Skip to content

Rigour 2.3.1

Choose a tag to compare

View all tags
@pudo pudo released this 30 Jul 13:53
· 76 commits to main since this release
v2.3.1
f13a181

A data-only patch release: it widens the person-name alias corpus that backs the NAME symbol tagger, mostly for Traditional Chinese renderings and Belarusian forms of Romance and Germanic name particles.

Highlights

Person-name aliases (rust/data/names/person_names.txt)

The corpus feeds the AC name tagger, so an added alias means that surface form now resolves to the same NAME symbol as its other spellings during analyze_names — and therefore matches across scripts and transliterations rather than being compared character by character.

  • Traditional Chinese renderings for entries that previously carried only the Simplified form: 馮 (von), 範 (van), 達 (da), 戴爾 (del), 赫爾 (hull), 庫馬爾 (kumar), 格奧爾基耶夫 (georgiev), 尼諾·德 (niño de). Also 德拉 for de la, covering the form written without an interpunct.
  • Belarusian forms of name particles, which were absent across the board: дэ and ды (de), дэль (del), дэльі (degli), ля (la), дзю (du).
  • New entry for the surname georgiev (Q105531174), alongside the already-present georgiew and georgiewa.
  • Merged the duplicate van entry (Q50354059) into the canonical one (Q1258618), which picked up its 來自 alias. "van" and its variants now emit a single NAME symbol instead of two.

Upgrade notes

No API or behaviour change beyond the widened alias set. Names containing the added forms will now produce NAME symbol spans where they previously produced none, which can raise name-matching scores for affected pairs.

Full changelog: v2.3.0...v2.3.1

Assets 2
Loading