Skip to main content

Library for computing measures of similarity for sequences of hashable data types

Project description


PyPI CI Documentation Status

Python library for computing measures of distance and similarity for sequences of hashable data types.


While developed as a general-purpose library, seqsim is mostly designed for usage in research within the field of cultural evolution, and particularly of the cultural evolution of textual traditions. Some methods act as a thin-wrapper to either the standard Python library or of to other libraries such as textdistance.


In any standard Python environment, seqsim can be installed with:

$ pip install seqsim


The library offers different methods to compare sequences of arbitrary hashable elements. It is possible to mix sequence and element types.

Full documentation is offered at ReadTheDocs and code with almost complete coverage is offered in the tests. For most common usages, a wrapper .distance() function can be used.

>>> import seqsim
>>> seqsim.edit.levenshtein_dist("kitten", "string")
>>> seqsim.edit.levenshtein_dist("kitten", "string", normal=True)
>>> 0.8333333333333334
>>> seqsim.sequence.ratcliff_obershelp([1,2,3,4], [2,4,3,5])
>>> seqsim.compression.entropy_ncd([1,2,3,4], [2,4,3,5])


The core of the library are the metrics for sequence distance/similarity on arbitrary data types, as in the table below.

Method "kitten" / "sitting" (1, 2, 3, 4) / (3, 4, 2, 1)
arith_ncd 1.25 0.888889
arith_ncd_normal 1.25 0.888889
birnbaum 0.666667 0.7
birnbaum_normal 0.666667 0.7
birnbaum_simil 7 3
birnbaum_simil_normal 0.25 0.3
bulk_delete 3 3
bulk_delete_normal 0.428571 0.75
damerau 3 4
damerau_normal 0.428571 1
entropy 0.101341 0
entropy_normal 0.101341 0
fragile_ends_simil 3 3.5
fragile_ends_simil_normal 0.5 1
jaccard 0.7 0
jaccard_normal 0.7 0
jaro 0.253968 0.5
jaro_normal 0.253968 0.5
jaro_winkler 0.253968 0.5
jaro_winkler_normal 0.253968 0.5
levenshtein 3 4
levenshtein_normal 0.428571 1
mmcwpa 0.538462 0.387628
mmcwpa_normal 0.538462 0.387628
ratcliff_obershelp 0.384615 0.5
ratcliff_obershelp_normal 0.384615 0.5
sorensen 0.384615 0
sorensen_normal 0.384615 0
stemmatological_simil 3 3
stemmatological_simil_normal 0.428571 0.75
subseq_jaccard 0.751556 0.547008
subseq_jaccard_normal 0.751556 0.547008


Version 0.3.1:

  • Fixed bug due to typo in one of the methods
  • Selected one Birnbaum implementation

Version 0.3:

  • Improvements to code quality, documentation, and references
  • Added new methods and scaffolding for future expansions

Version 0.2:

  • First release for new roadmap supporting sequences of any hashable Python datatype, importing code from other projects (mostly from titivillus)

Community guidelines

While the authors can be contacted directly for support, it is recommended that third parties use GitHub standard features, such as issues and pull requests, to contribute, report problems, or seek support.

Contributing guidelines, including a code of conduct, can be found in the file.

Authors and citation

The library is developed in the context of "Cultural Evolution of Text", project, with funding from the Riksbankens Jubileumsfond (grant agreement ID: MXM19-1087:1).

If you use seqsim, please cite it as:

Tresoldi, Tiago; Maurits, Luke; Dunn, Michael. (2021). seqsim, a library for computing measures of distance and similarity for sequences of hashable data types. Version 0.3.2. Uppsala: Uppsala universitet. Available at:

In BibTeX:

  author = {Tresoldi, Tiago; Maurits, Luke; Dunn, Michael},
  title = {seqsim, a library for computing measures of distance and similarity for sequences of hashable data types. Version 0.3.2},
  howpublished = {\url{}},
  address = {Uppsala},
  publisher = {Uppsala universitet},
  year = {2021},


The image at the top of this file is derived from Yves de Saint-Denis, Vie et martyre de saint Denis et de ses compagnons, versions latine et française. It is available in high resolution from Bibliothèque nationale de France, Département des Manuscrits, Français 2090, fol. 12v.

References to the various implementation are available in the source code comments and in the online documentation.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

seqsim-0.3.2.tar.gz (22.0 kB view hashes)

Uploaded source

Built Distribution

seqsim-0.3.2-py2.py3-none-any.whl (23.4 kB view hashes)

Uploaded py2 py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page