Elamur Groupantsworld-school-scenario-assessment
The groupant beside world/src/school.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 | wave34-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 |
|---|---|---|---|
world-school-scenario-assessment-behavior-complement | behavior_complement | partial source model | behavior-complement |
world-school-scenario-assessment-reverse-reconstruction | reverse_reconstruction | partial source model | reverse-reconstruction |
world-school-scenario-assessment-coupling | constraint_relation | partial source model | coupling |
world-school-scenario-assessment-observation-fiber | information_loss | partial source model | observation-fiber |
Parts
Model text
model
world-school-scenario-assessment
forward
Scenario generators create labeled Problems; scenario retries negative answers. Courses designate exercise and scoring sources but call parser/training/I/O. actor_learn_answer enumerates up to six effects for at most three verbs and votes only coordinates shared by all fitting assignments. assessment_fold uses wrapping-u64 repeat hash mixing or u128(index*prime+repeat) modulo folds; zero folds yields None.
invariant
Positive folds assign exactly one check fold; all other folds learn that row. Equal hashes stay together regardless of index. Noncoprime multiplier/fold pairs may produce empty folds; hash equality is not semantic/provenance identity. Elementary exact template separation does not imply scenario pool separation.
cost
Fold O(1), no allocation. Effect enumeration at most 6^3=216 solves. Scenario retry lacks a general bound for arbitrary parameters. Parser/teacher/corpus/optimizer costs opaque.
composition
cross_check_k uses one helper for learn/check predicates. Native u128/masked-u64 oracle tests repeats 0..8, folds 1/2/5/7, extreme indices and equal hashes. assessment_fold is public for parent joins. No dataset, teacher, course, heldout or live parser access.
epoch-memory
2026-10-30 initial school source-bound epoch. Prospective model: keep O(1) arithmetic/partition cost, widen aggregation scalars, and expose actual codecs without filesystem effects. Old native witnesses found unchecked operation, score, marginal vote, partition and duplicate-vote arithmetic; legacy serialization synthesized zero accumulators. Review found that dropping explicit zero accumulators also changed the prediction table. Both activation directions now have named tests. No speedup, approximate-error profile or general semantic proof is claimed. Unchecked update/world arithmetic retains representability premises.
checks
cargo test -p world --lib school::wave12_models -- --test-threads=1. Root evidence.md binds exact source/evidence hashes, repeated debug/release outcomes and a compiled production mutation. Named fixture domains appear in composition. Syntax is a source-bound partial projection; macros remain opaque. No semantic AST or live school execution.
behavior-complement
Subject/revision: manifest hash and byte scope. Universe: supplied finite inputs, fixed dependencies, successful allocation and the explicit arithmetic premises of this scope. Inner absent-behavior relation: A row cannot satisfy both part==fold and part!=fold. Group repeat multiplication cannot panic or differ by debug/release overflow handling. No claim excludes curriculum overlap or external dataset leakage. These are source-derived constraints with finite named witnesses where indicated; external behavior remains unknown.
reverse-reconstruction
Subject/revision: manifest hash. Operator: compatible preimage/converse, holding dependencies and admissibility fixed. For fixed repeat/folds retain every index/hash giving the observed residue; affine maps can be many-to-one. Group hashes lose text. For a learned coordinate retain every labeled reading whose fitting assignments unanimously constrain it, keeping all alternatives for unconstrained coordinates. Partial relational description, not a complete inverse search or a unique causal history.
coupling
Subject/revision: manifest hash. Constraint relation on stage identities and ordered intermediate states: cross_check_k uses one helper for learn/check predicates. Native u128/masked-u64 oracle tests repeats 0..8, folds 1/2/5/7, extreme indices and equal hashes. assessment_fold is public for parent joins. No dataset, teacher, course, heldout or live parser access. Supplied native joins and source inspection are distinct evidence.
observation-fiber
Subject/revision: manifest hash. Domain: scoped inputs with the declared observation. The information-loss relation equates inputs whose observation agrees: A correct answer can constrain one effect side and leave others free. Scores cannot recover assignments or certify assessment provenance. GIVE_SAY[1] appears in both ordinary scenario pools; this overlap is explicitly retained. Finite tests do not establish unrestricted completeness.
wave34-fold-residues
Current nested child world-school-scenario-assessment-fold-residues: folds=0 returns None. Otherwise ungrouped result is (indexp[repeat mod 5]+repeat) mod folds in u128. Grouped result is (hash XOR wrapping64(repeatC)) mod folds; index is ignored.
Parent interaction: cross_check_k applies the same helper to row selection for both learn !=fold and check ==fold. No dataset or training is executed; helper membership does not certify the provenance of hashed text or extra training rows.
Earlier checks and epoch observations above remain historical, not new execution receipts. See file-root wave34-evidence.md for current selections.
wave34-composition
Wave34 composition
fold-residues feeds both !=fold learning and ==fold checking in cross_check_k using identical index/hash/repeat/folds. Zero folds has no helper result and the loop has no iterations. This decomposition models partition selection only, without executing datasets, extra-row acquisition, training or evaluation; it does not prove assessment provenance.
The previous model.md and its composition remain historical epoch memory. Current scoped relations use the manifest source hash; partial syntax and finite helper tests do not prove semantic completeness.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/world/src/school.rs.groupant/scenario-assessment.groupant/