Elamur Groupantsonto-tests
The groupant beside onto/src/tests.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 |
| composition | partial | composition |
| cost | partial | Cost |
| descriptant | present | descriptant |
| invariant | partial | model.md |
| 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 |
fixture_transport | fixture_transport | partial source derived model | fixture_transport |
layer_projection | layer_projection | partial source derived model | layer_projection |
proof_witness | proof_witness | partial source derived model | proof_witness |
scope_boundary | scope_boundary | partial source derived model | scope_boundary |
Parts
Model text
descriptant
onto-tests
mini constructs17-concept sorted dictionary/CSR fixture, whale senses[person,animal], then store::check. kb adds layers/defaults/strict seeds and Nixon diamond. Fourteen tests assert snapshot gates, inheritance/proofs, query/eval/bridge/pronoun/HS and completeness behavior; tempfile cleanup best effort.
Epoch 2026-10-31
Prospective review preceded inline test edits. Tests used bounded synthetic CPU fixtures only. Missing knowledge distinct from false/undetermined. leak_all_layers is not a compiled mutation. Fixed synthetic fixtures; initial query runs allowed ambient prior lookup, corrected by four isolated reruns with absent ONTO_PRIOR (prior-isolation.json).
No production optimization, live teacher, corpus, assessment or database mutation. Registry presence denotes partial modeling; syntax is lexical and generated twice identically.
approximant
Forward model: onto-tests
mini constructs17-concept sorted dictionary/CSR fixture, whale senses[person,animal], then store::check. kb adds layers/defaults/strict seeds and Nixon diamond. Fourteen tests assert snapshot gates, inheritance/proofs, query/eval/bridge/pronoun/HS and completeness behavior; tempfile cleanup best effort.
Domain
Finite input arrays/text with resource availability and the source-specific bounds and representation assumptions in invariant.md. Exact categorical semantics; no executable percent surrogate is supplied. 1/2/5/10% caps need a meaningful metric and are not claimed.
Relations
mini yields C=17 and whale senses=[15,8]. Given K=kb(), prop(K,crow,CanSpeak,0)=false; prop(K,crow,CanSpeak,1)=prop(K,crow,CanSpeak,2)=true. prop(K,Nixon,Pacifist,0)=Undetermined with two one-edge extra proofs. unpack(pack(mini()))=mini() on the fixture.
invariant
Invariant and cost
Missing knowledge distinct from false/undetermined. leak_all_layers is not a compiled mutation. Fixed synthetic fixtures; initial query runs allowed ambient prior lookup, corrected by four isolated reruns with absent ONTO_PRIOR (prior-isolation.json).
Cost
Fixed17 concepts and small temp TSVs; inference/storage delegated; no throughput claim. Temp paths PID-based, cleanup errors ignored.
No allocation, integer-overflow, malformed public-state or external-effect guarantee beyond explicitly described branches.
composition
Composition
Synthetic Graph → store → Kb → query/eval/pron/HS, with explicit fixture-only assertions; core completeness tests exercise parser without real core data.
Child fragments specialize concrete byte scopes and share the parent syntax projection; they do not duplicate whole-file guarantees. No composition-wide approximation error budget is established.
models
Typed obernant catalog
Partial source-derived relations, not proofs. Complement excludes incompatible observations; reverse describes compatible antecedents.
behavior_complement
Passing examples exclude guessing for blorft and false for missing fish wing. Hash test excludes changed fixture bytes; finite pass cannot exclude all malformed inputs.
reverse_reconstruction
Outcomes admit many implementations/graphs with same observed answers. Snapshot Graph equality constrains round-trip state, not arbitrary format implementations.
fixture_transport
store validates mini; roundtrip compares all Graph arrays; selected-byte flips fail unpack.
layer_projection
Real crow false speaking; fable/tale true. leak_all_layers makes real true showing sensitive observation.
proof_witness
Penguin direct exception empty chain beats bird default; Nixon two one-edge extra proofs give Undetermined.
scope_boundary
Synthetic hash file, not held-out assessment. Same-process concurrent duplicate tests could collide in PID tempfile names.
tests
Evidence, 2026-10-31
Selected command: cargo test --offline --locked -p onto --bin onto -- --exact --test-threads=1 tests::snapshot_round_trip_and_corruption tests::defaults_exceptions_most_specific_wins tests::layers_do_not_leak tests::nixon_diamond_is_undetermined_with_both_proofs tests::gates_are_not_mute tests::queries_answer_with_proofs_and_refuse_to_guess tests::scoring tests::frozen_hash_gate tests::world_command_fields tests::prop_names_parse tests::pronoun_api tests::core_completeness_gate tests::wsd_stems tests::hs_isa_gate
Named outcomes:
- tests::snapshot_round_trip_and_corruption
- tests::defaults_exceptions_most_specific_wins
- tests::layers_do_not_leak
- tests::nixon_diamond_is_undetermined_with_both_proofs
- tests::gates_are_not_mute
- tests::queries_answer_with_proofs_and_refuse_to_guess
- tests::scoring
- tests::frozen_hash_gate
- tests::world_command_fields
- tests::prop_names_parse
- tests::pronoun_api
- tests::core_completeness_gate
- tests::wsd_stems
- tests::hs_isa_gate
Two debug and two release/optimized executions returned zero and identical sorted nonempty named outcomes. Source-generated syntax pairs matched byte-for-byte. No dedicated compiled runtime control for this test-fixture source; production controls are mapped separately.
Finite synthetic examples only. Full ingestion, live workers, real corpus, allocation/I/O fault injection, GPU and full-school behavior remain unexecuted. No universal or executable-approximant assurance. External receipts: octopus-wave29-ontocore-2026-10-31 (native-commands.json, source-test-map.json, outcomes.json, controls.json and numeric exit/log files).
Prior isolation correction
Initial repeated query tests omitted ONTO_PRIOR; query ranking may initialize wsd::global_prior from ambient/default path. Whether that file existed was not established; initial query outcomes are not fully isolated evidence. The affected queries_answer_with_proofs_and_refuse_to_guess name was rerun twice debug and twice release with ONTO_PRIOR asserted absent before/after. All four pass. Native command specs now set this environment. See prior-isolation.json.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/onto/src/tests.rs.groupant/