Bullseye and Bookworm flavors

• Added support for Bullseye and Bookworm based docker images

New major release : v 2.0 !

Version 2.0 introduces two major changes:

1. A new method to traverse the composition graph, which dramatically improves the overall speed, especially when the sequences are long contain many errors. We have files that took 25 minutes to align before that can now take about 7 seconds. This is especially noticeable with the adapted composition (the default).
2. Some smarts were introduced when --use-case and --use-punctuation are enabled. Now, by default, punctuation symbols can only be substituted by other punctuation symbols (or deleted/inserted). Also, words that differ only by the first letter case will be preffered for substitution.

These behavior, as well as the beam size (that has a default value of 50.0) can be controlled with the following new parameters:

  --disable-strict-punctuation: Disable strict punctuation alignment (which prevents punctuation aligning with words).
  --disable-favored-subs        Disable favored substitutions (which makes alignment favor substitutions between words which differ only by case).
  --favored-sub-cost FLOAT      Cost for favored substitutions (e.g., case diff). Default: 0.1

See the README.md for more details.

1.14.0

What's Changed

• Allow flexible forwarding of NLP columns to SBS by @nishchalb in #56

Full Changelog: 1.13.0...1.14.0

fstalign-1.13.0

What's Changed

• Update license copyright year(s) by @github-actions in #53
• Bug: WER sidecar info not appearing in SBS by @nishchalb in #55

Full Changelog: 1.12.0...1.13.0

1.12.0

What's Changed

• FSTALIGN-37: Add flag in fstalign to allow for case-sensitive testing by @pique0822 in #51
• FSTALIGN-37: Add flag in fstalign to allow for case-sensitive testing (cont.) by @pique0822 in #52

Full Changelog: 1.11.0...1.12.0

fstalign-1.11.0

Release notes - Rev FST Aligner - 1.11.0

Improvement

FSTALIGN-61 Minimize explicit heap allocations (new) and leverage more managed memory sematics

FSTALIGN-62 Migrate to the Debian Buster and newer OpenFST

FSTALIGN-63 Add a way to preserve inserts in NLP output.

fstalign-1.10.0

What's Changed

• Updated nlp file that was missing confidence column by @dchen579 in #43
• Accept references in FST format by @dchen579 in #44
• FSTALIGN-61: Fewer explicit heap allocations by @dchen579 in #45

New Contributors

• @dchen579 made their first contribution in #43

Full Changelog: 1.9.0...1.10.0

fstalign-1.9.0

What's Changed

• Update license copyright year(s) by @github-actions in #38
• Update FST usage to mention composition approach by @nishchalb in #41
• Pass through confidence to aligned NLP output by @nishchalb in #42
• Nerd-1422: Add flag for reading punctuation from nlp as separate tokens by @ajhinsvark in #10

New Contributors

• @github-actions made their first contribution in #38

Full Changelog: 1.8.0...1.9.0

fstalign-1.8.0

What's Changed

• Ignore extra columns in NLP by @nishchalb in #33
• Allow alright in synonym reference by @nishchalb in #34
• Add sentence WER metrics for NLP input by @nishchalb in #35

Full Changelog: 1.7.0...1.8.0

fstalign-1.7.0

What's Changed

• New QOL github action workflows by @qmac in #29
• FSTALIGN-52: Remove some synonyms by @nishchalb in #30
• Put wer tag entity type in SBS output by @nishchalb in #32

Full Changelog: 1.6.1...1.7.0