feat(od-ontology): pull namespace inheritance + richer DOLCE classifier from hydrators

[AdaWorldAPI]

What

Phase-2 continued (per REPATRIATION-FRAME.md — frame #12, schema #13,
alignment table #14). Pulled the Layer-1 hydrator content from
lance_graph_ontology::hydrators::{odoo, dolce_odoo} into odoo-rs's
alignment module — three concrete things land:

1. The literal "Odoo inheriting classes" declaration

pub const ODOO_BUNDLE_ID:           OntologyBundleId = OntologyBundleId { graph: 0x0002, version: 1 };
pub const ODOO_INHERITS_FROM_FIBOFND_V1: OntologyBundleId = OntologyBundleId { graph: 0x0007, version: 1 };
pub const ODOO_NAMESPACE_IRI:       &str = "https://ada.world/onto/odoo#";

This IS the namespace-level statement that "Odoo's classes inherit from FIBO
Foundations" — the literal pattern your original complaint named ("instead of
inheriting classes"). The per-class owl:equivalentClass rows in ODOO_SEED
are the instances of this inheritance; these constants are the
declaration.

2. The cascade edge whitelist (compiler-AST symbol-table predicates)

pub const ODOO_EDGE_WHITELIST: &[&str] = &[
    "http://www.w3.org/2000/01/rdf-schema#subClassOf",      // facet subsumption (LOAD-BEARING)
    "http://www.w3.org/2002/07/owl#equivalentClass",        // Layer-2 alignment (LOAD-BEARING)
    "http://www.w3.org/2000/01/rdf-schema#subPropertyOf",   // field-level subsumption
    "http://www.w3.org/2002/07/owl#equivalentProperty",     // field-level alignment
];

Per the compiler-AST frame (#13): these ARE the AST-edge types the
symbol-table follows during transitive resolution. rdfs:subClassOf carries
Odoo's own facet subsumption; owl:equivalentClass is the route into FIBO /
schema.org / QUDT slots (the ODOO_SEED rows resolve through this).

3. The richer DOLCE classifier (merged from TWO disagreeing lance-graph classifiers)

Category	PR #14 (alignment.rs)	Hydrator (this PR)	Merged here
Perdurant suffixes	8 (.move, .move.line, …)	12 (adds .message, .activity, .attendance, .transition, .event, .log, .history, .transaction, .scrap)	15 (union)
Quality suffixes	uom.* only	.tag, .type, .group, .category, .tax	uom.* + .tag, .type, .groups, .group, .category
Abstract suffixes	.tax, fiscal.position, reconcile.model, payment.term, .pricelist*	.template, .config, .policy, .rule, .formula	Both lists merged + .settings
Endurant special-cases	implicit via product.*	product.template only	product.template AND account.account.template
Default	Unknown	Endurant	Unknown (surfaces gaps; consumer decides)

Plus IRI handling — accepts bare names, odoo: prefix, or full
https://ada.world/onto/odoo# IRI.

Two-classifier disagreement (deliberate reconciliations)

lance-graph hosts two disagreeing DOLCE classifiers; the doc table on
dolce_odoo records every reconciliation choice with rationale, including
.tax (kept Abstract; the hydrator says Quality — Phase-3 OGAR-side decides),
hr.* (kept Unknown — the alignment seed and the hydrator disagree, surfacing
the gap rather than picking one), and the default (kept Unknown — hiding gaps
behind an Endurant fallback obscures the surface).

Real findings surfaced via cross-validation

Two silent bugs in the hydrator the cross-validation caught:

1. .config rule had a misleading comment claiming to match *.config.settings
   but it doesn't — res.config.settings ends with .settings, not .config.
   Added .settings to the Abstract list to catch the real Odoo shape.
2. .group rule similarly claimed to match res.groups (plural — the
   canonical Odoo class) but didn't. Added .groups to the Quality list.

The hydrator's own tests didn't surface either bug because they don't test
the actual class names the comments claim to match.

Tests

150+ test results green across the suite: 39 lib + 27 alignment-pin (was
14 in #14 → 13 new pins) + 10 corpus-pin + 13 typed-lift + 10 ogar-parallel +
9 ogar-convergence + 7 bin + others. clippy on my new code clean. fmt-clean.

11 new pins in tests/alignment_pin.rs:

• prefixed-IRI handling (3 forms: bare / odoo: / full namespace)
• strip_odoo_prefix roundtrip
• new Perdurant suffix rules (7 assertions)
• new Quality suffix rules (4 assertions)
• new Abstract suffix rules (2 assertions)
• Endurant special-cases (product.template, account.account.template)
• .tax disagreement marker (documented Abstract choice)
• namespace IRI / bundle ID / FIBO Foundations inheritance
• edge whitelist (2 load-bearing + 2 property variants)
• TTL sources list

Frame status

Builds on the consistent across-PR governance:

• docs(od-ontology): repatriation frame — lance-graph → (OGAR + odoo-rs + ruff) #12 (frame — north star)
• docs(od-ontology): pull SPO schema from lance-graph + name the compiler-AST frame #13 (Phase 1: SPO schema + compiler-AST frame)
• feat(od-ontology): pull alignment symbol-table from lance-graph-callcenter #14 (Phase 2a: alignment symbol-table)
• feat(od-ontology): pull namespace inheritance + richer DOLCE classifier from hydrators #15 (this PR): Phase 2b: namespace inheritance + richer classifier

Remaining Phase 2: the odoo_blueprint/{l1..l15}.rs 14K-LOC pull (bounded per
file). Phase 3 (push empowerments to OGAR): the OwlPivot shape, the
suffix-classifier algorithm, the family_default_style → StyleCluster
inheritance, and resolving the documented disagreements (especially .tax).

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.