Elamur Groupantsmova-serve-engine
The groupant beside tools/mova-serve/src/engine.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 | Approximant |
| Ast | partial | syntax.json |
| Cost | symbolic source derived | Cost |
| Descriptant | present | Descriptant |
| Invariant | partial | Invariant |
| Tests | executed bounded checks | Tests |
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.
| relation | type | label | where |
|---|---|---|---|
| Behavior complement | behavior complement | partial source derived model | Behavior complement |
| Reverse reconstruction | reverse reconstruction | partial source derived model | Reverse reconstruction |
| Model presence | model presence | partial source derived model | Model presence |
Model text
Descriptant
The analysis: mova-parse's dispatcher (auto or a chosen language), CoNLL-U, tokens read back from it, and for English sentences prag's frames (one discourse context a text), implicatures, presuppositions, maxims, irony and the speech-act perceptron's labels with margins; model versions by SHA-256.
Approximant
analyse(t, c) = (sents = parse_c(t); conllu(sents); per English parsed sentence: frames_of_tree, infer, irony; Labeler.label(English sentences)).
Invariant
Deterministic in (text, choice, model files); JSON tokens equal the CoNLL-U rows; frames and pragmatics only for English sentences whose every token has a head.
Cost
≈18 ms cold for 100 words, 0.66 s for 5 000 words (2026-11-06), dominated by the two annotations (parse and the SLM's features).
Behavior complement
No frames or pragmatics for French; no Wiktionary expression lens; no coreference beyond prag's salience context.
Reverse reconstruction
The CoNLL-U recovers the tokens and the language comments; the versions recover the model files.
Model presence
Without an English model or silver run there are no labels; without a model a sentence is reported unparsed.
Tests
engine::tests (unparsed French without a model; the SLM rows' parsing); server::tests through the API; two bench
processes with identical digests.
Provenance
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/mova-serve/src/engine.rs.groupant/