Source Data Model
Canonical source rows are the authority layer. Derived indexes, API envelopes, MCP plans, AI orchestration, and visible UI may return only pointers, evidence IDs, receipt pointers, states, and indexability until evidence gates explicitly open.
- Route Family
- Docs
- Robots Policy
- Public route
- Sitemap Inclusion
- included
- Source Gate
- Landing honesty gate
- Receipt Pointer
- none
- Closed Claim
- Documentation only
Authority Boundary
Logoi treats canonical source data as the authoritative layer: canonical source rows are the authority layer. Semantic vector search, sparse lexical search, route JSON, MCP plans, AI orchestration, and UI read surfaces are derived projections. They can point back to canonical records, but they cannot create source rows, evidence rows, claims, receipt exports, review promotions, or sitemap/indexability promotions.
Canonical Store Status
This panel reads the internal canonical store adapter. It reports authority rows and pointer relationships only; derived projections remain downstream mirrors and cannot write, promote, export receipts, or create claims.
Source Of Truth
Canonical fixture rows are the source of truth for identity, route state, pointer state, and blocker state.
Derived, Not Authoritative
API envelopes, route registries, search indexes, JSON-LD, and future agent surfaces may point back to canonical rows. They are not allowed to become source authority or promotion machinery.
src/lib/source/canonicalDataModel.ts | Boundary | Visible Status | Machine State |
|---|---|---|
| Source of truth | Read-only fixture Canonical fixture rows | canonical_source_data |
| Derived projections | Read-only fixture 3 derived projections | derived_index_projection |
| Storage adapter | Disabled No storage adapter | none |
| Runtime calls | False Not callable | false |
| Fixture mutation | False Closed | false |
| Claim creation | False Closed | false |
| Receipt export | False Closed | false |
Current Source-Policy Status
This compact status surface reflects the existing fixture and static policy posture. It does not add source metadata, expose a new API, export receipts, ingest text, promote routes, or open claim-bearing output.
Audited Homer HTML source display is limited to Iliad 1.1-5. The manifest permits website HTML display for that fixture only.
Public API source-text bodies, receipt export, indexing, morphology, translation, lexical range, and etymology stay blocked.
Custody-only Homer snapshots keep checksum and provenance visible while public passage/source route display stays closed.
Custody rows do not unlock public passage display, translation, morphology, etymology, receipt export, API source-text bodies, or indexability.
Hebrew WLC primary-text review and OSHB/MorphHB provider review remain separated metadata-only lanes.
No Hebrew verse text, morphology features, translation, lexical range, etymology, occurrence evidence, receipt export, or index promotion is opened.
Latin Vulgate, classical Latin, English translation, and etymology authority lanes remain visibly separate.
No Latin source text, English translation evidence, morphology, etymology authority content, lexical fallback content, cross-language equivalence, or promoted receipt evidence is opened.
SBLGNT stays the Greek NT primary-text candidate and MorphGNT stays the provider candidate; the lane is manifest-only.
No Greek NT source text, MorphGNT provider text, morphology features, passage rows, occurrence rows, translation/gloss rows, receipt rows, API/MCP payloads, chunks, indexes, or public claim is opened.
Translations are reader aids and reception witnesses. They are not primary evidence for Greek, Hebrew, or Latin claims.
A translation cannot promote source-language, morphology, etymology, lexical, equivalence, corpus, receipt, API/MCP, or indexability gates.
Etymology authorities are citation/audit metadata lanes only; authority selection, root-fallacy warnings, and usage-boundary warnings remain separate.
No origin statement, root or derivation relation, cognacy, historical-development account, lexical range, translation inference, usage inference, or receipt export is opened.
| Output Family | Current Posture |
|---|---|
| translation text | Blocked |
| morphology features | Blocked |
| etymology statements | Blocked |
| lexical range, gloss, usage, frequency, or equivalence claims | Blocked |
| receipt export | Blocked |
| API or MCP runtime | Blocked |
| ingestion, chunking, indexing, sitemap inclusion, and index promotion | Blocked |
Source Activation Boundary
documentation only metadata authority onlyThe source data model can describe activation order, but it is not an activation event. Review rows, manifest rows, projection manifests, and route links stay as metadata authority until a later audited packet explicitly opens a narrower evidence surface.
| Layer | Allowed Now | Cannot Authorize |
|---|---|---|
| Canonical Rows | Identity, state, route, blocker, and provenance pointers. | Evidence claims, public receipts, index promotion, or runtime calls. |
| Manifest Review | Source acquisition review, candidate comparison, checksum/provenance inspection, and next-validator planning. | Source display, ingestion, chunking, search indexes, sitemap promotion, or deployment. |
| Derived Projections | Future manifest inspection for exact-reference, sparse lexical, and semantic vector layers. | Authority creation, claim derivation, storage binding, API execution, or MCP execution. |
| Public And Agent Surfaces | Routes, source IDs, evidence IDs, receipt pointers, state labels, blockers, and review links. | Generated answers, teaching artifacts, exported receipts, or closed source-language claims. |
Hard Stop
A schema row, manifest row, projection manifest, validator pass, or visible route link is not permission to display, ingest, chunk, index, export, call API tools, call MCP tools, generate teaching output, or publish evidence claims.
