORSPACE = {
'/DeviceGray': Colorspace.gray,
'/CalGray': Colorspace.gray,
'/DeviceRGB': Colorspace.rgb,
'/CalRGB': Colorspace.rgb,
'/DeviceCMYK': Colorspace.cmyk,
'/Lab': Colorspace.lab,
'/ICCBased': Colorspace.icc,
'/Indexed': Colorspace.index,
'/Separation': Colorspace.sep,
'/DeviceN': Colorspace.devn,
'/Pattern': Colorspace.pattern,
'/G': Colorspace.gray, # Abbreviations permitted in inline images
'/RGB': Colorspace.rgb,
'/CMYK': Colorspace.cmyk,
'/I': Colorspace.index,
}
FRIENDLY_ENCODING = {
'/CCITTFaxDecode': Encoding.ccitt,
'/DCTDecode': Encoding.jpeg,
'/JPXDecode': Encoding.jpeg2000,
'/JBIG2Decode': Encoding.jbig2,
'/CCF': Encoding.ccitt, # Abbreviations permitted in inline images
'/DCT': Encoding.jpeg,
'/AHx': Encoding.asciihex,
'/A85': Encoding.ascii85,
'/LZW': Encoding.lzw,
'/Fl': Encoding.flate,
'/RL': Encoding.runlength,
}
FRIENDLY_COMP = {
Colorspace.gray: 1,
Colorspace.rgb: 3,
Colorspace.cmyk: 4,
Colorspace.lab: 3,
Colorspace.index: 1,
}
UNIT_SQUARE = (1.0, 0.0, 0.0, 1.0, 0.0, 0.0)
def _is_unit_square(shorthand):
values = map(float, shorthand)
pairwise = zip(values, UNIT_SQUARE)
return all([isclose(a, b, rel_tol=1e-3) for a, b in pairwise])
XobjectSettings = namedtuple('XobjectSettings', ['name', 'shorthand', 'stack_depth'])
InlineSettings = namedtuple('InlineSettings', ['iimage', 'shorthand', 'stack_depth'])
ContentsInfo = namedtuple(
'ContentsInfo', ['xobject_settings', 'inline_images', 'found_vector']
)
TextboxInfo = namedtuple('TextboxInfo', ['bbox', 'is_visible', 'is_corrupt'])
class VectorInfo:
def __init__(self):
pass
def _normalize_stack(graphobjs):
"""Convert runs of qQ's in the stack into single graphobjs"""
for operands, operator in graphobjs:
operator = str(operator)
if re.match(r'Q*q+$', operator): # Zero or more Q, one or more q
for char in operator: # Split into individual
yield ([], char) # Yield individual
else:
yield (operands, operator)
def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE):
"""Interpret the PDF content stream.
The stack represents the state of the PDF graphics stack. We are only
interested in the current transformation matrix (CTM) so we only track
this object; a full implementation would need to track many other items.
The CTM is initialized to the mapping from user space to device space.
PDF units are 1/72". In a PDF viewer or printer this matrix is initialized
to the transformation to device space. For example if set to
(1/72, 0, 0, 1/72, 0, 0) then all units would be calculated in inches.
Images are always considered to be (0, 0) -> (1, 1). Before drawing an
image there should be a 'cm' that sets up an image coordinate system
where drawing from (0, 0) -> (1, 1) will draw on the desired area of the
page.
PDF units suit our needs so we initialize ctm to the identity matrix.
According to the PDF specification, the maximum stack depth is 32. Other
viewers tolerate some amount beyond this. We issue a warning if the
stack depth exceeds the spec limit and set a hard limit beyond this to
bound our memory requirements. If the stack underflows behavior is
undefined in the spec, but we just pretend nothing happened and leave the
CTM unchanged.
"""
stack = []
ctm = PdfMatrix(initial_shorthand)
xobject_settings = []
inline_images = []
found_vector = False
vector_ops = set('S s f F f* B B* b b*'.split())
image_ops = set('BI ID EI q Q Do cm'.split())
operator_whitelist = ' '.join(vector_ops | image_ops)
for n, graphobj in enumerate(
_normalize_stack(
pikepdf.parse_content_stream(contentstream, operator_whitelist)
)
):
operands, operator = graphobj
if operator == 'q':
stack.append(ctm)
if len(stack) > 32: # See docstring
if len(stack) > 128:
raise RuntimeError(
"PDF graphics stack overflowed hard limit, operator %i" % n
)
warn("PDF graphics stack overflowed spec limit")
elif operator == 'Q':
try:
ctm = stack.pop()
except IndexError:
# Keeping the ctm the same seems to be the only sensible thing
# to do. Just pretend nothing happened, keep calm and carry on.
warn("PDF graphics stack underflowed - PDF may be malformed")
elif operator == 'cm':
ctm = PdfMatrix(operands) @ ctm
elif operator == 'Do':
image_name = operands[0]
settings = XobjectSettings(
name=image_name, shorthand=ctm.shorthand, stack_depth=len(stack)
)
xobject_settings.append(settings)
elif operator == 'INLINE IMAGE': # BI/ID/EI are grouped into this
iimage = operands[0]
inline = InlineSettings(
iimage=iimage, shorthand=ctm.shorthand, stack_depth=len(stack)
)
inline_images.append(inline)
elif operator in vector_ops:
found_vector = True
return ContentsInfo(
xobject_settings=xobject_settings,
inline_images=inline_images,
found_vector=found_vector,
)
def _get_dpi(ctm_shorthand, image_size):
"""Given the transformation matrix and image size, find the image DPI.
PDFs do not include image resolution information within image data.
Instead, the PDF page content stream describes the location where the
image will be rasterized, and the effective resolution is the ratio of the
pixel size to raster target size.
Normally a scanned PDF has the paper size set appropriately but this is
not guaranteed. The most common case is a cropped image will change the
page size (/CropBox) without altering the page content stream. That means
it is not sufficient to assume that the image fills the page, even though
that is the most common case.
A PDF image may be scaled (always), cropped, translated, rotated in place
to an arbitrary angle (rarely) and skewed. Only equal area mappings can
be expressed, that is, it is not necessary to consider distortions where
the effective DPI varies with position.
To determine the image scale, transform an offset axis vector v0 (0, 0),
width-axis vector v0 (1, 0), height-axis vector vh (0, 1) with the matrix,
which gives the dimensions of the image in PDF units. From there we can
compare to actual image dimensions. PDF uses
row vector * matrix_tranposed unlike the traditional
matrix * column vector.
The offset, width and height vectors can be combined in a matrix and
multiplied by the transform matrix. Then we want to calculated
magnitude(width_vector - offset_vector)
and
magnitude(height_vector - offset_vector)
When the above is worked out algebraically, the effect of translation
cancels out, and the vector magnitudes become functions of the nonzero
transformation matrix indices. The results of the derivation are used
in this code.
pdfimages -list does calculate the DPI in some way that is not completely
naive, but it does not get the DPI of rotated images right, so cannot be
used anymore to validate this. Photoshop works, or using Acrobat to
rotate the image back to normal.
It does not matter if the image is partially cropped, or even out of the
/MediaBox.
"""
a, b, c, d, _, _ = ctm_shorthand
# Calculate the width and height of the image in PDF units
image_drawn_width = hypot(a, b)
image_drawn_height = hypot(c, d)
# The scale of the image is pixels per unit of default user space (1/72")
scale_w = image_size[0] / image_drawn_width
scale_h = image_size[1] / image_drawn_height
# DPI = scale * 72
dpi_w = scale_w * 72.0
dpi_h = scale_h * 72.0
return dpi_w, dpi_h
class ImageInfo:
DPI_PREC = Decimal('1.000')
def __init__(self, *, name='', pdfimage=None, inline=None, shorthand=None):
self._name = str(name)
self._shorthand = shorthand
if inline is not None:
self._origin = 'inline'
pim = inline.iimage
elif pdfimage is not None:
self._origin = 'xobject'
pim = pikepdf.PdfImage(pdfimage)
self._width = pim.width
self._height = pim.height
# If /ImageMask is true, then this image is a stencil mask
# (Images that draw with this stencil mask will have a reference to
# it in their /Mask, but we don't actually need that information)
if pim.image_mask:
self._type = 'stencil'
else:
self._type = 'image'
self._bpc = int(pim.bits_per_component)
try:
self._enc = FRIENDLY_ENCODING.get(pim.filters[0], 'image')
except IndexError:
self._enc = '?'
try:
self._color = FRIENDLY_COLORSPACE.get(pim.colorspace, '?')
except NotImplementedError:
self._color = '?'
if self._enc == Encoding.jpeg2000:
self._color = Colorspace.jpeg2000
self._comp = FRIENDLY_COMP.get(self._color, '?')
