feat(NTO/WorkOrder): add WorkOrder namespace from AdaWorldAPI/WoA transcode (15 entities + 12 verbs)

[AdaWorldAPI]

Summary

Adds the WorkOrder namespace under NTO/, transcoded from
AdaWorldAPI/WoA's SQLAlchemy
models.py. 15 entity TTLs + 12 distinct verb TTLs, all OGIT-shaped,
all dcterms:source-annotated back to the originating Python class.

This is purely additive to the fork. WorkOrder-specific terms live
in this fork only; the upstream almatoai/OGIT repository is not
affected. The session that produced this PR ran a 12-agent specialist
ensemble (workspace-primer, container-architect, family-codec-smith,
bus-compiler, ripple-architect, host-glove-designer, scenario-world,
mirror-kernel-synthesist, truth-architect, palette-engineer,
certification-officer, perspective-weaver) plus an integration-lead
meta-agent that validated end-to-end before this PR opened.

What lands

Entities (15) under NTO/WorkOrder/entities/:

• Tenant (42 triples) — multi-tenant root.
• Customer (96 triples) — Kunde + supplier roles.
• Order (80 triples) — workorder / offer / order / invoice / credit
  via docType discriminator (the source WorkOrder.doc_type enum is
  preserved as a single class with the variant tracked in the
  description and docType predicate).
• Article (58 triples) — Artikelstamm.
• Position (40), Activity (29), Picture (29), HistoryEntry (30).
• User (75), LogbookEntry (93), NumberSequence (33), Setting (37).
• CustomerPortalUser (35), PasswordEntry (75), TimeSheet (61).

Verbs (12) under NTO/WorkOrder/verbs/, anti-flattening preserved
(every relationship is a distinct verb, not a generic related):

• Issued — Customer → Order
• HasPosition / HasActivity / HasPicture / HasHistory — Order → children
• Assigned — User ↔ Order
• RefersToArticle — Position → Article
• AccessesPortal — CustomerPortalUser → Customer
• OwnsPasswords — Customer → PasswordEntry
• LogsTime — User → TimeSheet
• Drives — User → LogbookEntry
• BelongsToTenant — multi-source → Tenant via 6 from-to pairs

Shape and conventions

Each entity follows the NTO/Network/entities/IPAddress.ttl reference
shape: a rdfs:Class ; rdfs:subClassOf ogit:Entity ; ogit:scope "NTO" ; ogit:parent ogit:Node with mandatory / optional / indexed attribute
lists and ogit:allowed [ ogit:relates | ogit:belongs ... ] clauses.

Each verb is a rdfs:Class ; rdfs:subClassOf ogit:Verb with an
ogit:from-to list per spec.

Namespace IRI: ogit.WorkOrder: <http://www.purl.org/ogit/WorkOrder/>.
Field-scoped predicates camelCased (passwordHash, mustChangePw,
settingKey/settingValue, lastLogin, passwortEnc, notizenEnc,
privatAnteil, startKm, endeKm, timerStart, abgerechnet, …).
Global predicates reused where they fit (ogit:id, ogit:name,
ogit:status, ogit:email, ogit:created-at).

dcterms:source = "AdaWorldAPI/WoA/models.py:<ClassName>" is set on
every class, plus per-attribute models.py:<ClassName>.<field>
provenance pointers where new predicates were introduced.

Validation

• All 27 TTL files parse cleanly via pyoxigraph 0.5.8 TURTLE
  parser. Zero syntax errors. 989 total triples.
• Round-trip verified through
  lance-graph-ontology::OntologyRegistry::hydrate_once_sync
  on the consumer branch claude/create-graph-ontology-crate-gkuJG
  in AdaWorldAPI/lance-graph — see new test
  hydrate_workorder_namespace_from_real_ogit in
  crates/lance-graph-ontology/tests/hydrate_real_ogit.rs. All 15
  entities resolve through the registry; WoaBridge scope-locks the
  namespace correctly.
• Carrier-method doctrine and SPO-1 federation (Option B) decisions
  documented in the consumer repo's
  .claude/DECISION_SPO_ARIGRAPH.md.

Test plan

• All 27 TTL files validated via pyoxigraph (no syntax errors).
• WorkOrder namespace hydrates through lance-graph-ontology
  registry against the real OGIT fork.
• WoaBridge::new(registry) constructs and g_lock() matches
  registry.namespace_id("WorkOrder").
• Every verb is a distinct relationship (anti-flattening
  preserved) — no collapse into a generic related predicate.
• dcterms:source provenance present on every class.
• Reviewer to confirm no overlap with planned upstream OGIT
  namespaces (almatoai/OGIT); WorkOrder is fork-only by design.

Out of scope

• dcterms:source literal values are captured by the consumer's TTL
  parser but not yet threaded through to
  MappingProposal::source_uri. Tracked as TTL-PROBE-5 in the
  consumer repo's TECH_DEBT — does not affect this PR's TTL
  correctness.
• PARSER-1 entropy row (4–6 parsers across the consumer workspace)
  is by design (cold-path vs hot-path); consolidation is post-migration
  work, not a blocker for this fork PR.

https://claude.ai/code/session_01WevBiZ3jzVocu8fBpTY8sq

---

Generated by Claude Code

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3871d37cb0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Declare WorkOrder verbs as object properties

All newly added WorkOrder verbs use a rdfs:Class; rdfs:subClassOf ogit:Verb, while the repo's established verb shape is a owl:ObjectProperty; rdfs:subPropertyOf ogit:Verb across the existing NTO/SGO verb files. Consumers that discover relationships using the existing ObjectProperty/subPropertyOf convention will skip these 12 WorkOrder relationships even though the Turtle syntax parses cleanly.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Define the referenced WorkOrder attributes

Several attributes in these new entity attribute lists are never declared as rdfs:Property anywhere in the added namespace: for this entity geraet, logbuch, and intern are unresolved, and ogit:created-at is also not an existing global attribute (the repo uses ogit:createdAt). The same unresolved-reference pattern appears in HistoryEntry, Position, and Picture, so ontology consumers can load the entities but cannot resolve metadata/provenance for those attributes.

Useful? React with 👍 / 👎.