koala73 / koala73/worldmonitor

feat(correlation): implement conditional fuzzy entity resolution

Open
#5,989 0 comments 0 reactions 0 assignees View on GitHub
area: AI/intel area:seeders codex feature P2
Dominant language
TypeScript
Stars
86.6k
Forks
13.1k
Avg merge
8h 4m
Merged PRs (30d)
825

Description

## Parent

#5981

## What to build

If the Phase 1b gate passes, resolve fuzzy or transliterated open-vocabulary names into existing registry entities without allowing raw names or false transitive groups to become card identity.

## Acceptance criteria

- [ ] Consume the approved cross-runtime unknown-name producer and honor its fallback contract.
- [ ] Score incoming names against each registry name and alias with the documented two-factor model and a held-out-tuned threshold.
- [ ] Preserve typed and length prefilters without dropping the single-token misspelling cases this feature targets.
- [ ] Require every card-bearing group to contain a registry entity and prevent unrelated open-vocabulary names from chaining together.
- [ ] Produce deterministic groups with reproducible score breakdowns and non-empty merge reasons.
- [ ] Keep `exact` mode behavior unchanged and prove browser/seeder parity in `fuzzy` mode.

## Blocked by

- #5984
- #5986
- #5988

Contributor guide

Open the contributing guide

Research direction

Begin with parent #5981 and blocked issues #5984, #5986, and #5988; confirm the approved unknown-name producer and fallback contract before implementation. Trace the browser and seeder entry points, then verify the acceptance criteria: deterministic fuzzy groups, reproducible score breakdowns, unchanged exact mode, and browser/seeder parity.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
data
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.