# Bit of a hack... infer grayscale if component count is uncertain
# but encoding must be monochrome. This happens if a monochrome image
# has an ICC profile attached. Better solution would be to examine
# the ICC profile.
if self._comp == '?' and self._enc in (Encoding.ccitt, 'jbig2'):
self._comp = FRIENDLY_COMP[Colorspace.gray]
@property
def name(self):
return self._name
@property
def type_(self):
return self._type
@property
def width(self):
return self._width
@property
def height(self):
return self._height
@property
def bpc(self):
return self._bpc
@property
def color(self):
return self._color
@property
def comp(self):
return self._comp
@property
def enc(self):
return self._enc
@property
def xres(self):
return _get_dpi(self._shorthand, (self._width, self._height))[0]
@property
def yres(self):
return _get_dpi(self._shorthand, (self._width, self._height))[1]
def __repr__(self):
class_locals = {
attr: getattr(self, attr, None)
for attr in dir(self)
if not attr.startswith('_')
}
return (
"<ImageInfo '{name}' {type_} {width}x{height} {color} "
"{comp} {bpc} {enc} {xres}x{yres}>"
).format(**class_locals)
def _find_inline_images(contentsinfo):
"Find inline images in the contentstream"
for n, inline in enumerate(contentsinfo.inline_images):
yield ImageInfo(
name='inline-%02d' % n, shorthand=inline.shorthand, inline=inline
)
def _image_xobjects(container):
"""Search for all XObject-based images in the container
Usually the container is a page, but it could also be a Form XObject
that contains images. Filter out the Form XObjects which are dealt with
elsewhere.
Generate a sequence of tuples (image, xobj container), where container,
where xobj is the name of the object and image is the object itself,
since the object does not know its own name.
"""
if '/Resources' not in container:
return
resources = container['/Resources']
if '/XObject' not in resources:
return
xobjs = resources['/XObject'].as_dict()
for xobj in xobjs:
candidate = xobjs[xobj]
if not '/Subtype' in candidate:
continue
if candidate['/Subtype'] == '/Image':
pdfimage = candidate
yield (pdfimage, xobj)
def _find_regular_images(container, contentsinfo):
"""Find images stored in the container's /Resources /XObject
Usually the container is a page, but it could also be a Form XObject
that contains images.
Generates images with their DPI at time of drawing.
"""
for pdfimage, xobj in _image_xobjects(container):
# For each image that is drawn on this, check if we drawing the
# current image - yes this is O(n^2), but n == 1 almost always
for draw in contentsinfo.xobject_settings:
if draw.name != xobj:
continue
if draw.stack_depth == 0 and _is_unit_square(draw.shorthand):
# At least one PDF in the wild (and test suite) draws an image
# when the graphics stack depth is 0, meaning that the image
# gets drawn into a square of 1x1 PDF units (or 1/72",
# or 0.35 mm). The equivalent DPI will be >100,000. Exclude
# these from our DPI calculation for the page.
continue
yield ImageInfo(name=draw.name, pdfimage=pdfimage, shorthand=draw.shorthand)
def _find_form_xobject_images(pdf, container, contentsinfo):
"""Find any images that are in Form XObjects in the container
The container may be a page, or a parent Form XObject.
"""
if '/Resources' not in container:
return
resources = container['/Resources']
if '/XObject' not in resources:
return
xobjs = resources['/XObject'].as_dict()
for xobj in xobjs:
candidate = xobjs[xobj]
if candidate['/Subtype'] != '/Form':
continue
form_xobject = candidate
for settings in contentsinfo.xobject_settings:
if settings.name != xobj:
continue
# Find images once for each time this Form XObject is drawn.
# This could be optimized to cache the multiple drawing events
# but in practice both Form XObjects and multiple drawing of the
# same object are both very rare.
ctm_shorthand = settings.shorthand
yield from _process_content_streams(
pdf=pdf, container=form_xobject, shorthand=ctm_shorthand
)
def _process_content_streams(*, pdf, container, shorthand=None):
"""Find all individual instances of images drawn in the container
Usually the container is a page, but it may also be a Form XObject.
On a typical page images are stored inline or as regular images
in an XObject.
Form XObjects may include inline images, XObject images,
and recursively, other Form XObjects; and also vector graphic objects.
Every instance of an image being drawn somewhere is flattened and
treated as a unique image, since if the same image is drawn multiple times
on one page it may be drawn at differing resolutions, and our objective
is to find the resolution at which the page can be rastered without
downsampling.
"""
if container.get('/Type') == '/Page' and '/Contents' in container:
initial_shorthand = shorthand or UNIT_SQUARE
elif container.get('/Type') == '/XObject' and container['/Subtype'] == '/Form':
# Set the CTM to the state it was when the "Do" operator was
# encountered that is drawing this instance of the Form XObject
ctm = PdfMatrix(shorthand) if shorthand else PdfMatrix.identity()
# A Form XObject may provide its own matrix to map form space into
# user space. Get this if one exists
form_shorthand = container.get('/Matrix', PdfMatrix.identity())
form_matrix = PdfMatrix(form_shorthand)
# Concatenate form matrix with CTM to ensure CTM is correct for
# drawing this instance of the XObject
ctm = form_matrix @ ctm
initial_shorthand = ctm.shorthand
else:
return
contentsinfo = _interpret_contents(container, initial_shorthand)
if contentsinfo.found_vector:
yield VectorInfo()
yield from _find_inline_images(contentsinfo)
yield from _find_regular_images(container, contentsinfo)
yield from _find_form_xobject_images(pdf, container, contentsinfo)
def _page_has_text(text_blocks, page_width, page_height):
"""Smarter text detection that ignores text in margins"""
pw, ph = float(page_width), float(page_height)
margin_ratio = 0.125
interior_bbox = (
margin_ratio * pw, # left
(1 - margin_ratio) * ph, # top
(1 - margin_ratio) * pw, # right
margin_ratio * ph, # bottom (first quadrant: bottom < top)
)
def rects_intersect(a, b):
"""
Where (a,b) are 4-tuple rects (left-0, top-1, right-2, bottom-3)
https://stackoverflow.com/questions/306316/determine-if-two-rectangles-overlap-each-other
Formula assumes all boxes are in first quadrant
"""
return a[0] < b[2] and a[2] > b[0] and a[1] > b[3] and a[3] < b[1]
has_text = False
for bbox in text_blocks:
if rects_intersect(bbox, interior_bbox):
has_text = True
break
return has_text
def simplify_textboxes(miner, textbox_getter):
"""Extract only limited content from text boxes
We do this to save memory and ensure that our objects are pickleable.
