Elamur Groupantsmova-whale-world-dossier
The groupant beside tools/whale-world/src/dossier.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 approximant |
| ast | partial | syntax.json |
| composition | partial | Composition |
| cost | partial | Cost |
| descriptant | present | Descriptant |
| invariant | partial | Invariant |
| tests | finite evidence | tests.md |
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 |
Model text
Descriptant
The dossier of one document (mova-whale-dossier-v1): an append-only texts.jsonl index and texts/<epoch>-<kind>-<n>.md files (originals zstd), with licence, licence class, public_ok, hash, words, created time, inputs and citation counts; selection by epoch and kind from the index alone; hash-checked reads. Source: tools/whale-world/src/dossier.rs.
Forward approximant
add: lock -> n = 1 + rows of (epoch, kind) -> refuse existing id -> create_new file (zstd for originals) -> fsync -> row (public_ok rule, citations scan) -> append + fsync. read: file -> decompress -> sha256 == row or refuse. citations: numeric and author-year markers against [n] lines and reference-section lines.
Invariant
Rows are never rewritten and files never replaced; public_ok implies kind original, an open library licence, not local-only and public_originals; a changed file is refused by every reader.
Cost
add and select are linear in the number of rows (the index is read whole); read is linear in the text; citations linear in text length times reference lines.
Composition
Parent: whale-world; child module import; used by world::add and main.
behavior_complement
Does not delete or edit texts, decide quality, resolve citations to documents, or publish anything; an author-year marker is matched only by surname and year. Evidence status: source-derived model with finite checks, not a proof.
reverse_reconstruction
From a row the stored text is recovered exactly (hash-checked); the source path or URL and inputs name where it came from; an original is recovered after zstd decompression. Citation counts do not recover the markers.
tests.md
cargo test --manifest-path tools/whale-world/Cargo.toml --offline with WORLD_EN_MODEL=$MOVA_DATA/download/0/0.4/en.bin (9 unit tests: dossier append-only/select/changed files, licence negative controls, citation markers, importers incl. a topic corpus with a Sea Spider page and a consultant draft, entity extraction, LRU, two byte-identical world builds with cited queries, incremental epochs with old worlds intact and a changed file refused; 2026-11-05). Live 2026-11-05 on six papers: two independent imports and builds byte-identical.
These are finite fixtures and live samples; not a proof.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/whale-world/src/dossier.rs.groupant/