Elamur Groupantsprag-eval-training
The groupant beside prag/src/eval.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 source model | forward |
| ast | partial | ../syntax.json |
| composition | partial source model | composition |
| cost | partial source model | cost |
| descriptant | partial source model | epoch-memory |
| invariant | partial source model | invariant |
| tests | partial source model | checks |
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 |
|---|---|---|---|
prag-eval-training-behavior-complement | behavior_complement | partial source model | behavior-complement |
prag-eval-training-reverse-reconstruction | reverse_reconstruction | partial source model | reverse-reconstruction |
prag-eval-training-specialization | observation_relation | partial source model | specialization |
Model text
forward
train_rows clones rows in idx order, replacing previous-act with gold[prev[i]].act or start. fit_set dispatches six model::fit calls using ys[k], Field::ALL[k].classes(), the same x, keys, dictionary and seed.
invariant
Input rows remain unchanged. Valid domain requires six aligned label vectors, valid labels, dictionary-compatible encoded rows and keys aligned with x; malformed training shapes are not newly validated.
cost
Row preparation O(selected*S) cloning plus string bytes. Training is sum_k T(kind,N,S,C_k), with six retained models and learner-specific temporary storage.
behavior-complement
A prepared row borrowing its own act rather than its predecessor, mutation of input rows, or swapping field labels/names is excluded in this domain. Domain and observation are those of forward; this is a source-derived partial exclusion model, not a general proof.
reverse-reconstruction
Observed prepared start may arise from no predecessor or absent predecessor gold. Original previous-act slot is overwritten and unrecoverable; compatible source rows differ freely in that slot. A fitted model generally has many compatible training histories. The relation is all compatible antecedents, with no uniqueness claim.
specialization
Field-routing oracle supplies constant last-class labels separately for all six fields and checks all three actual learner kinds on a two-row fixture.
composition
Children share source identity and dictionary/sample or record indices. Parent order and domains constrain joins; leaf observations alone do not establish external orchestration. Overlapping scopes must not double-count costs.
checks
Seven wave14_models tests: four evaluation and three data; supplied inputs only. Debug twice and release twice; compiled production mutation per source must fail a named assertion. Syntax generated twice identically. Receipts summarize final evidence.
epoch-memory
2026-10-30: first coverage of these previously unmodeled files. Prospective review predicted no valid-domain metric change and O(N) extra validation. Old source accepted mismatched vectors; the rejection witness failed (six other tests passed). Narrow assertions preserve tuple API and reject invalid domains. Data production behavior remains unchanged. Tiny two-row fitting exercises three learners without any training job.
limitations
Partial syntax projection, opaque macros, no semantic AST or universal proof. No live parser, DB subprocess, external corpus, teacher, GPU or training job. External IO, report formatting, selection corner cases, malformed Set/Train and empty tale windows remain outside executed scope. No performance claim.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/prag/src/eval.rs.groupant/training.groupant/