"""
for box in textbox_getter(miner):
first_line = box._objs[0]
first_char = first_line._objs[0]
visible = first_char.rendermode != 3
corrupt = first_char.get_text() == '\ufffd'
yield TextboxInfo(box.bbox, visible, corrupt)
def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext):
pageinfo = {}
pageinfo['pageno'] = pageno
pageinfo['images'] = []
page = pdf.pages[pageno]
mediabox = [Decimal(d) for d in page.MediaBox.as_list()]
width_pt = mediabox[2] - mediabox[0]
height_pt = mediabox[3] - mediabox[1]
if xmltext is not None:
bboxes = ghosttext.page_get_textblocks(
fspath(infile), pageno, xmltext=xmltext, height=height_pt
)
pageinfo['bboxes'] = bboxes
else:
pscript5_mode = str(pdf.docinfo.get('/Creator')).startswith('PScript5')
miner = get_page_analysis(infile, pageno, pscript5_mode)
pageinfo['textboxes'] = list(simplify_textboxes(miner, get_text_boxes))
bboxes = (box.bbox for box in pageinfo['textboxes'])
pageinfo['has_text'] = _page_has_text(bboxes, width_pt, height_pt)
userunit = page.get('/UserUnit', Decimal(1.0))
if not isinstance(userunit, Decimal):
userunit = Decimal(userunit)
pageinfo['userunit'] = userunit
pageinfo['width_inches'] = width_pt * userunit / Decimal(72.0)
pageinfo['height_inches'] = height_pt * userunit / Decimal(72.0)
try:
pageinfo['rotate'] = int(page['/Rotate'])
except KeyError:
pageinfo['rotate'] = 0
userunit_shorthand = (userunit, 0, 0, userunit, 0, 0)
contentsinfo = [
ci
for ci in _process_content_streams(
pdf=pdf, container=page, shorthand=userunit_shorthand
)
]
pageinfo['has_vector'] = False
if any(isinstance(ci, VectorInfo) for ci in contentsinfo):
pageinfo['has_vector'] = True
pageinfo['images'] = [im for im in contentsinfo if isinstance(im, ImageInfo)]
if pageinfo['images']:
xres = Decimal(max(image.xres for image in pageinfo['images']))
yres = Decimal(max(image.yres for image in pageinfo['images']))
pageinfo['xres'], pageinfo['yres'] = xres, yres
pageinfo['width_pixels'] = int(round(xres * pageinfo['width_inches']))
pageinfo['height_pixels'] = int(round(yres * pageinfo['height_inches']))
return pageinfo
def _pdf_get_all_pageinfo(infile, detailed_analysis=False, log=None):
if not log:
log = Mock()
pdf = pikepdf.open(infile) # Do not close in this function
if pdf.is_encrypted:
pdf.close()
raise EncryptedPdfError() # Triggered by encryption with empty passwd
if detailed_analysis:
pages_xml = None
else:
pages_xml = ghosttext.extract_text_xml(infile, pdf, pageno=None, log=log)
pages = []
for n in range(len(pdf.pages)):
page_xml = pages_xml[n] if pages_xml else None
page = PageInfo(pdf, n, infile, page_xml, detailed_analysis)
pages.append(page)
return pages, pdf
class PageInfo:
def __init__(self, pdf, pageno, infile, xmltext, detailed_analysis=False):
self._pageno = pageno
self._infile = infile
self._pageinfo = _pdf_get_pageinfo(pdf, pageno, infile, xmltext)
self._detailed_analysis = detailed_analysis
@property
def pageno(self):
return self._pageno
@property
def has_text(self):
return self._pageinfo['has_text']
@property
def has_corrupt_text(self):
if not self._detailed_analysis:
raise NotImplementedError('Did not do detailed analysis')
return any(tbox.is_corrupt for tbox in self._pageinfo['textboxes'])
@property
def has_vector(self):
return self._pageinfo['has_vector']
@property
def width_inches(self):
return self._pageinfo['width_inches']
@property
def height_inches(self):
return self._pageinfo['height_inches']
@property
def width_pixels(self):
return int(round(self.width_inches * self.xres))
@property
def height_pixels(self):
return int(round(self.height_inches * self.yres))
@property
def rotation(self):
return self._pageinfo.get('rotate', None)
@rotation.setter
def rotation(self, value):
if value in (0, 90, 180, 270, 360, -90, -180, -270):
self._pageinfo['rotate'] = value
else:
raise ValueError("rotation must be a cardinal angle")
@property
def images(self):
return self._pageinfo['images']
def get_textareas(self, visible=None, corrupt=None):
def predicate(obj, want_visible, want_corrupt):
result = True
if want_visible is not None:
if obj.is_visible != want_visible:
result = False
if want_corrupt is not None:
if obj.is_corrupt != want_corrupt:
result = False
return result
if 'textboxes' not in self._pageinfo:
if visible is not None and corrupt is not None:
raise NotImplementedError('Ghostscript textboxes cannot be classified')
return self._pageinfo['bboxes']
return (
obj.bbox
for obj in self._pageinfo['textboxes']
if predicate(obj, visible, corrupt)
)
@property
def xres(self):
return self._pageinfo.get('xres', None)
@property
def yres(self):
return self._pageinfo.get('yres', None)
@property
def userunit(self):
return self._pageinfo.get('userunit', None)
@property
def min_version(self):
if self.userunit is not None:
return '1.6'
else:
return '1.5'
def __repr__(self):
return (
'<PageInfo ' 'pageno={} {}"x{}" rotation={} res={}x{} has_text={}>'
).format(
self.pageno,
self.width_inches,
self.height_inches,
self.rotation,
self.xres,
self.yres,
self.has_text,
)
class PdfInfo:
"""Get summary information about a PDF"""
def __init__(self, infile, detailed_page_analysis=False, log=None):
self._infile = infile
self._pages, pdf = _pdf_get_all_pageinfo(
infile, detailed_page_analysis, log=log
)
self._needs_rendering = pdf.root.get('/NeedsRendering', False)
self._has_acroform = False
if '/AcroForm' in pdf.root:
if len(pdf.root.AcroForm.get('/Fields', [])) > 0:
self._has_acroform = True
elif '/XFA' in pdf.root.AcroForm:
self._has_acroform = True
pdf.close()
@property
def pages(self):
return self._pages
@property
def min_version(self):
# The minimum PDF is the maximum version that any particular page needs
return max(page.min_version for page in self.pages)
@property
def has_userunit(self):
return any(page.userunit != 1.0 for page in self.pages)
@property
def has_acroform(self):
return self._has_acroform
@property
def filename(self):
if not isinstance(self._infile, (str, Path)):
raise NotImplementedError("can't get filename from stream")
return self._infile
@property
def needs_rendering(self):
return self._needs_rendering
def __getitem__(self, item):
return self._pages[item]
def __len__(self):
return len(self._pages)
def __repr__(self):
return f"<PdfInfo('...'), page count={len(self)}>"
def main():
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('infile')
args = parser.parse_args()
info = _pdf_get_all_pageinfo(args.infile)
from pprint import pprint
pprint(info)
if __name__ == '__main__':
main()
from .info import PdfInfo, Colorspace, Encoding
src/ocrmypdf/pdfinfo/ghosttext.py
+4 -1
View File
@@ -15,11 +15,14 @@
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
import logging
import re
import xml.etree.ElementTree as ET
from ..exec import ghostscript
gslog = logging.getLogger()
# Forgive me for I have sinned
# I am using regular expressions to parse XML. However the XML in this case,
# generated by Ghostscript, is self-consistent enough to be parseable.
@@ -74,7 +77,7 @@ def page_get_textblocks(infile, pageno, xmltext, height):
return [block for block in joined_blocks()]
def extract_text_xml(infile, pdf, pageno=None, log=None):
def extract_text_xml(infile, pdf, pageno=None, log=gslog):
existing_text = ghostscript.extract_text(infile, pageno=None)
existing_text = regex_remove_char_tags.sub(b' ', existing_text)
src/ocrmypdf/pdfinfo/info.py
+814
View File
@@ -0,0 +1,814 @@
#!/usr/bin/env python3
# © 2015 James R. Barlow: github.com/jbarlow83
#
# This file is part of OCRmyPDF.
#
# OCRmyPDF is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OCRmyPDF is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with OCRmyPDF. If not, see <http://www.gnu.org/licenses/>.
from collections import namedtuple
from decimal import Decimal
from enum import Enum
import logging
from math import hypot, isclose
from os import fspath
from pathlib import Path
from warnings import warn
import re
from pikepdf import PdfMatrix
import pikepdf
from tqdm import tqdm
from . import ghosttext
from .layout import get_page_analysis, get_text_boxes
from ocrmypdf.exceptions import EncryptedPdfError, MissingDependencyError
logger = logging.getLogger()
Colorspace = Enum('Colorspace', 'gray rgb cmyk lab icc index sep devn pattern jpeg2000')
Encoding = Enum(
'Encoding', 'ccitt jpeg jpeg2000 jbig2 asciihex ascii85 lzw flate ' + 'runlength'
)
FRIENDLY_COLORSPACE = {
'/DeviceGray': Colorspace.gray,
'/CalGray': Colorspace.gray,
'/DeviceRGB': Colorspace.rgb,
'/CalRGB': Colorspace.rgb,
'/DeviceCMYK': Colorspace.cmyk,
'/Lab': Colorspace.lab,
'/ICCBased': Colorspace.icc,
'/Indexed': Colorspace.index,
'/Separation': Colorspace.sep,
'/DeviceN': Colorspace.devn,
'/Pattern': Colorspace.pattern,
'/G': Colorspace.gray, # Abbreviations permitted in inline images
'/RGB': Colorspace.rgb,
'/CMYK': Colorspace.cmyk,
'/I': Colorspace.index,
}
FRIENDLY_ENCODING = {
'/CCITTFaxDecode': Encoding.ccitt,
'/DCTDecode': Encoding.jpeg,
'/JPXDecode': Encoding.jpeg2000,
'/JBIG2Decode': Encoding.jbig2,
'/CCF': Encoding.ccitt, # Abbreviations permitted in inline images
'/DCT': Encoding.jpeg,
'/AHx': Encoding.asciihex,
'/A85': Encoding.ascii85,
'/LZW': Encoding.lzw,
'/Fl': Encoding.flate,
'/RL': Encoding.runlength,
}
FRIENDLY_COMP = {
Colorspace.gray: 1,
Colorspace.rgb: 3,
Colorspace.cmyk: 4,
Colorspace.lab: 3,
Colorspace.index: 1,
}
UNIT_SQUARE = (1.0, 0.0, 0.0, 1.0, 0.0, 0.0)
def _is_unit_square(shorthand):
values = map(float, shorthand)
pairwise = zip(values, UNIT_SQUARE)
return all([isclose(a, b, rel_tol=1e-3) for a, b in pairwise])
XobjectSettings = namedtuple('XobjectSettings', ['name', 'shorthand', 'stack_depth'])
InlineSettings = namedtuple('InlineSettings', ['iimage', 'shorthand', 'stack_depth'])
ContentsInfo = namedtuple(
'ContentsInfo', ['xobject_settings', 'inline_images', 'found_vector']
)
TextboxInfo = namedtuple('TextboxInfo', ['bbox', 'is_visible', 'is_corrupt'])
class VectorInfo:
def __init__(self):
pass
def _normalize_stack(graphobjs):
"""Convert runs of qQ's in the stack into single graphobjs"""
for operands, operator in graphobjs:
operator = str(operator)
if re.match(r'Q*q+$', operator): # Zero or more Q, one or more q
for char in operator: # Split into individual
yield ([], char) # Yield individual
else:
yield (operands, operator)
def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE):
"""Interpret the PDF content stream.
