citeproc-py

citeproc-py is a CSL processor for Python. It aims to implement the CSL 1.0.1 specification. citeproc-py can output styled citations and bibliographies in a number of different output formats. Currently supported are plain text, reStructuredText and HTML. Other formats can be added easily.

citeproc-py uses semantic versioning. Currently, its major version number is still at 0, meaning the API is not yet stable. However, you should not expect to see any major API changes soon.

Requirements

citeproc-py supports Python 3.9 and up. It depends on lxml for parsing and navigating the CSL style and locale files.

Installation

The recommended way of installing citeproc-py is by using pip:

pip install citeproc-py

If lxml isn't installed, pip will try to install it for you.

If you insist, you can manually install citeproc-py from distribution packages hosted at PyPI. Please ignore the release archives offered by GitHub.

Getting Started

To get started with citeproc-py, take a look at the examples under examples/. Two examples are provided, one parsing references from a JSON representation of references as supported by citeproc-js, another parsing the references from a BibTeX file. Both show and explain how to cite references and render the bibliography.

CSL Compatibility

Currently, citeproc-py passes almost 60% of the (relevant) tests in the citeproc-test suite. However, it is more than 60% complete, as citeproc-py doesn't take care of double spaces and repeated punctuation marks yet, making a good deal of the tests fail. In addition, the following features have not yet been implemented (there are probably some I forgot though):

• disambiguation/year-suffix
• et-al-subsequent-min/et-al-subsequent-use-first
• collapsing
• punctuation-in-quote
• display

Also, some citeproc-js functionality that is not part of the CSL spec is not (yet) supported:

• raw dates
• static-ordering
• literal names

Contributing

citeproc-py is 100% volunteer maintained, and new contributions are always welcome. If you would like to contribute, please follow the guidelines in the CONTRIBUTING.md file.

Local Install and Running the Tests

First clone the citeproc-py repository and install the submodules with

git submodule update --init

Then install with

pip install -e .

Then move to the tests directory and run

python citeproc-test.py

Run

citeproc-test.py --help

to see its usage information. The first time you run the script it will clone the citeproc-test suite repository into the tests directory and checkout the last tested version. By default failed tests are automatically added into the failing_tests.txt file and aren't shown when running the test suite again.

If you want git to fully ignore the submodule, you can type

git update-index --assume-unchanged citeproc/data/schema

v0.8.2 (Wed Mar 12 2025)

🏠 Internal

• Explicitly specify utf-8 encoding while reading top level .md files for description #162 (@yarikoptic)
• Instruct that long description is in markdown and not ReST #160 (@yarikoptic)

Authors: 1

• Yaroslav Halchenko (@yarikoptic)

v0.8.1 (Wed Mar 12 2025)

🏠 Internal

• release: checkout with submodules #159 (@yarikoptic)

Authors: 1

• Yaroslav Halchenko (@yarikoptic)

v0.8.0 (Wed Mar 12 2025)

🚀 Enhancement

• release: use GitHUb App token for checkout and push #158 (@tmorrell)

⚠️ Pushed to master

• Rename CHANGES to CHANGELOG (@yarikoptic)

🏠 Internal

• release: switch to using GitHub App to overcome branch protection #157 (@tmorrell)
• Set up auto to automate releases #153 (@jwodder @yarikoptic)

📝 Documentation

• Convert CHANGES.rst, CONTRIBUTING.rst, and README.md to markdown #154 (@tmorrell)

Authors: 3

• John T. Wodder II (@jwodder)
• Tom Morrell (@tmorrell)
• Yaroslav Halchenko (@yarikoptic)

Release 0.7.0 (2025-02-19)

Just to get a release out after long period.

Bug fixes

• Better handling of ordinals
• Fix locator conditions (resolves #142)
• Make family name optional
• Allow date parts to not be integers
• Support space macros
• Fix multiple capitals
• Fix parsing BibTex integer values

Other changes:

• Removed support for Python 3.6, 3.7, 3.8 and added for 3.10 - 3.13
• Switch versioning to versioneer

Release 0.6.0 (2021-05-27)

Bug fixes

• Various issues on Windows: testing, explicit utf-8 encoding
• Strings with unescaped \ declared r"aw"

Other changes:

• Removed support for Python 2.7, 3.5 and added for 3.9
• Travis CI is removed in favor of GitHub actions testing across all 3 major OSes (MacOS, Windows, GNU/Linux Ubuntu)
• CSL test-suite progressed from 5779a8c to c3db429

Release 0.5.1 (2020-03-06)

Bug fixed:

• avoid rnc2rng 2.6.2 which breaks installation of citeproc-py

Release 0.5.0 (2020-02-09)

Enhancements:

• handle commas and ampersands in a functional style
• Number: handle commas and ampersands
• added symbol for textquotesingle
• specify fallback locales for fr-CA and es-CL
• improved page number and ranges parsing

Bug fixed:

• don't fail on empty page ranges (#90) (bbm)
• detect end of file while parsing incorrect bib (#59) (John Vandenberg)

Other changes:

• Removed 3.2-3.4 and added 3.7, 3.8 to supported Pythons
• Refactored locales handling

Release 0.4.0 (2017-06-23)

New features:

• allow specifying the encoding of a BibTeX database file (#20 and #25)
• BibTeX 'month' field: support integers and " " values
• BibTeX 'pages' field: support "10", "10+", "10-12" and "10--12" formats
• BibTeX entry types: map the non-standard 'thesis' and 'report' entries
• update the CSL schema to version 1.0.1 (#5)
• update the CSL locales to commit 49bf3fc0

Bug fixed:

• avoid crash when there is nothing to affix (David Lesieur)
• fix BibTeX month to CSL month mapping (#24)
• strip leading/trailing whitespace from BibTeX values (#37)

Release 0.3.0 (2014-11-07)

Major improvements to the BibTeX parser:

• split names into parts and assign them to the equivalent CSL name parts
• fixed handling of accent macros and escaped characters
• more compatible (La)TeX macro handling in general (but still basic)
• handle standard Computer Modern ligatures such as --, ---, and <<
• added unit tests for the BibTeX and LaTeX parsers

Other changes:

• disable RelaxNG validation of CSL styles by default (API change)

Release 0.2.0 (2014-10-25)

• bad cite callback function can determine how a bad cite is displayed (hetsch)
• added option to disable RelaxNG validation (Jasper Op de Coul)
• distutils was replaced with setuptools (Joshua Carp)
• bug fixes (Yaroslav Halchenko, David Lesieur)
• CitationStylesBibliography.bibliography() now returns the list of entries