Skip to content

fix: keep non Latin letters when normalizing person names - #67

Merged
VolodymyrLinuxovich merged 1 commit into
mainfrom
fix/non-latin-person-names
Oct 31, 2026
Merged

VolodymyrLinuxovich merged 1 commit into
mainfrom
fix/non-latin-person-names

Conversation

@VolodymyrLinuxovich

Copy link
Copy Markdown
Owner

Fixes #66

Summary

Name normalization now keeps letters from every script. It still lowercases and drops accents, so José and Jose share a key.

Provider dedupe uses this for the person name and the organization name. Two different Cyrillic named investors at one fund are both kept. A true repeat of the same person is still removed.

Provider organization lookup by name uses the same key. Two Cyrillic fund names no longer share an empty key.

Validation

  • New tests in people-search-pipeline.test.ts. They fail on main.
  • npm test passes.
  • npm run typecheck and lint pass.
  • npm run eval:retrieval gives the same numbers as main.

normalizedName and the provider dedupe key kept only a to z, so Cyrillic or CJK names became empty and different people at one fund were dropped as duplicates. Names now keep letters from every script and drop accents. Provider organization lookup uses the same key.

Fixes #66
@vercel

vercel Bot commented Oct 13, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
largevcmodel Ready Ready Preview Oct 13, 2026 3:30am UTC

@VolodymyrLinuxovich
VolodymyrLinuxovich merged commit 5eef95b into main Oct 31, 2026
6 checks passed

This branch was successfully deployed

1 active deployment
Preview — 7e5b2644 Deployed Oct 13, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Provider dedupe drops people whose names are not in Latin script

1 participant