The stack represents the state of the PDF graphics stack. We are only
interested in the current transformation matrix (CTM) so we only track
this object; a full implementation would need to track many other items.
The CTM is initialized to the mapping from user space to device space.
PDF units are 1/72". In a PDF viewer or printer this matrix is initialized
to the transformation to device space. For example if set to
(1/72, 0, 0, 1/72, 0, 0) then all units would be calculated in inches.
Images are always considered to be (0, 0) -> (1, 1). Before drawing an
image there should be a 'cm' that sets up an image coordinate system
where drawing from (0, 0) -> (1, 1) will draw on the desired area of the
page.
PDF units suit our needs so we initialize ctm to the identity matrix.
According to the PDF specification, the maximum stack depth is 32. Other
viewers tolerate some amount beyond this. We issue a warning if the
stack depth exceeds the spec limit and set a hard limit beyond this to
bound our memory requirements. If the stack underflows behavior is
undefined in the spec, but we just pretend nothing happened and leave the
CTM unchanged.
"""
stack = []
ctm = PdfMatrix(initial_shorthand)
xobject_settings = []
inline_images = []
found_vector = False
vector_ops = set('S s f F f* B B* b b*'.split())
image_ops = set('BI ID EI q Q Do cm'.split())
operator_whitelist = ' '.join(vector_ops | image_ops)
for n, graphobj in enumerate(
_normalize_stack(
pikepdf.parse_content_stream(contentstream, operator_whitelist)
)
):
operands, operator = graphobj
if operator == 'q':
stack.append(ctm)
if len(stack) > 32: # See docstring
if len(stack) > 128:
raise RuntimeError(
"PDF graphics stack overflowed hard limit, operator %i" % n
)
warn("PDF graphics stack overflowed spec limit")
elif operator == 'Q':
try:
ctm = stack.pop()
except IndexError:
# Keeping the ctm the same seems to be the only sensible thing
# to do. Just pretend nothing happened, keep calm and carry on.
warn("PDF graphics stack underflowed - PDF may be malformed")
elif operator == 'cm':
ctm = PdfMatrix(operands) @ ctm
elif operator == 'Do':
image_name = operands[0]
settings = XobjectSettings(
name=image_name, shorthand=ctm.shorthand, stack_depth=len(stack)
)
xobject_settings.append(settings)
elif operator == 'INLINE IMAGE': # BI/ID/EI are grouped into this
iimage = operands[0]
inline = InlineSettings(
iimage=iimage, shorthand=ctm.shorthand, stack_depth=len(stack)
)
inline_images.append(inline)
elif operator in vector_ops:
found_vector = True
return ContentsInfo(
xobject_settings=xobject_settings,
inline_images=inline_images,
found_vector=found_vector,
)
def _get_dpi(ctm_shorthand, image_size):
"""Given the transformation matrix and image size, find the image DPI.
PDFs do not include image resolution information within image data.
Instead, the PDF page content stream describes the location where the
image will be rasterized, and the effective resolution is the ratio of the
pixel size to raster target size.
Normally a scanned PDF has the paper size set appropriately but this is
not guaranteed. The most common case is a cropped image will change the
page size (/CropBox) without altering the page content stream. That means
it is not sufficient to assume that the image fills the page, even though
that is the most common case.
A PDF image may be scaled (always), cropped, translated, rotated in place
to an arbitrary angle (rarely) and skewed. Only equal area mappings can
be expressed, that is, it is not necessary to consider distortions where
the effective DPI varies with position.
To determine the image scale, transform an offset axis vector v0 (0, 0),
width-axis vector v0 (1, 0), height-axis vector vh (0, 1) with the matrix,
which gives the dimensions of the image in PDF units. From there we can
compare to actual image dimensions. PDF uses
row vector * matrix_tranposed unlike the traditional
matrix * column vector.
The offset, width and height vectors can be combined in a matrix and
multiplied by the transform matrix. Then we want to calculated
magnitude(width_vector - offset_vector)
and
magnitude(height_vector - offset_vector)
When the above is worked out algebraically, the effect of translation
cancels out, and the vector magnitudes become functions of the nonzero
transformation matrix indices. The results of the derivation are used
in this code.
pdfimages -list does calculate the DPI in some way that is not completely
naive, but it does not get the DPI of rotated images right, so cannot be
used anymore to validate this. Photoshop works, or using Acrobat to
rotate the image back to normal.
It does not matter if the image is partially cropped, or even out of the
/MediaBox.
"""
a, b, c, d, _, _ = ctm_shorthand
# Calculate the width and height of the image in PDF units
image_drawn_width = hypot(a, b)
image_drawn_height = hypot(c, d)
# The scale of the image is pixels per unit of default user space (1/72")
scale_w = image_size[0] / image_drawn_width
scale_h = image_size[1] / image_drawn_height
# DPI = scale * 72
dpi_w = scale_w * 72.0
dpi_h = scale_h * 72.0
return dpi_w, dpi_h
class ImageInfo:
DPI_PREC = Decimal('1.000')
def __init__(self, *, name='', pdfimage=None, inline=None, shorthand=None):
self._name = str(name)
self._shorthand = shorthand
if inline is not None:
self._origin = 'inline'
pim = inline.iimage
elif pdfimage is not None:
self._origin = 'xobject'
pim = pikepdf.PdfImage(pdfimage)
self._width = pim.width
self._height = pim.height
# If /ImageMask is true, then this image is a stencil mask
# (Images that draw with this stencil mask will have a reference to
# it in their /Mask, but we don't actually need that information)
if pim.image_mask:
self._type = 'stencil'
else:
self._type = 'image'
self._bpc = int(pim.bits_per_component)
try:
self._enc = FRIENDLY_ENCODING.get(pim.filters[0], 'image')
except IndexError:
self._enc = '?'
try:
self._color = FRIENDLY_COLORSPACE.get(pim.colorspace, '?')
except NotImplementedError:
self._color = '?'
if self._enc == Encoding.jpeg2000:
self._color = Colorspace.jpeg2000
self._comp = FRIENDLY_COMP.get(self._color, '?')
# Bit of a hack... infer grayscale if component count is uncertain
# but encoding must be monochrome. This happens if a monochrome image
# has an ICC profile attached. Better solution would be to examine
# the ICC profile.
if self._comp == '?' and self._enc in (Encoding.ccitt, Encoding.jbig2):
self._comp = FRIENDLY_COMP[Colorspace.gray]
@property
def name(self):
return self._name
@property
def type_(self):
return self._type
@property
def width(self):
return self._width
@property
def height(self):
return self._height
@property
def bpc(self):
return self._bpc
@property
def color(self):
return self._color
@property
def comp(self):
return self._comp
@property
def enc(self):
return self._enc
@property
def xres(self):
return _get_dpi(self._shorthand, (self._width, self._height))[0]
@property
def yres(self):
return _get_dpi(self._shorthand, (self._width, self._height))[1]
def __repr__(self):
class_locals = {
attr: getattr(self, attr, None)
for attr in dir(self)
if not attr.startswith('_')
}
return (
"<ImageInfo '{name}' {type_} {width}x{height} {color} "
"{comp} {bpc} {enc} {xres}x{yres}>"
).format(**class_locals)
def _find_inline_images(contentsinfo):
"Find inline images in the contentstream"
for n, inline in enumerate(contentsinfo.inline_images):
yield ImageInfo(
name='inline-%02d' % n, shorthand=inline.shorthand, inline=inline
)
def _image_xobjects(container):
"""Search for all XObject-based images in the container
Usually the container is a page, but it could also be a Form XObject
that contains images. Filter out the Form XObjects which are dealt with
elsewhere.
Generate a sequence of tuples (image, xobj container), where container,
where xobj is the name of the object and image is the object itself,
since the object does not know its own name.
"""
if '/Resources' not in container:
return
resources = container['/Resources']
if '/XObject' not in resources:
return
xobjs = resources['/XObject'].as_dict()
for xobj in xobjs:
candidate = xobjs[xobj]
if not '/Subtype' in candidate:
continue
if candidate['/Subtype'] == '/Image':
pdfimage = candidate
yield (pdfimage, xobj)
def _find_regular_images(container, contentsinfo):
"""Find images stored in the container's /Resources /XObject
Usually the container is a page, but it could also be a Form XObject
that contains images.
Generates images with their DPI at time of drawing.
