2.8 KiB
2.8 KiB
Contributions to spencermountain/compromise
modest natural-language processing
This repository holds only my commits to the upstream project, with their original author, date and diff. Each of my commits is preceded by a snapshot of the upstream files it touched, so every diff shows exactly what I changed.
Merged pull requests
- find payloads by term id so they survive a .remove() (2026-10-06)
- keep a titlecase comparative before 'than' as an adjective (2026-10-06)
- keep a closing bracket in the word when it closes one inside it (2026-10-05)
- match lexicon words that start with # or @ (2026-10-05)
- leave a spaced dash out of a match's text (2026-10-05)
- contract each match on its own (2026-10-05)
- don't add an implicit 'to' after 'wanna' before a noun (2026-10-05)
- split hyphens in a match the same way the tokenizer does (2026-10-05)
- contract 'he is not' to he's not instead of he'sn't (2026-10-05)
- drop does/did when turning a negative verb positive (2026-10-05)
- contract 'can not' to can't (2026-10-05)
- dates: read singular possessive durations (2026-10-05)
- don't merge an -ent syllable back into an open first syllable (2026-10-05)
- run the regex-boundaries tests on the full build (2026-10-05)
- parse 'fortnight' as two weeks (2026-10-05)
- Treat 's before a copula as possessive (2026-09-29)
- Keep the case in toSingular (2026-09-29)
- Keep "never" when going to the infinitive (2026-09-29)
- Skip the verb of a relative clause following the subject (2026-09-19)
- Recognize more openers of a secondary clause in mainClause (2026-09-19)
- Only treat subordinators as secondary-clause signals before the verb (2026-09-14)