Elamur Groupantsstudy-prepare-finish
The groupant beside tools/enrich/src/study.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 | forward |
| ast | not materialized | |
| descriptant | present | epoch-memory |
| invariant | partial | invariant |
| tests | partial bounded checks | 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 |
|---|---|---|---|
study-prepare-finish-forward | forward_relation | partial source model | forward |
study-prepare-finish-behavior-complement | behavior_complement | partial source model | behavior-complement |
study-prepare-finish-reverse-reconstruction | reverse_reconstruction | partial source model | reverse-reconstruction |
study-prepare-finish-coupling | interaction_relation | partial source model | coupling |
Model text
Partial source model, 2026-10-29. The raw Rust file remains authoritative. AST is not materialized; byte ranges are snapshot anchors, not semantic proofs.
forward
prepare_next requires staged parent, bound seven-output receipt, valid context, at least three nonfrozen sentences and current dictionary before prompt budgeting and directory creation. finish_improve validates baseline/review/candidate lineage, binds four assessment observations and emits an evaluated-candidate receipt.
behavior-complement
Success with a wrong receipt parent, modified named output, insufficient nonfrozen capacity, changed prior-state hash or an assessment bound to another model is excluded by guards. A preexisting partial directory is possible and fails verification; it is not impossible behavior.
reverse-reconstruction
A cycle receipt constrains two learner identities, two assessment identities and four grades plus their compatible artifact histories. Existing learned/cross-exam files may have arisen earlier: receipt observations cannot reconstruct unique process attempts or prove a teacher request occurred in this invocation.
coupling
Sequential relation is prepare; external review/stage; finish, with retained study between review and candidate assessment. Cached artifacts are verified on reentry. Failure retains a prefix, with no rollback; final receipt equality is required if already present. teacher_calls=1 describes outer-cycle accounting rather than an observed call inside finish.
composition
Scope spans prepare_next, improve wrappers and finish_improve. Filesystem/process failure partitions and environment races remain unmodeled. Cost includes reading/hashing artifacts, optional learner construction and up to two missing cross-exam subprocesses. Existing binding and assessment tests cover portions; no new end-to-end finish or teacher test was executed.
invariant
Forward and reverse observations refer to the same fixed source snapshot and dependency behavior. Ordering is deterministic for fixed inputs; filesystem/environment stability is assumed where relevant. The complement is an inner bound relative to this fragment's input/output universe.
checks
See the file-level tests.md for executed bounded checks and limits. The first three model relations are specifications; tests sample correspondence rather than proving all inputs. No expanded semantic AST or executable general inverse solver exists.
epoch-memory
2026-10-29: retained the file-level 23 models and added this finer scope. An initial incidence oracle incorrectly treated three-byte fox as a token; the existing tokenizer rejects it. Corrected the independent membership matrix rather than changing production behavior. This is a model correction, not an algorithm defect.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/enrich/src/study.rs.groupant/prepare-finish.groupant/