"""
for pdfimage, xobj in _image_xobjects(container):
# For each image that is drawn on this, check if we drawing the
# current image - yes this is O(n^2), but n == 1 almost always
for draw in contentsinfo.xobject_settings:
if draw.name != xobj:
continue
if draw.stack_depth == 0 and _is_unit_square(draw.shorthand):
# At least one PDF in the wild (and test suite) draws an image
# when the graphics stack depth is 0, meaning that the image
# gets drawn into a square of 1x1 PDF units (or 1/72",
# or 0.35 mm). The equivalent DPI will be >100,000. Exclude
# these from our DPI calculation for the page.
continue
yield ImageInfo(name=draw.name, pdfimage=pdfimage, shorthand=draw.shorthand)
def _find_form_xobject_images(pdf, container, contentsinfo):
"""Find any images that are in Form XObjects in the container
The container may be a page, or a parent Form XObject.
"""
if '/Resources' not in container:
return
resources = container['/Resources']
if '/XObject' not in resources:
return
xobjs = resources['/XObject'].as_dict()
for xobj in xobjs:
candidate = xobjs[xobj]
if candidate['/Subtype'] != '/Form':
continue
form_xobject = candidate
for settings in contentsinfo.xobject_settings:
if settings.name != xobj:
continue
# Find images once for each time this Form XObject is drawn.
# This could be optimized to cache the multiple drawing events
# but in practice both Form XObjects and multiple drawing of the
# same object are both very rare.
ctm_shorthand = settings.shorthand
yield from _process_content_streams(
pdf=pdf, container=form_xobject, shorthand=ctm_shorthand
)
def _process_content_streams(*, pdf, container, shorthand=None):
"""Find all individual instances of images drawn in the container
Usually the container is a page, but it may also be a Form XObject.
On a typical page images are stored inline or as regular images
in an XObject.
Form XObjects may include inline images, XObject images,
and recursively, other Form XObjects; and also vector graphic objects.
Every instance of an image being drawn somewhere is flattened and
treated as a unique image, since if the same image is drawn multiple times
on one page it may be drawn at differing resolutions, and our objective
is to find the resolution at which the page can be rastered without
downsampling.
"""
if container.get('/Type') == '/Page' and '/Contents' in container:
initial_shorthand = shorthand or UNIT_SQUARE
elif container.get('/Type') == '/XObject' and container['/Subtype'] == '/Form':
# Set the CTM to the state it was when the "Do" operator was
# encountered that is drawing this instance of the Form XObject
ctm = PdfMatrix(shorthand) if shorthand else PdfMatrix.identity()
# A Form XObject may provide its own matrix to map form space into
# user space. Get this if one exists
form_shorthand = container.get('/Matrix', PdfMatrix.identity())
form_matrix = PdfMatrix(form_shorthand)
# Concatenate form matrix with CTM to ensure CTM is correct for
# drawing this instance of the XObject
ctm = form_matrix @ ctm
initial_shorthand = ctm.shorthand
else:
return
contentsinfo = _interpret_contents(container, initial_shorthand)
if contentsinfo.found_vector:
yield VectorInfo()
yield from _find_inline_images(contentsinfo)
yield from _find_regular_images(container, contentsinfo)
yield from _find_form_xobject_images(pdf, container, contentsinfo)
def _page_has_text(text_blocks, page_width, page_height):
"""Smarter text detection that ignores text in margins"""
pw, ph = float(page_width), float(page_height)
margin_ratio = 0.125
interior_bbox = (
margin_ratio * pw, # left
(1 - margin_ratio) * ph, # top
(1 - margin_ratio) * pw, # right
margin_ratio * ph, # bottom (first quadrant: bottom < top)
)
def rects_intersect(a, b):
"""
Where (a,b) are 4-tuple rects (left-0, top-1, right-2, bottom-3)
https://stackoverflow.com/questions/306316/determine-if-two-rectangles-overlap-each-other
Formula assumes all boxes are in first quadrant
"""
return a[0] < b[2] and a[2] > b[0] and a[1] > b[3] and a[3] < b[1]
has_text = False
for bbox in text_blocks:
if rects_intersect(bbox, interior_bbox):
has_text = True
break
return has_text
def simplify_textboxes(miner, textbox_getter):
"""Extract only limited content from text boxes
We do this to save memory and ensure that our objects are pickleable.
"""
for box in textbox_getter(miner):
first_line = box._objs[0]
first_char = first_line._objs[0]
visible = first_char.rendermode != 3
corrupt = first_char.get_text() == '\ufffd'
yield TextboxInfo(box.bbox, visible, corrupt)
def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext):
pageinfo = {}
pageinfo['pageno'] = pageno
pageinfo['images'] = []
page = pdf.pages[pageno]
mediabox = [Decimal(d) for d in page.MediaBox.as_list()]
width_pt = mediabox[2] - mediabox[0]
height_pt = mediabox[3] - mediabox[1]
if xmltext is not None:
bboxes = ghosttext.page_get_textblocks(
fspath(infile), pageno, xmltext=xmltext, height=height_pt
)
pageinfo['bboxes'] = bboxes
else:
pscript5_mode = str(pdf.docinfo.get('/Creator')).startswith('PScript5')
miner = get_page_analysis(infile, pageno, pscript5_mode)
pageinfo['textboxes'] = list(simplify_textboxes(miner, get_text_boxes))
bboxes = (box.bbox for box in pageinfo['textboxes'])
pageinfo['has_text'] = _page_has_text(bboxes, width_pt, height_pt)
userunit = page.get('/UserUnit', Decimal(1.0))
if not isinstance(userunit, Decimal):
userunit = Decimal(userunit)
pageinfo['userunit'] = userunit
pageinfo['width_inches'] = width_pt * userunit / Decimal(72.0)
pageinfo['height_inches'] = height_pt * userunit / Decimal(72.0)
try:
pageinfo['rotate'] = int(page['/Rotate'])
except KeyError:
pageinfo['rotate'] = 0
userunit_shorthand = (userunit, 0, 0, userunit, 0, 0)
contentsinfo = [
ci
for ci in _process_content_streams(
pdf=pdf, container=page, shorthand=userunit_shorthand
)
]
pageinfo['has_vector'] = False
if any(isinstance(ci, VectorInfo) for ci in contentsinfo):
pageinfo['has_vector'] = True
pageinfo['images'] = [im for im in contentsinfo if isinstance(im, ImageInfo)]
if pageinfo['images']:
xres = Decimal(max(image.xres for image in pageinfo['images']))
yres = Decimal(max(image.yres for image in pageinfo['images']))
pageinfo['xres'], pageinfo['yres'] = xres, yres
pageinfo['width_pixels'] = int(round(xres * pageinfo['width_inches']))
pageinfo['height_pixels'] = int(round(yres * pageinfo['height_inches']))
return pageinfo
def _pdf_get_all_pageinfo(infile, detailed_analysis=False, log=None, progbar=False):
pdf = pikepdf.open(infile) # Do not close in this function
if pdf.is_encrypted:
pdf.close()
raise EncryptedPdfError() # Triggered by encryption with empty passwd
if detailed_analysis:
pages_xml = None
else:
pages_xml = ghosttext.extract_text_xml(infile, pdf, pageno=None, log=log)
pages = []
for n, _ in tqdm(
enumerate(pdf.pages),
total=len(pdf.pages),
desc="Scan",
unit='page',
disable=not progbar,
):
page_xml = pages_xml[n] if pages_xml else None
page = PageInfo(pdf, n, infile, page_xml, detailed_analysis)
pages.append(page)
return pages, pdf
class PageInfo:
def __init__(self, pdf, pageno, infile, xmltext, detailed_analysis=False):
self._pageno = pageno
self._infile = infile
self._pageinfo = _pdf_get_pageinfo(pdf, pageno, infile, xmltext)
self._detailed_analysis = detailed_analysis
@property
def pageno(self):
return self._pageno
@property
def has_text(self):
return self._pageinfo['has_text']
@property
def has_corrupt_text(self):
if not self._detailed_analysis:
raise NotImplementedError('Did not do detailed analysis')
return any(tbox.is_corrupt for tbox in self._pageinfo['textboxes'])
@property
def has_vector(self):
return self._pageinfo['has_vector']
@property
def width_inches(self):
return self._pageinfo['width_inches']
@property
def height_inches(self):
return self._pageinfo['height_inches']
@property
def width_pixels(self):
return int(round(self.width_inches * self.xres))
@property
def height_pixels(self):
return int(round(self.height_inches * self.yres))
@property
def rotation(self):
return self._pageinfo.get('rotate', None)
@rotation.setter
def rotation(self, value):
if value in (0, 90, 180, 270, 360, -90, -180, -270):
self._pageinfo['rotate'] = value
else:
raise ValueError("rotation must be a cardinal angle")
@property
def images(self):
return self._pageinfo['images']
