Elamur Groupants⌕
☰

prag-frames-md

The groupant beside prag/src/frames_md.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
astpartialsyntax.json
compositionpartial source modelcomposition
costpartial source modelcost
descriptantpartial source modelepoch-memory
invariantpartial source modelinvariant
testsbounded native and inspectedevidence

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-frames-md-behavior-complementbehavior_complementpartial source modelbehavior-complement
prag-frames-md-reverse-reconstructionreverse_reconstructionpartial source modelreverse-reconstruction
prag-frames-md-column-transportfield_projectionpartial source modelcolumn-transport
prag-frames-md-provenance-lossprovenance_fiberpartial source modelprovenance-loss

Parts

Model text

model

English/Hebrew supplied-frame Markdown transport

Subject: prag/src/frames_md.rs SHA-256 023ca98fa461e9d1545915abd8071ecb542cc61999bbf5a054b712e5edf9b5e3. All entries share this exact revision. Status: partial source-derived relations, finite native evidence for pure seams, source inspection for effects. Domain: finite supplied inputs under the stated resource/arithmetic premises. Exclusions are inner source-derived families, not a complete behavioral complement. Reverse relations retain compatible antecedents but no complete inverse solver is implemented.

forward

English scans non-Bare frames in order and renders ten columns with original one-based indices. Hebrew projects accepted CoNLL-U fields, joins normalized MIds, and renders six columns. Empty inputs produce the respective no-frame note. These observations are bytes, not linguistic truth.

behavior-complement

Universe: finite supplied frame/word vectors paired with proposed bytes. Excluded families include a Bare frame rendered as its own row, dense renumbering after skipped Bare frames, a Hebrew row without nonempty Frame metadata and an earlier duplicate MId winning a participant lookup. These are source-derived inner exclusions, not exhaustive absent behavior.

reverse-reconstruction

For output y retain every vector whose rendered projections yield y. Varying English src/cause/next/quote fields alone leaves y unchanged. Hebrew accent removal, rejected CoNLL-U lines and overwritten MIds create additional fibers. Empty notes admit arbitrary Bare-only or frameless vectors. No original parse is uniquely recoverable.

invariant

Inputs are unchanged; visible English coordinates retain original indices. Escaping does not remove line breaks; complete Markdown sanitation is not guaranteed.

cost

English O(sum entity subtree bytes across recursive ancestors + output + oblique/role lookup work); chain copying can be quadratic. Hebrew O(B+W log W+VW+P log W+output) for bytes B, words W, verbs V and participant IDs P.

column-transport

Fix roles and flags: who/what/to-whom carry agent/patient/recipient respectively; unsaid applies only to filled slots. Native exact-column checks preserve this wiring.

provenance-loss

Holding rendered fields fixed while changing src/cause/next/quote leaves output equal. The explicit two-antecedent native fixture falsifies unique provenance recovery.

composition

English composes phrase, ent, TAM and escaping. Hebrew composes hwords with hebrew_table. Inspected roundtrip site_code::tables removes the exact trailing note; its Hebrew sites call hwords then hebrew_table. External caller execution is not claimed.

epoch-memory

Wave14 first modeling epoch, 2026-10-30: preserves rendering and tuning selection semantics. Two public pure tuning seams are now actual production-path calls. Six rendering tests and four tuning tests record field loss, tolerance/duplicate behavior, strict selection thresholds and domain limits. No bug repair or speed claim. Existing semantics retained include newline passthrough, first key/last MId collision, zip truncation and unchecked score arithmetic. Evidence is finite; no parser, teacher, filesystem corpus, journal-backed tune, optimizer training or GPU execution.

limitations

Partial syntax binds source bytes; macro bodies remain opaque and semantic AST is absent. Native/structural checks are not universal proofs or full-system coverage. No approximation error budget or automatic substitution is claimed. File evidence distinguishes executed pure paths and inspected effects.

evidence

Evidence, 2026-10-30

Source SHA-256: 023ca98fa461e9d1545915abd8071ecb542cc61999bbf5a054b712e5edf9b5e3.

Receipt set octopus-wave14-formats-2026-10-30. Six frames_md and four tune tests pass twice debug/twice release with identical ten named outcomes. Compiled production controls: 0591 -> 0592 fails text_escape_and_accent_projection_have_distinct_fibers; p<0.05 -> p<=0.05 fails rejection_thresholds_and_configuration_order. Sources restore exact bytes before positives. Both partial syntax projections repeat identically. Finite families: 169 escape pairs, 27 neighbor cases, 60 selection cases in both orders, explicit boundaries/duplicates/NaN and exact columns. No repair, performance or universal correctness claim; unrelated existing compiler warnings remain.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/prag/src/frames_md.rs.groupant/