Elamur Groupantsmath-steps-binding
The groupant beside math/src/steps.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
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | approximant |
| ast | partial | ../syntax.json |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | source review only runtime gap | tests |
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.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial source derived model | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
modifier_compatibility | modifier_compatibility | partial source derived model | modifier_compatibility |
environment_dispatch | environment_dispatch | partial source derived model | environment_dispatch |
Model text
descriptant
math-steps-binding
ignore_set v1 collects real quantities lacking question-context overlap, returning them only if>=2 real quantities remain. v2 uses target head/modifier matches over quantities with units; returns unmatched unit-bearing indices only if at least1 bound item, a nonempty ignored set and>=2 real survivors.
Epoch 2026-10-31: replaced duplicate catalog sentences with concrete exclusions, compatible-antecedent examples and two distinct scoped behavioral views. Source and prior execution evidence are unchanged; this revision ran no tests.
approximant
Scoped forward relation
ignore_set v1 collects real quantities lacking question-context overlap, returning them only if>=2 real quantities remain. v2 uses target head/modifier matches over quantities with units; returns unmatched unit-bearing indices only if at least1 bound item, a nonempty ignored set and>=2 real survivors.
This source-derived relation is partial; no executable percentage surrogate or replacement is supplied.
invariant
Scoped domain and invariants
Returned indices belong to real input quantities. Ignore selection is a heuristic feature/selection boundary, not proof that excluded numbers are irrelevant.
Cost
Membership checks repeatedly scan target/modifier/context lists; collecting bound/ignored indices uses Vec::contains. Training BindModel separately costs epochsitemsfeatures and is not executed here.
models
behavior_complement
A const-prefixed index cannot enter the returned ignore set. A result leaving fewer than2 real quantities is excluded. Missing-unit quantities in v2 cannot be ignored merely for lacking a unit.
reverse_reconstruction
With two bound apples plus one orange, [orange_index] can result for many subject/verb/number values. An empty set admits no targets, no unmatched units, or the survivor-count gate failing; it does not identify the reason.
modifier_compatibility
v2 treats empty target modifiers or empty description modifiers as compatible; otherwise any shared modifier suffices. It does not demand full modifier-set equality.
environment_dispatch
MATH_IGNORE_V=1 selects context-overlap v1; all other values use v2. Identical Prob can yield different ignore sets under these environments; constants are excluded in both.
tests
Scoped evidence
Static source review only; no ignore_set or BindModel execution.
Named actual source fixtures: none.
Parent tests.md records commands and paired profiles. No additional coverage inferred from parent fixtures outside this byte range; source-scoped numeric/parse properties remain partial.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/math/src/steps.rs.groupant/binding.groupant/