pypandoc

Pypandoc provides a thin wrapper for pandoc, a universal document converter.

Installation

Pypandoc uses pandoc, so it needs an available installation of pandoc. For some common cases (wheels, conda packages), pypandoc already includes pandoc (and pandoc-citeproc) in its prebuilt package.

If pandoc is already installed (i.e. pandoc is in the PATH), pypandoc uses the version with the higher version number, and if both are the same, the already installed version. See Specifying the location of pandoc binaries for more.

To use pandoc filters, you must have the relevant filters installed on your machine.

Installing via pip

Install via pip install pypandoc.

Prebuilt wheels for Windows and Mac OS X include pandoc. If there is no prebuilt binary available, you have to install pandoc yourself.

If you use Linux and have your own wheelhouse, you can build a wheel which include pandoc with python setup.py download_pandoc; python setup.py bdist_wheel. Be aware that this works only on 64bit intel systems, as we only download it from the official releases.

Installing via conda

Pypandoc is included in conda-forge. The conda packages will also install the pandoc package, so pandoc is available in the installation.

Install via conda install -c conda-forge pypandoc.

You can also add the channel to your conda config via conda config --add channels conda-forge. This makes it possible to use conda install pypandoc directly and also lets you update via conda update pypandoc.

Installing pandoc

If you don't get pandoc installed via a prebuild wheel which includes pandoc or via the conda package dependencies, you need to install pandoc by yourself.

Installing pandoc via pypandoc

Installing via pypandoc is possible on Windows, Mac OS X or Linux (Intel-based, 64-bit):

# expects an installed pypandoc: pip install pypandoc
from pypandoc.pandoc_download import download_pandoc
# see the documentation how to customize the installation path
# but be aware that you then need to include it in the `PATH`
download_pandoc()

The default install location is included in the search path for pandoc, so you don't need to add it to the PATH.

By default, the latest pandoc version is installed. If you want to specify your own version, say 1.19.1, use download_pandoc(version='1.19.1') instead.

Installing pandoc manually

Installing manually via the system mechanism is also possible. Such installation mechanism make pandoc available on many more platforms:

Ubuntu/Debian: sudo apt-get install pandoc
Fedora/Red Hat: sudo yum install pandoc
Arch: sudo pacman -S pandoc
Mac OS X with Homebrew: brew install pandoc pandoc-citeproc Caskroom/cask/mactex
Machine with Haskell: cabal-install pandoc
Windows: There is an installer available here
FreeBSD with pkg: pkg install hs-pandoc
Or see Pandoc - Installing pandoc

Be aware that not all install mechanisms put pandoc in the PATH, so you either have to change the PATH yourself or set the full PATH to pandoc in PYPANDOC_PANDOC. See the next section for more information.

Specifying the location of pandoc binaries

You can point to a specific pandoc version by setting the environment variable PYPANDOC_PANDOC to the full PATH to the pandoc binary (PYPANDOC_PANDOC=/home/x/whatever/pandoc or PYPANDOC_PANDOC=c:\pandoc\pandoc.exe). If this environment variable is set, this is the only place where pandoc is searched for.

In certain cases, e.g. pandoc is installed but a web server with its own user cannot find the binaries, it is useful to specify the location at runtime:

import os
os.environ.setdefault('PYPANDOC_PANDOC', '/home/x/whatever/pandoc')

Usage

There are two basic ways to use pypandoc: with input files or with input strings.

import pypandoc

# With an input file: it will infer the input format from the filename
output = pypandoc.convert_file('somefile.md', 'rst')

# ...but you can overwrite the format via the `format` argument:
output = pypandoc.convert_file('somefile.txt', 'rst', format='md')

# alternatively you could just pass some string. In this case you need to
# define the input format:
output = pypandoc.convert_text('# some title', 'rst', format='md')
# output == 'some title\r\n==========\r\n\r\n'

convert_text expects this string to be unicode or utf-8 encoded bytes. convert_* will always return a unicode string.

It's also possible to directly let pandoc write the output to a file. This is the only way to convert to some output formats (e.g. odt, docx, epub, epub3, pdf). In that case convert_*() will return an empty string.

import pypandoc

output = pypandoc.convert_file('somefile.md', 'docx', outputfile="somefile.docx")
assert output == ""

It's also possible to specify multiple input files to pandoc, either as absolute paths, relative paths or file patterns.

import pypandoc

# convert all markdown files in a chapters/ subdirectory.
pypandoc.convert_file('chapters/*.md', 'docx', outputfile="somefile.docx")

# convert all markdown files in the book1 and book2 directories.
pypandoc.convert_file(['book1/*.md', 'book2/*.md'], 'docx', outputfile="somefile.docx")