def get_textareas(self, visible=None, corrupt=None):
def predicate(obj, want_visible, want_corrupt):
result = True
if want_visible is not None:
if obj.is_visible != want_visible:
result = False
if want_corrupt is not None:
if obj.is_corrupt != want_corrupt:
result = False
return result
if 'textboxes' not in self._pageinfo:
if visible is not None and corrupt is not None:
raise NotImplementedError('Ghostscript textboxes cannot be classified')
return self._pageinfo['bboxes']
return (
obj.bbox
for obj in self._pageinfo['textboxes']
if predicate(obj, visible, corrupt)
)
@property
def xres(self):
return self._pageinfo.get('xres', None)
@property
def yres(self):
return self._pageinfo.get('yres', None)
@property
def userunit(self):
return self._pageinfo.get('userunit', None)
@property
def min_version(self):
if self.userunit is not None:
return '1.6'
else:
return '1.5'
def __repr__(self):
return (
'<PageInfo ' 'pageno={} {}"x{}" rotation={} res={}x{} has_text={}>'
).format(
self.pageno,
self.width_inches,
self.height_inches,
self.rotation,
self.xres,
self.yres,
self.has_text,
)
class PdfInfo:
"""Get summary information about a PDF"""
def __init__(self, infile, detailed_page_analysis=False, log=logger, progbar=False):
self._infile = infile
self._pages, pdf = _pdf_get_all_pageinfo(
infile, detailed_page_analysis, log=log, progbar=progbar
)
self._needs_rendering = pdf.root.get('/NeedsRendering', False)
self._has_acroform = False
if '/AcroForm' in pdf.root:
if len(pdf.root.AcroForm.get('/Fields', [])) > 0:
self._has_acroform = True
elif '/XFA' in pdf.root.AcroForm:
self._has_acroform = True
pdf.close()
@property
def pages(self):
return self._pages
@property
def min_version(self):
# The minimum PDF is the maximum version that any particular page needs
return max(page.min_version for page in self.pages)
@property
def has_userunit(self):
return any(page.userunit != 1.0 for page in self.pages)
@property
def has_acroform(self):
return self._has_acroform
@property
def filename(self):
if not isinstance(self._infile, (str, Path)):
raise NotImplementedError("can't get filename from stream")
return self._infile
@property
def needs_rendering(self):
return self._needs_rendering
def __getitem__(self, item):
return self._pages[item]
def __len__(self):
return len(self._pages)
def __repr__(self):
return f"<PdfInfo('...'), page count={len(self)}>"
def main():
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('infile')
args = parser.parse_args()
info = _pdf_get_all_pageinfo(args.infile)
from pprint import pprint
pprint(info)
if __name__ == '__main__':
main()
src/ocrmypdf/pdfinfo/layout.py
+2 -1
View File
@@ -172,6 +172,7 @@ class LTStateAwareChar(LTChar):
- the Unicode mapping is known, and both have the same render mode
- the Unicode mapping is unknown but both are part of the same font
"""
# pylint: disable=protected-access
both_unicode_mapped = isinstance(self._text, str) and isinstance(obj._text, str)
try:
if both_unicode_mapped:
@@ -184,7 +185,7 @@ class LTStateAwareChar(LTChar):
def get_text(self):
if isinstance(self._text, tuple):
return '�'
return '\ufffd' # standard 'Unknown symbol'
return self._text
def __repr__(self):
tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/pdf.bin → tests/cache/2400dpi/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin → tests/cache/2400dpi/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stdout.bin → tests/cache/2400dpi/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin → tests/cache/2400dpi/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/txt.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/hocr.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/hocr.bin
Vendored
+8 -8
View File
@@ -9,23 +9,23 @@
<meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word ocrp_wconf'/>
</head>
<body>
<div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.jz4b2s0s/000001.ocr.png"; bbox 0 0 1000 800; ppageno 0'>
<div class='ocr_page' id='page_1' title='image "/var/folders/2s/7t022mgj0h5cprbq0dtb1ksm0000gn/T/com.github.ocrmypdf.xam82ph5/000001_ocr.png"; bbox 0 0 1000 800; ppageno 0'>
<div class='ocr_carea' id='block_1_1' title="bbox 296 96 704 504">
<p class='ocr_par' id='par_1_1' lang='eng' title="bbox 296 96 704 504">
<span class='ocr_line' id='line_1_1' title="bbox 296 96 704 504; baseline 0 296; x_size 169.33333; x_descenders 42.333332; x_ascenders 42.333336">
<span class='ocrx_word' id='word_1_1' title='bbox 296 96 704 504; x_wconf 95'><strong><em> </em></strong></span>
<span class='ocrx_word' id='word_1_1' title='bbox 296 96 704 504; x_wconf 95'> </span>
</span>
</p>
</div>
<div class='ocr_carea' id='block_1_2' title="bbox 150 592 841 622">
<p class='ocr_par' id='par_1_2' lang='eng' title="bbox 150 592 841 622">
<span class='ocr_line' id='line_1_2' title="bbox 150 592 841 622; baseline 0 -6; x_size 30; x_descenders 6; x_ascenders 8">
<span class='ocrx_word' id='word_1_2' title='bbox 150 592 230 616; x_wconf 96'><strong><em>This</em></strong></span>
<span class='ocrx_word' id='word_1_3' title='bbox 260 592 384 616; x_wconf 95'><strong><em>should</em></strong></span>
<span class='ocrx_word' id='word_1_4' title='bbox 413 592 449 616; x_wconf 95'><strong><em>be</em></strong></span>
<span class='ocrx_word' id='word_1_5' title='bbox 479 600 493 616; x_wconf 95'><strong><em>a</em></strong></span>
<span class='ocrx_word' id='word_1_6' title='bbox 523 592 668 622; x_wconf 95'><strong><em>perfect</em></strong></span>
<span class='ocrx_word' id='word_1_7' title='bbox 698 592 841 616; x_wconf 55'><strong><em>circle:</em></strong></span>
<span class='ocrx_word' id='word_1_2' title='bbox 150 592 230 616; x_wconf 96'>This</span>
<span class='ocrx_word' id='word_1_3' title='bbox 260 592 384 616; x_wconf 95'>should</span>
<span class='ocrx_word' id='word_1_4' title='bbox 413 592 449 616; x_wconf 95'>be</span>
<span class='ocrx_word' id='word_1_5' title='bbox 479 600 493 616; x_wconf 95'>a</span>
<span class='ocrx_word' id='word_1_6' title='bbox 523 592 668 622; x_wconf 95'>perfect</span>
<span class='ocrx_word' id='word_1_7' title='bbox 698 592 841 616; x_wconf 55'>circle:</span>
</span>
</p>
</div>
tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/stderr.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stdout.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/stdout.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/txt.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/pdf.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/stderr.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/stdout.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/txt.bin → tests/cache/aspect/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004__hocr__txt/hocr.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/hocr.bin
Vendored
+731 -731
View File
File diff suppressed because it is too large Load Diff
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_hocr__hocr__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/pdf.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/txt.bin → tests/cache/cardinal/__-l__eng__000001_ocr.png__000001_ocr_tess__pdf__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/hocr.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_hocr__hocr__txt/hocr.bin
Vendored
+731 -731
View File
File diff suppressed because it is too large Load Diff
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_hocr__hocr__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_hocr__hocr__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/txt.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_hocr__hocr__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004__hocr__txt/txt.bin → tests/cache/cardinal/__-l__eng__000002_ocr.png__000002_ocr_tess__pdf__txt/txt.bin
Vendored
+3 -3
View File
@@ -103,7 +103,7 @@ ERASE, REPEAT, PLAY/STOP, or LOCATE.
© Will sync to standard LinnDrum or Linn 9000 sync tone.
© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
® Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
(even drop frame!)
@@ -115,9 +115,9 @@ on the TAP TEMPO button.
¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
¢ Any TIME SIGNATURE may be used, and may be changed within a song.
linn
Linn Electronics, Inc.
nn
Linn Electronics, Inc.
