Elamur GroupantsโŒ•
โ˜ฐ

prag-feats-features-selection

The groupant beside prag/src/feats.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

componentlabelwhere
approximantpartial source modelforward
astpartial../../syntax.json
compositionpartial source modelcomposition
costpartial source modelcost
descriptantpartial source modelepoch-memory
invariantpartial source modelinvariant
testspartial source modelchecks

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.

idtypelabelwhere
prag-feats-features-selection-behavior-complementbehavior_complementpartial source modelbehavior-complement
prag-feats-features-selection-reverse-reconstructionreverse_reconstructionpartial source modelreverse-reconstruction
prag-feats-features-selection-couplingconstraint_relationpartial source modelcoupling
prag-feats-features-selection-observation-fiberinformation_losspartial source modelobservation-fiber

Model text

Filter quotation as speech, concatenate utterance and retain narrator frame.

Source snapshot: 2a1a0bd6e43dd16f7284e993f30d0f4971c447c9e2080616033d19ac8693611f. All relations concern this snapshot on the stated admissible domain. Models are source-derived partial relations; finite fixtures are evidence, not general proofs. Complement is a conservative inner constraint model. Reverse reconstruction preserves all compatible antecedents conceptually; no enumerating solver is implemented.

forward

Initial frame speech cue admits all candidate spans. Otherwise admit carried span at byte zero, at least four whitespace words, no preceding alphanumeric, or !/? ending. Join trimmed retained contents with spaces; if none retain entire text and empty frame. quote is cont if carried, else none/whole/framed according to retained spans and alphabetic frame.

behavior-complement

Universe: well-formed ordered byte spans of item text and carry flag to (utterance,frame,quote). Excludes selecting an interior one-word title without speech cue, and quote=whole when retained spans leave alphabetic narrator text. quote=cont has priority even with no retained span.

reverse-reconstruction

For selected utterance/frame retain compatible delimiter spellings and whitespace variants inside span edges; joining multiple spans loses individual boundaries. quote=none admits unquoted texts and rejected short quotations.

invariant

frame_of requires ordered nonoverlapping valid UTF-8 bounds; public quote_spans provides them. This private function does not validate arbitrary supplied intervals.

cost

Text/frame scans and string copies O(B) plus narrator cue matching; candidate whitespace scans total O(B) on nonoverlapping spans. No measured speedup or executable approximation error bound is claimed.

coupling

Speech detection uses initial candidate frame; final sayv later uses frame after filtering. These frames may differ; retain both in the state model.

observation-fiber

Trim and join discard boundary whitespace and number of quotation spans. The speech heuristic does not distinguish every title from direct speech.

composition

Speech detection uses initial candidate frame; final sayv later uses frame after filtering. These frames may differ; retain both in the state model. No child fragments; this is one scoped relation.

checks

supplied_ud_projection_and_quote_selection supplies framed speech, rejected Book title, whole Hello and carried utterance with exact closure input assertions. See the file-level evidence.md for exact log hashes and repeated run outcomes.

epoch-memory

2026-10-30 wave13: initial source-bound model. Prospective model preceded edits. Existing normalization and quote rules were retained; tests expose their information loss. Source identity witness justified qualifying predecessor keys by Src. Annotation is isolated behind a private closure used by the production row wrapper; fixtures do not call or certify tokenizer/parser. AST is a partial syntax projection, macro bodies opaque; semantic AST, live model/corpus and general proof remain absent.

This page in Markdown โ†’ https://md.groupants.elamur.ai/free/en/mova/prag/src/feats.rs.groupant/features.groupant/selection.groupant/