# convert the front from another drive, and all markdown files in the chapter directory.
pypandoc.convert_file(['D:/book_front.md', 'book2/*.md'], 'docx', outputfile="somefile.docx")

In addition to format, it is possible to pass extra_args. That makes it possible to access various pandoc options easily.

output = pypandoc.convert_text(
    '<h1>Primary Heading</h1>',
    'md', format='html',
    extra_args=['--atx-headers'])
# output == '# Primary Heading\r\n'
output = pypandoc.convert_text(
    '# Primary Heading',
    'html', format='md',
    extra_args=['--base-header-level=2'])
# output == '<h2 id="primary-heading">Primary Heading</h2>\r\n'

pypandoc now supports easy addition of pandoc filters.

filters = ['pandoc-citeproc']
pdoc_args = ['--mathjax',
             '--smart']
output = pypandoc.convert_file(filename,
                               to='html5',
                               format='md',
                               extra_args=pdoc_args,
                               filters=filters)

Please pass any filters in as a list and not as a string.

Please refer to pandoc -h and the official documentation for further details.

Dealing with Formatting Arguments

Pandoc supports custom formatting though -V parameter. In order to use it through pypandoc, use code such as this:

output = pypandoc.convert_file('demo.md', 'pdf', outputfile='demo.pdf',
  extra_args=['-V', 'geometry:margin=1.5cm'])

Note: it's important to separate -V and its argument within a list like that or else it won't work. This gotcha has to do with the way subprocess.Popen works.

Logging Messages

Pypandoc logs messages using the Python logging library. By default, it will send messages to the console, including any messages generated by Pandoc. If desired, this behaviour can be changed by adding handlers to the pypandoc logger before calling any functions. For example, to mute all logging add a null handler:

import logging
logging.getLogger('pypandoc').addHandler(logging.NullHandler())

Getting Pandoc Version

As it can be useful sometimes to check what pandoc version is available at your system or which particular pandoc binary is used by pypandoc. For that, pypandoc provides the following utility functions. Example:

print(pypandoc.get_pandoc_version())
print(pypandoc.get_pandoc_path())
print(pypandoc.get_pandoc_formats())

pydocverter is a client for a service called Docverter, which offers pandoc as a service (plus some extra goodies).
See pyandoc for an alternative implementation of a pandoc wrapper from Kenneth Reitz. This one hasn't been active in a while though.
See panflute which provides convert_text similar to pypandoc's. Its focus is on writing and running pandoc filters though.

Contributing

Contributions are welcome. When opening a PR, please keep the following guidelines in mind:

Before implementing, please open an issue for discussion.
Make sure you have tests for the new logic.
Make sure your code passes flake8 pypandoc/*.py tests.py
Add yourself to contributors at README.md unless you are already there. In that case tweak your contributions.

Note that for citeproc tests to pass you'll need to have pandoc-citeproc installed. If you installed a prebuilt wheel or conda package, it is already included.

Contributors

Nicklas Tegner - New maintainer as of 1. Juli 2021
Valentin Haenel - String conversion fix
Daniel Sanchez - Automatic parsing of input/output formats
Thomas G. - Python 3 support
Ben Jao Ming - Fail gracefully if pandoc is missing
Ross Crawford-d'Heureuse - Encode input in UTF-8 and add Django example
Michael Chow - Decode output in UTF-8
Janusz Skonieczny - Support Windows newlines and allow encoding to be specified.
gabeos - Fix help parsing
Marc Abramowitz - Make setup.py fail hard if pandoc is missing, Travis, Dockerfile, PyPI badge, Tox, PEP-8, improved documentation
Daniel L. - Add extra_args example to README
Amy Guy - Exception handling for unicode errors
Florian Eßer - Allow Markdown extensions in output format
Philipp Wendler - Allow Markdown extensions in input format
Jan Schulz - Handling output to a file, Travis to work on newer version of pandoc, return code checking, get_pandoc_version. Helped to fix the Travis build, new convert_* API
Aaron Gonzales - Added better filter handling
David Lukes - Enabled input from non-plain-text files and made sure tests clean up template files correctly if they fail
valholl - Set up licensing information correctly and include examples to distribution version
Cyrille Rossant - Fixed bug by trimming out stars in the list of pandoc formats. Helped to fix the Travis build.
Paul Osborne - Don't require pandoc to install pypandoc.
Felix Yan - Added installation instructions for Arch Linux.
Kolen Cheung - Implement _get_pandoc_urls for installing arbitrary version as well as the latest version of pandoc. Minor: README, Travis, setup.py.
Rebecca Heineman - Added scanning code for finding pandoc in Windows
Andrew Barraford - Download destination.

License

Pypandoc is available under MIT license. See LICENSE for more details. Pandoc itself is available under the GPL2 license.

scottclowe/pypandoc