Elamur Groupants⌕
☰

cycle-main

The groupant beside tools/cycle/src/main.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
approximantpartialmodel.md
astpartialsyntax.json
compositionpartialComposition
costpartialinvariant-and-cost
descriptantpresentdescriptant
invariantpartialinvariant-and-cost
testsbounded or representationtests

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
behavior_complementbehavior_complementpartial source derived modelbehavior_complement
reverse_reconstructionreverse_reconstructionpartial source derived modelreverse_reconstruction
resume_artifactsresume_artifactspartial source derived modelresume_artifacts
parser_coordinatesparser_coordinatespartial source derived modelparser_coordinates
model_publicationmodel_publicationpartial source derived modelmodel_publication
log_schemalog_schemapartial source derived modellog_schema

Parts

Model text

descriptant

cycle-main

Epoch 2026-10-31

Prospective inspection identified this relation before additions: CLI validates Codex-only teacher for training commands then loads Cfg. pre/post flows skip existing artifact stages and call external tools; final trees append, check and optional train precede row. Helpers convert CoNLL-U to parser ids/tokens, extract log metrics, choose newest eligible model and report exact aligned-token changes.

The production implementation is unchanged; only selected inline tests are appended. Source-bound syntax is generated twice and compared. Native checks exercise bounded helpers; entry orchestration, teacher calls, models and corpora remain source-derived. Partial syntax is not semantic proof. Preserve failure prefixes and ambiguity when refining this model.

Epoch 2026-11-01

Added two local reviewed DB commands; existing round/teacher/training paths retain their dispatch behavior. Reviewed orchestration is modeled in cycle-reviewed and the parent composition.

approximant

Forward relation

CLI validates Codex-only teacher for training commands then loads Cfg. pre/post flows skip existing artifact stages and call external tools; final trees append, check and optional train precede row. Helpers convert CoNLL-U to parser ids/tokens, extract log metrics, choose newest eligible model and report exact aligned-token changes.

Composition

build supplies default path. pipeline calls ids_of/parser/log/report/model helpers, DB calls ids_of. jobs wraps draft commands. Child scopes separate parser representation from change-report observation; external models and teacher unresolved.

Reviewed DB CLI — 2026-11-01

Dispatch promote-reviewed TREES PRAG DECISIONS and rollback-reviewed HASH to local reviewed module. CYCLE_DB determines target. Cfg load remains ahead of dispatch and requires configured MOVA_BIG; these commands never invoke teacher/model training.

catalog

Typed observations: cycle-main

Subject: tools/cycle/src/main.rs, SHA-256 c901ffdda1a67fa69ed553c5d0c8e7e83d1aa0c23ce8a7bc74d53f30e5c5b108. Domain: bounded source inputs, declared filesystem/environment state and normal allocations. Each view is a partial source-derived relation, not a universal proof. Observation is one helper call or the explicitly described trace. Checks are in tests.md; unexecuted effects remain gaps.

behavior_complement

Invalid configured training teacher is rejected before Cfg load. parser_input count mismatch excludes file writes for ids/tokens but not directory creation. Token-form mismatch excludes that sentence's change counts.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

reverse_reconstruction

Report reconstructs only compatible counts/ranked relation transitions, not original trees; unreported sentences, words beyond top examples and lexical inputs remain ambiguous. Missing metrics NaN merges many malformed/missing logs.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

resume_artifacts

Each existing marker skips its stage; existence is not validation or a transaction. A final artifact can survive failure before table row and affect restart.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

parser_coordinates

Only ten-column rows without '-' or '.' id contribute forms; sent_id metadata enters independently. Equal counts validate cardinality, not semantic id/token pairing.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

model_publication

gparse_for maximizes (mtime,path) for model.safetensors excluding .part and round j≥k; unrelated names qualify. part_dir creates sibling suffix, train success precedes rename.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

log_schema

apply_metrics first '| final | Rust |' numeric cells indices 6/7, first cost marker and first ': +' integer; missing UAS/LAS/cost NaN, missing added zero. num_after consumes only digits/dots.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in tests.md; no finite check establishes all inputs.

tests

Checks and limits

tests::changes_counts_moves; tests::part_dir_is_a_sibling; tests::metrics_from_logs; tests::parser_input_matches_ids. No round/training/report CLI.

Exact native outcomes and command metadata are retained in the external wave33 cycle receipt bundle. Selected names run twice in debug and release; syntax is independently regenerated twice. Compiled mutations and source restoration receipts distinguish behavior changes from assertion-only controls. These claims are finalized in evidence.md. No live teacher, real model/corpus, benchmark CLI or lesson-learning CLI is run.

evidence

Wave33 evidence — 2026-10-31

Across this nine-source slice, 14 exact native names passed twice in debug and twice in release (56 runs); sorted target-qualified outcomes agree. Six compiled behavior controls fail named assertions and exact restored bytes pass again. One exploratory parser-count control failed by unwrap panic; it was restored and replaced by a form-column assertion control. No assertion-only control is counted.

Nine syntax projections were generated twice and byte compared. Production source prefixes match the base exactly; four files append only formatted cfg(test) blocks. Local groupants validation reports valid=true, zero diagnostics. Counts: nine source roots, ten scoped children, 114 typed views.

See tests.md for this source's precise executed or representation-only coverage. Native evidence does not apply to ignored live teacher or lesson execution. No teacher/API, model/corpus, learning or benchmark CLI was run. Receipts retain numeric statuses, exact names, log/source hashes, original backups, unique mutation replacements and restoration assertions. The external bundle is named octopus-wave33-cycle-2026-10-31.

invariant-and-cost

Invariant and cost

parser_input requires id count=nonempty token groups, creates dirs before validation; model .part excluded from selection and renamed after successful train. changes compares exact token form arrays, denominator max(words,1). Existing files signal completion without content validation.

Parser/report O(bytes+words+distinct changes log distinct changes); model selection O(directory entries). Wrapper cost sums subprocess costs and filesystem bytes, unmeasured. Stages can leave durable prefixes on failure.

No executable percentage surrogate is claimed: effects, categorical acceptance and report text have no justified approximate metric here. Future profiles require a useful metric, zero case, reference and evidence at caps 1/2/5/10%.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/cycle/src/main.rs.groupant/