Elamur Groupantsmath-solution-memory
The groupant beside math/src/memory.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 |
| composition | partial | composition |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | partial bounded evidence | 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 |
failure_precedence | failure_precedence | partial source derived model | failure_precedence |
representation_transport | representation_transport | partial source derived model | representation_transport |
resource_boundary | resource_boundary | partial source derived model | resource_boundary |
domain_boundary | domain_boundary | partial source derived model | domain_boundary |
Parts
Model text
descriptant
Source-bound description
numbers scans unsigned digits/listed words (half=>2). Features stems words plus number/magnitude/markers. train sorts ids/classes, averaged perceptron; guess lexical tie wins. Template excludes0/1, replaces matching literal with FIRST index, requires every index used; instantiate parenthesizes values. Similarity2LCS/(a+b), zero on empty or>400 length.
Epoch 2026-10-31: prospective model before bounded synthetic tests; production behavior unchanged. No external memory load/generalization. load drops malformed lines and read errors; learned accuracy unmeasured.
approximant
Forward relation
numbers scans unsigned digits/listed words (half=>2). Features stems words plus number/magnitude/markers. train sorts ids/classes, averaged perceptron; guess lexical tie wins. Template excludes0/1, replaces matching literal with FIRST index, requires every index used; instantiate parenthesizes values. Similarity2LCS/(a+b), zero on empty or>400 length.
Domain: Synthetic only; no-panic comparison domain requires non-NaN comparable accumulated scores as well as comparable weights; finite input weights alone do not prevent score overflow or NaN; usize tick overflow unguarded. Coordinates are chars, not bytes.
Executable Rust approximant unsupported: categorical/effect observations have no useful percentage metric yet. Future caps1/2/5/10% only; no automatic substitution.
invariant
Invariant and cost
Empty classes excludes Guess. Empty/>400 skeleton excludes positive similarity. Missing slot excludes Some instantiate; similarity alone does not verify plan.
Ordering: Reject0/1→remove root id/option→map except metadata→require complete slots.
Cost: Text scans+big rational work; training O(epochsmemosclassesfeatures), LCS O(ab) time/space capped400; template scans plus numeric matching.
composition
Composition and epoch memory
Verified B admitted even incorrect; run_c gates instantiated template. Text substitution is lossy, not bijection.
No external memory load/generalization. load drops malformed lines and read errors; learned accuracy unmeasured.
Partial syntax and finite tests do not establish all caller assumptions.
models
behavior_complement
Empty classes excludes Guess. Empty/>400 skeleton excludes positive similarity. Missing slot excludes Some instantiate; similarity alone does not verify plan.
reverse_reconstruction
Templates cannot recover removed values/id/option or numeric JSON type; repeated equal question numbers map first slot and fail completeness. Features and winning class admit many texts/weights.
failure_precedence
Reject0/1→remove root id/option→map except metadata→require complete slots.
representation_transport
Verified B admitted even incorrect; run_c gates instantiated template. Text substitution is lossy, not bijection.
resource_boundary
Text scans+big rational work; training O(epochsmemosclassesfeatures), LCS O(ab) time/space capped400; template scans plus numeric matching.
domain_boundary
Synthetic only; no-panic comparison domain requires non-NaN comparable accumulated scores as well as comparable weights; finite input weights alone do not prevent score overflow or NaN; usize tick overflow unguarded. Coordinates are chars, not bytes.
tests
Evidence, 2026-10-31
Selected actual-source checks: cargo test -p math --lib memory::. Exact commands, numeric statuses, normalized outcomes, compiled assertion controls and restoration receipts are in the external wave28 mathcontrol run.
No external memory load/generalization. load drops malformed lines and read errors; learned accuracy unmeasured.
No live teacher/API, trained model, corpus/held-out, GPU, DB or expensive benchmark run. Parent syntax generated twice identically. Scope remains partial.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/math/src/memory.rs.groupant/