18720 Oxnard Street, Tarzana, CA 91356
(818) 708-8131 TELEX #298949 LINN UR
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/hocr.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_hocr__hocr__txt/hocr.bin
Vendored
+731 -731
View File
File diff suppressed because it is too large Load Diff
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_hocr__hocr__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_hocr__hocr__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/txt.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_hocr__hocr__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003__hocr__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003__hocr__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000003_ocr.png__000003_ocr_tess__pdf__txt/txt.bin
Vendored
+128
View File
@@ -0,0 +1,128 @@
2A NNI‘I 6F6867# XATALL IE18-80L (818)
9SEI6 VO “BUBZIRY, “J0aNS PIPUXO OZLEI
“Uy ‘soTUOMOI,q UUrT
uut]
“‘SUOS B UIJIM pasueyo oq ABU pue ‘posn oq AWW AYN IVNOIS AWLL AUV
“parlsop Jr SUOTIISUBI} YIOOUIS YIM “BoueNbas eB OJUI pourtueIZOId 9q ABU SFONWHO OdINAL e
‘uonng OdNAL dV L 9) uO
sojou Jayienb Suiddy} Aq 10 ‘syUSTIOIOUI oINUTIAI-J8g-Jesg & JO sys} UL ofquisn(pe ‘ATTeouIAUINU paiajus oq ABU OdINALL e
(jouer doup u3a9)
“puooes Jed souely O€ 10 “SZ “pz 18 [LVAG-MAd-SHN VU 10 ALOANIWAAd-SLVAd U! patyoeds aq kewl OAL «
‘uoTe1odo [SVx JO} Aj[eusoyUT JoyndUIOd 11g 9] 98108 ZHI 8g ‘poeds-ysry Bann soz] e
"9U0} DUAS 0006 UUL] Jo wNIqUUr] prepue}s 0} OUAS [ITAA ©
“ONYBA 9}OU poloapes Aue Je sas—nd jndyno 07 pewureigold 3q ACW SL Ad LNO YADONAL OML
"ALVOOT 10 GOLS/AV 1d ‘LWddad “ASV
SUIpNpoUr ‘suOTIOUN] posn A[UOUILUOS 94] JO AUBUT [O1]UOD AJ9]OWIAI 0} PousIsse oq ACUI ST AdNI HOLIMSLOO OME «
“SUIPIONAI I[IYM P2sesd JOU Iv $3}OU BUTISIXO—ZUIPIOIA SATON.ASOP-UON
‘suoneurldxa peuoyippe sdeydsip uowng g1TqH
oy] ‘pepsau JI ‘suoneiodo [ye yYsnosy] NOA sapins ApIespo Avfdsip QO] Joey Z7¢ 9y3—uoeIodo Urea] 0} Ased ‘aus «
jUorel]suowtap & IO} Aepol Jayeap uur’] INOA dag ‘dISHUL
INOA 0} UONUS}]¥ PaplAIPUN INOA SUTJOASp ITY ps pue
p1osai ‘asoduod no Jay 0} pausisap st 1s0uenbesuur’] oy)
Aum Aposiooid $,Jeu], ‘SS9d0Id SATTBS1D OY} YIM SOIOJIOIUT
yey) xo]dwWI0d Os dq JOA9U P[NoUsS osn NOA AZopOuYdE} oy
ISTUMOIAUIO?) NOAA UOHISOdWIO)
"NOSpr] B Oy ‘AONUTJUT yada 0} seq Maz Se] BY] Jas UdAd
uvd NOA ‘palisap JJ ‘souanbes Mou ¥B OVUT sjied ou] [Te Adoo
ATesrewO Ne WI) [IM ONOS ALVAAO JeyIe80} wey}
,deyd,, 0} UOTOUNJ ONOS ALVA ou] asn usy] ‘saouanbes
JENPIAIpUt UI (“949 ‘snJOYD ‘aS1OA) UOTIDIS JIseq Yes
Pl0da1 OF ST ABM JOuIOUY “(812g 666 01 dn) ysnory) ABM
dU} [fe YORI] YORs p10991 0} ST SUOS B 9789I9 0} ABM SUG,
SUOS & SUTVAID
*suoT}oes poJUBMUN
SAOUIOI 0} ABM SWS dU} SoyeIodo SUV ALATAaG
“OBPLIq dy} PUB SNIOY PUOdAS dT]] Ud9MIAQ SIDA ISI
ay) Jo Adoo B JJasuT WYSE NOAA ‘afdwexs 10.f ‘UO JUSIN]JIP
B IO aouaNbas sues OY} UI—JOY OUP 0} UOTIEIO] 9UO WOT]
$1Bq JAOUI OF NOA sMOTIe WOTIOUNS AdOO/IMASNI OULL
‘SUIPIONAI JIVIS Udy) “OQuINU eq porisop ay} puy
0} CNIMAY 10 ‘CYVM Od LSWA “AEVOOT esn Apduns
sainjeay [PUOHIPPY
‘gouanbas & UTYIIM s]UTOd a1y1dads 3¥ $9100 QnPIOAO OL "UOT}
-ouns dALLS ATONIS 24) Suisn pasueyo Jo ‘pasesa ‘pappe
aq osye ABUT S9]ON ‘U0 9q ]IIM 1 “yoeq podeyd uayM
—aouanbas oy] ul skeyd 71 a10J9q Isnf posers oq 0} d]0U ayy
ssaid pue ASvwug ploy Aydunis ‘jou Suomm & aseso OL
sunipa
jsesdueyo ureisoid pue ‘fepod ureysns
‘yonoplalje ‘AWOOTOA ‘UOTyeTNpow ‘pusg youd Surpnyour
pep10del are $199JJ2 TCTIN [WV iPeqqnpseao aq Aeur syoen
Ze 07 dn ‘Kem sie Uy *(foeI} JOyOUR OJOS 10 ALLAN
NOA ssofum) duAS yOaysod ul Avy [[IM Yow] ISI 93 “prooar
NOA 3[IYM—SUIPIOIA LIBIS PU YORI) TUdIOTJIP B JOaTas
*y1ed MOU B QNPIsA0 OL, “SuIps0daJ-jods 10} aouanbes mno0k
UI UOHBIO] Aue ssad0e ATYOIND 0} owt} Aue ye pasn aq AvUE
SJONUOD FLIVOOT pur ‘ANIMA ‘CYVMaYOd LSVd
{SUIPIOSAI {IY posesa JOU se So]OU SuTsTXO—
yous} 3U} OUT poppe aq JIM poteyd sajou yeuonippe Auy
*(povesjap 10 poysn{pe oq ABW UOTIIII0D BUTUTT]) j{paqoeLI09
2q ][IM S1OLIe Sur [fe ATUO—patey]d nod Jey Jedy ]],NOA
‘] req 0] punose yoeq sdoo] sduanbas ay] Udy AA “YOu Yor
§,sa0uaNbas at} O] SUIT) UI preogday [IW] INO Avy usy3
AV'1d pue (YOON ssoid Ayduus ‘aousnbes & p1o09es OF,
g0uaNbas & SUIP10I0y]
‘JONWOD s}JouNaI TeuONdGO e
"UOTJEZIUOIYUAS OPOS UIT} FLAWS [euondo e
‘sou .sulddoys, noyyM sayelodo pue yoegdvyd ZuLINp S¥IOM NOLLOANNYOO ONIWILL e
‘onqea ory AY
pojoojes-oid & ye sajou pyoy Aue syeadas ATTeONewWO Ne UOTOUNS [WAdAY OAISNOX e
‘LSVJ SUnIpS soyeu UOTOUN ASV UA OUlN-[eal SAISNIOXY e
‘Koy B JO YONO} 941 12 CASOdSNVALL 0g ABU Syde] [Te 10 9UC e
i ASIP Jed
S9}0U OOO‘OTT JOA SpfOy puv SpUOdeS UT SBUOS Xa[AUIOD So10}S DALIP YSIP , 74 € ISCJ-CNIN
jSIOZISOUJUAS
stuoydAjod of 0} dn skeyd A[snoourynuls ‘spouueYd [IW 9T JO duo 0} pousisse oq
ABUL YORI] YOR ‘syous) oruoydAjod ‘snoouelnurs 7¢ SuTeJUOS ssouUaNbas QO] OY} JO YORA e
‘SJONUOS ATWOOT pur ‘GNIMAY ‘GaVM OA
LSVd ‘GYOOde AOLS ‘AV Td YIM Jopsocas ade} Yowsj-N[NU O} eps st UOTLISdO @
LOPNOUT SaINjeoy s[quyIeUlss AUB S.JJ ‘OSN pue UIes] 0} o[duns A[suIzeUe JOA ‘PnJsomod APOUIOITXO
St 1] “UeIOIsNUL feUOIssajoid oY} 10 JOO} soUBULIOJIJAd pue UOTIsOduIOS 11e-dY1-JO-9}e)s B SI IONUANbDaguUT] ay
JOps1odady soUINbIS [GTI YVAL ZE
Jgouanbaguury oy
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004.text__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003__hocr__txt/hocr.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_hocr__hocr__txt/hocr.bin
Vendored
+731 -731
View File
File diff suppressed because it is too large Load Diff
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004.text__pdf__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_hocr__hocr__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004.text__pdf__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_hocr__hocr__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/txt.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_hocr__hocr__txt/txt.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_tess__pdf__txt/pdf.bin
Vendored
BIN
View File
Binary file not shown.
