Skip to main content

A runtime environment for querying FSTs stored in the AT&T `.att` format.

Project description

fst-runtime

The fst_runtime python package is written as a light-weight front-end for finite-state transducers compiled to AT&T's .att format via programs like foma, hfst, etc. This runtime reads the FST into memory from the .att file, and publishes two main query functions, down_generation and up_analysis, that walk the FST either down (wal+VERB+GER -> walking) or up (walking -> wal+VERB+GER), respectively.

There are also methods for bulk querying, down_generations and up_analyses.

Documentation for this project can be found at https://culturefoundryca.github.io/fst-runtime/.

Installation Instructions

This package is published on PyPI and can be installed via pip install fst_runtime or poetry add fst_runtime, etc.

Weighted FSTs

This runtime supports weighted FSTs, where the weights are defined under a semiring. Common semirings are provided via fst_runtime.semiring.

Example Usage

from fst_runtime.fst import Fst

fst = Fst('/home/username/fsts/walk.att')

generations = fst.down_generation('wal', suffixes=[['+VERB'], ['+GER', '+INF']])

for generation in generations:
    print(generation)

This example, based off of tests/data/fst4.att, would then return the results of wal+VERB+GER and wal+VERB+INF which would be ['walking', 'walk']. If you simply called fst.down_generation('wal'), it would generate all possible wordforms of it in the FST.

Similarly with up:

analyses = fst.up_analysis('walking')

for analysis in analyses:
    print(analysis)

In this case, we only get one result back, wal+VERB+GER.

Acknowledgements

We would like to thank Dr. Miikka Silfverberg for his help in deciding what this application should look like, and for providing test FSTs for us to use to test the application.

We would also like to thank UBC's ELF-Lab for the use of the waabam walk through their fully-fledged Ojibwe FST. You can find their repo here: https://github.com/ELF-Lab/OjibweMorph.

Also thank you to Sandra Radic for some early application code.

Development

For working on this project, please consult the wiki for the project's architecture and dev setup. PRs welcome.

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

fst_runtime-0.2.1.tar.gz (17.4 kB view details)

Uploaded Source

Built Distribution

fst_runtime-0.2.1-py3-none-any.whl (18.2 kB view details)

Uploaded Python 3

File details

Details for the file fst_runtime-0.2.1.tar.gz.

File metadata

  • Download URL: fst_runtime-0.2.1.tar.gz
  • Upload date:
  • Size: 17.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.3 CPython/3.12.6 Linux/6.8.0-1014-azure

File hashes

Hashes for fst_runtime-0.2.1.tar.gz
Algorithm Hash digest
SHA256 adeed4e11d06d6b925d9135268b424cfef284d8ab99722de8648217f96e620e4
MD5 471ffdedae4082a4fe1f20e300a2801c
BLAKE2b-256 0f58366fff1598eca0ca00d25577dfa416e6f5951f090b28d16a893c0ecb1d47

See more details on using hashes here.

File details

Details for the file fst_runtime-0.2.1-py3-none-any.whl.

File metadata

  • Download URL: fst_runtime-0.2.1-py3-none-any.whl
  • Upload date:
  • Size: 18.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.3 CPython/3.12.6 Linux/6.8.0-1014-azure

File hashes

Hashes for fst_runtime-0.2.1-py3-none-any.whl
Algorithm Hash digest
SHA256 369faabbb40ed013e2d0ec64030873c64b368170afe04d10ff7809ccb353267f
MD5 a4993ea71c9050dbcf4bf037e6c2d3da
BLAKE2b-256 896edcebc42f063cb4505a79877e610108cb2eb7fbf7cc1dbb09aff874e8aa6d

See more details on using hashes here.

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