📐onto-hsmap

The groupant beside onto/src/hsmap.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.

Facts

Components

Obernants

The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.

Parts

• onto-hsmap-selection
• onto-hsmap-isa-gate

Model text

descriptant

onto-hsmap

Candidates search longest NP and spelling variants, shorter preference-fit can override first nonfit. pick=WSD+state+chapter with >1e-9 replacement threshold. Build pass1 chapter distribution/pass2 scoring and constrained confirmations; ordered IsA fallback gates; sequential TSV derivative writes. judge/confirm invoke teacher and remain unexecuted.

Modeling epoch 2026-10-31

Prospective model saved before edits; nine existing sources read fully. Added bounded inline tests; store reader overflow repaired only after compiled old-code failing witness. No live corpus, teacher, model or database effects executed. Source syntax generated twice identically; partial lexical projection only.

approximant

Forward relation

Candidates search longest NP and spelling variants, shorter preference-fit can override first nonfit. pick=WSD+state+chapter with >1e-9 replacement threshold. Build pass1 chapter distribution/pass2 scoring and constrained confirmations; ordered IsA fallback gates; sequential TSV derivative writes. judge/confirm invoke teacher and remain unexecuted.

Equations

s(c)=WSD(c)+W_HS·1[lf(c)∈prefer]+W_CHAP·chapter_share(lf(c)); b_{i+1}=i if s_i>s_b+1e-9 else b. Qmap(s)=first accepted numeric row; Confirm(k)=last row.

Domain

Picks stay in supplied senses; confirmed c can retain old margin/why metadata. Tables trusted and may panic malformed; writes nontransactional.

Categorical behavior and effect boundaries have no validated percentage surrogate. Executable Rust approximant unsupported; 1/2/5/10% future caps require a useful metric first.

invariant

Invariant and domain

Picks stay in supplied senses; confirmed c can retain old margin/why metadata. Tables trusted and may panic malformed; writes nontransactional.

Limits

Finite synthetic fixtures are not semantic proof. Public invalid indices, allocation failures, huge counters, environment races and live effects outside explicit checks remain unverified.

cost

Cost approximation

Scoring O(S), WSD delegated; two mapped passes plus quadratic neighbour pairs per heading and feature scans; retains all rows/descriptions/output. No resource bound.

Symbolic costs only; no timing or throughput claim.

composition

Composition

s(c)=WSD(c)+W_HS·1[lf(c)∈prefer]+W_CHAP·chapter_share(lf(c)); b_{i+1}=i if s_i>s_b+1e-9 else b. Qmap(s)=first accepted numeric row; Confirm(k)=last row.

Graph builder→store shape check→Kb; Kb→query→bridge/describe/pron. hs Parsed→hsmap preference/scoring/gates. Core SynMap last-duplicate vs Graph first-duplicate requires unique synset keys for interchangeable lookup. Syntax children below share parent projection; exact byte ranges bind original source semantics.

• selection: score(c)=WSD(c)+1[preferred]+chapter_share(lf(c)); best replaces only >old+1e-9

• isa-gate: Verdict=first witness in ancestry,other-sense,part,taxon,modifier,missing-head,made,source,LCA

models

behavior_complement

Unaccepted Q cannot load; later accepted duplicate cannot replace first. Confirmation outside candidates cannot override. Direct ancestry excludes weaker fallback verdict.

reverse_reconstruction

Pick preimage is score vectors selecting c under scan threshold with reported second/margin. TSV loses source text, dedup levels/bounds; loader histories not recoverable.

score_ties

First best retained unless >1e-9 improvement; second total_cmp then low index; singleton9.99 sentinel.

graph_verdict_order

Yes→OtherSense→Part→Taxon(first60)→Modifier→Unknown head→Made→Source→Near/No.

mapping_identity

Q first accepted numeric row; confirmations last(phrase,chapter/state). live/food keys separate.

effect_boundary

load_rows external CSV/model; build sequential output; teacher paths excluded; render excludes scores/verdicts.

tests

Actual-code checks

• hsmap::wave29_tests::alignment_first_and_confirmation_last
• hsmap::wave29_tests::threshold_preferences_and_graph_join

Evidence 2026-10-31

Each selected test passed twice debug and twice release after exact mutation restoration (15 named suite checks; all four sorted outcome sets equal, exits0). Compiled source control: negative-hsmap.log (compiled source mutation assertion failure101). Controls compile successfully and fail an assertion with exit101. Store control is the original unsafe length addition before repair.

Pure selection/graph gate/load temp helpers; live build/teacher/model/CSV unexecuted

Whole CLI, live teacher/corpus/model/prior loading, filesystem failure/durability and huge inputs were not executed. No exact wrapper is advertised as a percentage surrogate. Added test blocks rustfmt-formatted independently; unrelated source formatting preserved. Strict broad Clippy not repeated: existing dependency/source warnings are outside this work.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/onto/src/hsmap.rs.groupant/