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004__hocr__txt/stderr.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_tess__pdf__txt/stderr.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000004.ocr.png__000004__hocr__txt/stdout.bin → tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_tess__pdf__txt/stdout.bin
Vendored
View File
tests/cache/cardinal/__-l__eng__000004_ocr.png__000004_ocr_tess__pdf__txt/txt.bin
Vendored
+124
View File
@@ -0,0 +1,124 @@
2A NNI‘I 6F6867# XATALL IE18-80L (818)
9SEI6 VO “BUBZIRY, “J0aNS PIPUXO OZLEI
“Uy ‘soTUOMOI,q UUrT
uu
“‘SUOS B UIJIM pasueyo oq ABU pue ‘posn oq AWW AYN IVNOIS AWLL AUV
“parlsop Jr SUOTIISUBI} YIOOUIS YIM “BoueNbas eB OJUI pourtueIZOId 9q ABU SFONWHO OdINAL e
‘uonng OdNAL dV L 9) uO
sojou Jayienb Suiddy} Aq 10 ‘syUSTIOIOUI oINUTIAI-J8g-Jesg & JO sys} UL ofquisn(pe ‘ATTeouIAUINU paiajus oq ABU OdINALL e
(jouer doup u3a9)
“puooes Jed souely O€ 10 “SZ “pz 18 [LVAG-MAd-SHN VU 10 ALOANIWAAd-SLVAd U! patyoeds aq kewl OAL «
‘uoTe1odo [SVx JO} Aj[eusoyUT JoyndUIOd 11g 9] 98108 ZHI 8g ‘poeds-ysry Bann soz] e
"9U0} DUAS 0006 UUL] Jo wNIqUUr] prepue}s 0} OUAS [ITAA ©
“ONYBA 9}OU poloapes Aue Je sas—nd jndyno 07 pewureigold 3q ACW SL Ad LNO YADONAL OML
"ALVOOT 10 GOLS/AV 1d ‘LWddad “ASV
SUIpNpoUr ‘suOTIOUN] posn A[UOUILUOS 94] JO AUBUT [O1]UOD AJ9]OWIAI 0} PousIsse oq ACUI ST AdNI HOLIMSLOO OME «
“SUIPIONAI I[IYM P2sesd JOU Iv $3}OU BUTISIXO—ZUIPIOIA SATON.ASOP-UON
‘suoneurldxa peuoyippe sdeydsip uowng g1TqH
oy] ‘pepsau JI ‘suoneiodo [ye yYsnosy] NOA sapins ApIespo Avfdsip QO] Joey Z7¢ 9y3—uoeIodo Urea] 0} Ased ‘aus «
jUorel]suowtap & IO} Aepol Jayeap uur’] INOA dag ‘dISHUL
INOA 0} UONUS}]¥ PaplAIPUN INOA SUTJOASp ITY ps pue
p1osai ‘asoduod no Jay 0} pausisap st 1s0uenbesuur’] oy)
Aum Aposiooid $,Jeu], ‘SS9d0Id SATTBS1D OY} YIM SOIOJIOIUT
yey) xo]dwWI0d Os dq JOA9U P[NoUsS osn NOA AZopOuYdE} oy
ISTUMOIAUIO?) NOAA UOHISOdWIO)
"NOSpr] B Oy ‘AONUTJUT yada 0} seq Maz Se] BY] Jas UdAd
uvd NOA ‘palisap JJ ‘souanbes Mou ¥B OVUT sjied ou] [Te Adoo
ATesrewO Ne WI) [IM ONOS ALVAAO JeyIe80} wey}
,deyd,, 0} UOTOUNJ ONOS ALVA ou] asn usy] ‘saouanbes
JENPIAIpUt UI (“949 ‘snJOYD ‘aS1OA) UOTIDIS JIseq Yes
Pl0da1 OF ST ABM JOuIOUY “(812g 666 01 dn) ysnory) ABM
dU} [fe YORI] YORs p10991 0} ST SUOS B 9789I9 0} ABM SUG,
SUOS & SUTVAID
*suoT}oes poJUBMUN
SAOUIOI 0} ABM SWS dU} SoyeIodo SUV ALATAaG
“OBPLIq dy} PUB SNIOY PUOdAS dT]] Ud9MIAQ SIDA ISI
ay) Jo Adoo B JJasuT WYSE NOAA ‘afdwexs 10.f ‘UO JUSIN]JIP
B IO aouaNbas sues OY} UI—JOY OUP 0} UOTIEIO] 9UO WOT]
$1Bq JAOUI OF NOA sMOTIe WOTIOUNS AdOO/IMASNI OULL
‘SUIPIONAI JIVIS Udy) “OQuINU eq porisop ay} puy
0} CNIMAY 10 ‘CYVM Od LSWA “AEVOOT esn Apduns
sainjeay [PUOHIPPY
‘gouanbas & UTYIIM s]UTOd a1y1dads 3¥ $9100 QnPIOAO OL "UOT}
-ouns dALLS ATONIS 24) Suisn pasueyo Jo ‘pasesa ‘pappe
aq osye ABUT S9]ON ‘U0 9q ]IIM 1 “yoeq podeyd uayM
—aouanbas oy] ul skeyd 71 a10J9q Isnf posers oq 0} d]0U ayy
ssaid pue ASvwug ploy Aydunis ‘jou Suomm & aseso OL
sunipa
jsesdueyo ureisoid pue ‘fepod ureysns
‘yonoplalje ‘AWOOTOA ‘UOTyeTNpow ‘pusg youd Surpnyour
pep10del are $199JJ2 TCTIN [WV iPeqqnpseao aq Aeur syoen
Ze 07 dn ‘Kem sie Uy *(foeI} JOyOUR OJOS 10 ALLAN
NOA ssofum) duAS yOaysod ul Avy [[IM Yow] ISI 93 “prooar
NOA 3[IYM—SUIPIOIA LIBIS PU YORI) TUdIOTJIP B JOaTas
*y1ed MOU B QNPIsA0 OL, “SuIps0daJ-jods 10} aouanbes mno0k
UI UOHBIO] Aue ssad0e ATYOIND 0} owt} Aue ye pasn aq AvUE
SJONUOD FLIVOOT pur ‘ANIMA ‘CYVMaYOd LSVd
{SUIPIOSAI {IY posesa JOU se So]OU SuTsTXO—
yous} 3U} OUT poppe aq JIM poteyd sajou yeuonippe Auy
*(povesjap 10 poysn{pe oq ABW UOTIIII0D BUTUTT]) j{paqoeLI09
2q ][IM S1OLIe Sur [fe ATUO—patey]d nod Jey Jedy ]],NOA
‘] req 0] punose yoeq sdoo] sduanbas ay] Udy AA “YOu Yor
§,sa0uaNbas at} O] SUIT) UI preogday [IW] INO Avy usy3
AV'1d pue (YOON ssoid Ayduus ‘aousnbes & p1o09es OF,
g0uaNbas & SUIP10I0y]
‘JONWOD s}JouNaI TeuONdGO e
"UOTJEZIUOIYUAS OPOS UIT} FLAWS [euondo e
‘sou .sulddoys, noyyM sayelodo pue yoegdvyd ZuLINp S¥IOM NOLLOANNYOO ONIWILL e
‘onqea ory AY
pojoojes-oid & ye sajou pyoy Aue syeadas ATTeONewWO Ne UOTOUNS [WAdAY OAISNOX e
‘LSVJ SUnIpS soyeu UOTOUN ASV UA OUlN-[eal SAISNIOXY e
‘Koy B JO YONO} 941 12 CASOdSNVALL 0g ABU Syde] [Te 10 9UC e
i ASIP Jed
S9}0U OOO‘OTT JOA SpfOy puv SpUOdeS UT SBUOS Xa[AUIOD So10}S DALIP YSIP , 74 € ISCJ-CNIN
jSIOZISOUJUAS
stuoydAjod of 0} dn skeyd A[snoourynuls ‘spouueYd [IW 9T JO duo 0} pousisse oq
ABUL YORI] YOR ‘syous) oruoydAjod ‘snoouelnurs 7¢ SuTeJUOS ssouUaNbas QO] OY} JO YORA e
‘SJONUOS ATWOOT pur ‘GNIMAY ‘GaVM OA
LSVd ‘GYOOde AOLS ‘AV Td YIM Jopsocas ade} Yowsj-N[NU O} eps st UOTLISdO @
LOPNOUT SaINjeoy s[quyIeUlss AUB S.JJ ‘OSN pue UIes] 0} o[duns A[suIzeUe JOA ‘PnJsomod APOUIOITXO
St 1] “UeIOIsNUL feUOIssajoid oY} 10 JOO} soUBULIOJIJAd pue UOTIsOduIOS 11e-dY1-JO-9}e)s B SI IONUANbDaguUT] ay
JOps1odady soUINbIS [GTI YVAL ZE
Jgouanbaguury oy
tests/cache/cardinal/__-l__osd__--psm__0__000001.ocr.preview.jpg__stdout/stderr.bin → tests/cache/cardinal/__-l__osd__--psm__0__000001_rasterize_preview.jpg__stdout/stderr.bin
Vendored
View File

Some files were not shown because too many files have changed in this diff Show More