Elamur Groupantsrt-mixed-heb-tokens
The groupant beside tools/roundtrip/src/mixed/heb.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 | invariant |
| 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.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial source derived model | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
raw_chunk | raw_chunk | partial source derived model | raw_chunk |
gap_index | gap_index | partial source derived model | gap_index |
wrap_boundary | wrap_boundary | partial source derived model | wrap_boundary |
tail_restore | tail_restore | partial source derived model | tail_restore |
Model text
descriptant
Descriptant
Prospective epoch
2026-10-31 source review predicted: tokens_down(tokens_up(p))=p under tokenizer coverage; wrap bytes and raw whitespace chunks restore original representation. Resource model: Tokenization/NFC O(B) plus ordered raw/gap map lookup; O(B+tokens) retained text. External prospective.md predates test additions and preserves APIs, effect order, hazards and interaction hypotheses.
Observed epoch
Bounded real helper checks in tests.md constrain these models; syntax is a partial lexical projection with no macro expansion, name resolution or proof. Production expressions unchanged; six inline test modules added. Compiled mutants are temporary and exactly restored. Existing malformed-input and nonfinite domains remain explicit, not silently repaired. Parent composes this scope with siblings without adding overlapping costs twice.
Narrow NFC token residual codec scope; finite checks do not cover all inputs. Parent effects and other functions are outside this fragment.
approximant
Forward relation
tokens_down(tokens_up(p))=p under tokenizer coverage; wrap bytes and raw whitespace chunks restore original representation.
Domain: Scope: NFC token residual codec. Valid UTF-8; words/forms lengths agree; positive group sizes sum to words; heads are zero or in range; generated scripts use unique valid Take indices. Residual metrics are finite nonnegative estimates, not compressed byte lengths.
invariant
Invariant and cost
Invariant
Scope: NFC token residual codec. Valid UTF-8; words/forms lengths agree; positive group sizes sum to words; heads are zero or in range; generated scripts use unique valid Take indices. Residual metrics are finite nonnegative estimates, not compressed byte lengths.
Cost
Tokenization/NFC O(B) plus ordered raw/gap map lookup; O(B+tokens) retained text.
Limits
Narrow NFC token residual codec scope; finite checks do not cover all inputs. Parent effects and other functions are outside this fragment. No percentage surrogate is implemented: categorical parsing/effect observations lack an admitted numerical error metric. Any future profile must define its metric/domain and use only1/2/5/10 percent caps.
models
Typed source-bound relations
Subject: exact source SHA-256 in manifest. Domain and resource premises are in invariant.md. Each model is partial, source-derived, over a single helper invocation or the described finite trace. Complements are proposed inner exclusions under those premises; reverse relations retain all compatible antecedents, with no unique inverse claim. The oracle is the named checks in tests.md; unexecuted cases remain hypotheses. No numerical approximation certificate.
behavior_complement
Within this scoped helper relation, outputs violating the stated forward equation or its preserved fields are excluded under the domain premises; this is a partial source-derived inner exclusion, not a claim about untested panics. raw map keys are whitespace-run ordinals, not token positions.
reverse_reconstruction
Original gaps/NFC chunks are absent from canonical tokens; these erased variants form the inverse fiber.
raw_chunk
raw map keys are whitespace-run ordinals, not token positions
gap_index
gaps indexed by (sentence,token) and default canonical_gap
wrap_boundary
all-wrapper text core returns lead=len and trail=0
tail_restore
tail appended before unnfc and exterior wrap
tests
Checks and evidence
Named source/helper checks: mixed::heb::tests::tokens_come_back, mixed::heb::tests::script_gives_the_tree_back.
Wave32 on2026-10-31:10 distinct exact tests passed twice debug and twice release, with equal sorted named outcomes. This node claims only its listed subset; empty subset means source review only. Five actual source-behavior mutants compiled, failed assertions, restored exact bytes/hash, and passed focused green reruns. No assertion-only mutants; existing swapped-input controls are separate. Receipt bundle: $MOVA_DATA/runs/octopus-wave32-rt_mixed-2026-10-31 (native-commands.json, source-test-map.json, mutation-results.json, .status/.log). No live corpus, teacher, frozen assessment, download or production data used.
Narrow NFC token residual codec scope; finite checks do not cover all inputs. Parent effects and other functions are outside this fragment.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/roundtrip/src/mixed/heb.rs.groupant/tokens.groupant/