c-fu/arah2
1
0
Fork 1
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
95 Commits 2 Branches 0 Tags
6368bad3d5150814dc50d55bbfa074e015d94a6c
Commit Graph
3 Commits
Author SHA1 Message Date
araharah-executor 6368bad3d5 feat(04-01): OSA Damerau distance + disjoint-tier scorer (the residual net) 
- distance.ts: hand-rolled 3-band OSA, adjacent transposition = 1 edit,
  zero imports (D-07); the row rotation is pinned by the D1-D9 vectors
- score.ts: TIER bands strictly disjoint descending — EXACT 1.00 /
  ALIAS_EXACT 0.90 / PREFIX 0.75 / TOKEN_SET 0.65 / FUZZY as the 0.60
  CEILING (a fuzzy score is fuzzySim * 0.60, never a 0.00 multiplier);
  FUZZY_ADMIT_SIMILARITY = 0.72 exported for source.ts + the corpus
  runner; token-level residual lets 'Sentrl' reach 'KL Sentral' while
  'Jln Tun Razak' cannot reach 'Jalan Ampang'
- scoreCandidate returns the RAW matched alias (corpus provenance);
  an exact ALIAS hit rides ALIAS_EXACT so a name-exact outranks it
- score.test.mjs S7 corrected during GREEN: admission belongs to
  source.ts — scoreCandidate asserts the sub-floor similarity instead

D1-D9 + S1-S7 green; typecheck clean.
 2026-10-24 00:32:44 +08:00
araharah-executor 25ee1e9c61 test(04-01): RED — OSA distance vectors (D1-D9) + disjoint-tier scorer suite (S1-S7) 
- 9 hand-computed Damerau-Levenshtein vectors (transpositions, empties,
  kitten/sitting sanity, the corpus's klcc/kclc + sunway/sunwya typos)
- 7 scorer laws: tier literals + disjointness, min-length guard before
  exact, exact-outranks-fuzzy, mid-word PREFIX vs token-boundary
  TOKEN_SET, fuzzy residual 0.75*0.60 arithmetic, the 0.72 admission
  boundary (Penang/Padang 0.667 refuses), matchedAlias raw provenance

All 16 new cases fail while src/distance.ts + src/score.ts are absent.
 2026-10-24 00:25:19 +08:00
araharah-executor ebcd542643 test(04-01): RED — search workspace scaffold + normalizer vector suite (T1-T9) 
- add the search/ npm workspace (root workspaces + search:test/search:typecheck delegates)
- zero-dependency runner/harness copied from the bot/test precedent
- 9 named normalizer laws: rule vectors, symmetry, whole-token-only,
  slash preservation, glued split, vowel-drop, idempotence, folding,
  table-is-data (>= 47 rows, one generic builder)

All 9 cases fail while src/normalize.ts is absent — the RED gate.
 2026-10-24 00:21:59 +08:00