Elamur Groupantsworld-primers-segmentation
The groupant beside world/src/primers.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 |
|---|---|---|---|
world-primers-segmentation-behavior-complement | behavior_complement | partial source model | behavior-complement |
world-primers-segmentation-reverse-reconstruction | reverse_reconstruction | partial source model | reverse-reconstruction |
world-primers-segmentation-coupling | constraint_relation | partial source model | coupling |
world-primers-segmentation-observation-fiber | information_loss | partial source model | observation-fiber |
Model text
forward
Split on line-start one-to-three digit labels followed by dot-space, concatenate continuation lines, then reject short/drill/garbled/no-question text. Drop a leading numeric label, clean spacing, choose last question mark and preceding dot/semicolon, capitalize first scalar.
invariant
Original chunk order survives; alphabetic-word minimum is six. Prefix material before the first label can form its own chunk. Decimal dots and multiple question marks can yield imperfect segmentation; no general sentence parser is claimed.
cost
O(B) scans and allocations per chunk, ignoring repeated string concatenation growth implementation details; total stored text O(B).
composition
problems outputs chunks; inline extraction calls word_problem while paged/chapter extraction calls story_and_question with separately associated numeric labels. Finite fixture checks the actual inline helper chain.
epoch-memory
2026-10-30 wave11. Synthetic old-code witnesses failed before narrow repairs: frame display lost attrs/cond; marker spelling priority selected an earlier answer; unchecked arithmetic panicked; dotted page answer 57 was mistaken for an item. Current tests preserve these counterexamples. No optimization, percentage-error surrogate or general correctness claim.
checks
cargo test -p world --lib primers::wave11_models -- --test-threads=1. See file-root evidence.md for repeated debug/release identities and compiled mutants. Syntax is a source-bound partial projection; macro bodies are opaque.
behavior-complement
Subject/revision: manifest source hash and this scope. Universe: bounded supplied inputs and terminating observations under fixed dependencies and successful allocation. Inner absent-behavior model excludes any transition violating: Original chunk order survives; alphabetic-word minimum is six. Prefix material before the first label can form its own chunk. Decimal dots and multiple question marks can yield imperfect segmentation; no general sentence parser is claimed. External effects and unmodeled malformed states are unknown, not excluded. Finite tests support only named fixtures.
reverse-reconstruction
Subject/revision: manifest source hash. Domain: the scoped forward input/output relation, fixed dependencies and successful allocation. Operator is the full compatible preimage: All line layouts and problem labels that collapse to the same story/question pair are compatible. Removed numbering, whitespace and post-question text cannot be reconstructed uniquely. This is a partial source model, not an implemented complete inverse search; no unique causal explanation is inferred.
coupling
Subject/revision: manifest source hash. Constraint relation on shared identities/intermediate states: problems outputs chunks; inline extraction calls word_problem while paged/chapter extraction calls story_and_question with separately associated numeric labels. Finite fixture checks the actual inline helper chain. Source inspection and actual finite joins are distinguished.
observation-fiber
Subject/revision: manifest source hash. Information-loss relation holds inputs in the same fiber when their declared observation agrees: Drill/fraction exclusion partitions a supplied chunk universe using character flags, word counts and punctuation. A rejected chunk does not reveal which discarded OCR text was intended. Finite correspondence only; unrestricted soundness/completeness is not established.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/world/src/primers.rs.groupant/segmentation.groupant/