Fail-Closed Activation Roadmap
Corpus and index activation must move in order. Later projection, runtime, or sitemap layers cannot compensate for skipped source custody, passage, occurrence, morphology, translation, receipt, or chunk gates.
| Surface | State | Boundary |
|---|---|---|
| BM25 sparse search | Blocked | Sparse projection remains a blocked manifest contract with no runtime search, snippets, counts, or dependency binding. |
| Chroma vector search | Blocked | Vector projection remains a blocked manifest contract with no model calls, embeddings, vector store, or semantic endpoint. |
| Search runtime | Blocked | Search may be described as an activation layer only; current study and public routes stay navigational and pointer-only. |
| Public API | Read-only fixture | Current public JSON may expose source IDs, evidence IDs, receipt pointers, indexability, and states only. |
| MCP | Planned, not callable | MCP remains a documented future contract with no server, resource registry, tool execution, or callable transport. |
Source custody
Passage audit
Occurrence and token audit
Morphology provider audit
Translation rights audit
Receipt promotion
Chunk eligibility
Sparse projection
Vector projection
API, MCP, and search runtime
Public indexability
Canonical Rows
These rows are fixture-backed authority records. Their presence proves identity, state, route, and blocker information only; every row keeps claim derivation closed.
| Entity | Fixture | Role | Rows | Claim Derivation |
|---|---|---|---|---|
| source | sources | authority_row | 36 | closed |
| corpus | corpora | authority_row | 6 | closed |
| work | works | authority_row | 1 | closed |
| passage | passages | authority_row | 3 | closed |
| term | terms | authority_row | 19 | closed |
| term_alias | termAliases | authority_row | 2 | closed |
| occurrence | occurrences | authority_row | 1 | closed |
| morphology_analysis | morphologyAnalyses | authority_row | 1 | closed |
| translation_gloss | translationGlosses | authority_row | 1 | closed |
| etymology_claim | etymologyClaims | authority_row | 19 | closed |
| evidence | evidence | authority_row | 11 | closed |
| receipt | receipts | authority_row | 5 | closed |
| journey | journeys | supporting_authority_row | 1 | closed |
| generated_teaching_artifact | generatedTeachingArtifacts | supporting_authority_row | 1 | closed |
| usage_summary | usageSummaries | supporting_authority_row | 1 | closed |
| kwic_row | kwicRows | supporting_authority_row | 1 | closed |
| chart_series | chartSeries | supporting_authority_row | 1 | closed |
| review_gate | reviewGates | review_gate_row | 17 | closed |
| page_registry | pageRegistry | route_boundary_row | 43 | closed |
| source_acquisition_manifest | sourceAcquisitionManifests | supporting_authority_row | 5 | closed |
Derived Projections
Exact-reference lookup, sparse lexical lookup, and semantic vector search are projections from canonical rows. They may return entity IDs, canonical paths, state pointers, and rank metadata; they may not become authority.
Exact reference projection manifest
Derived from source, corpus, work, passage, evidence, receipt, page_registry.
Sparse lexical projection manifest
Derived from source, corpus, work, passage, term, term_alias, occurrence, evidence, receipt.
Semantic vector projection manifest
Derived from source, passage, term, evidence, receipt, review_gate.
Projection Readiness
The exact-reference, sparse lexical, and semantic vector manifests are inspection records only. They describe the checksum, invalidation, and blocker contract a future index build would need; this page does not run search, bind storage, expose an API, or build projection output.
Exact reference projection manifest
Sparse lexical projection manifest
Semantic vector projection manifest
Projection Non-Authority Rules
Future sparse-search or vector-search readiness can be inspected here only as a manifest contract. Projection rows cannot become source, translation, morphology, etymology, lexical, usage, equivalence, teaching, receipt-export, review, or indexability authority.
| Projection | Creates Canonical Rows | Creates Evidence Rows | Creates Claims | Exports Receipts | Promotes Indexability |
|---|---|---|---|---|---|
| exact_reference | false | false | false | false | false |
| sparse_lexical | false | false | false | false | false |
| semantic_vector | false | false | false | false | false |
API, MCP, And AI Boundary
API/MCP/AI surfaces must preserve the evidence gates. The allowed response is pointer-first: source IDs, evidence IDs, receipt pointers, states, and indexability. A route, JSON envelope, or future MCP resource cannot bypass the canonical rows.
| Surface | Current Use | Callable | Required Inputs | Bypass Gates | Export Receipts |
|---|---|---|---|---|---|
| AI orchestration boundary | route_and_pointer_planning | false | source_ids, evidence_ids, receipt_pointer, indexability, states | false | false |
| Public API boundary | read_only_pointer_json | true | source_ids, evidence_ids, receipt_pointer, indexability, states | false | false |
| Future MCP boundary | documentation_only_not_callable | false | source_ids, evidence_ids, receipt_pointer, indexability, states | false | false |
Allowed Output Shape
Public documentation can describe the contract, and read-only machine surfaces can mirror it. They do not widen the runtime API, publish MCP, export receipts, or open source-claim fields.
Allowed Pointer Fields
source_idsevidence_idsreceipt_pointerindexabilitystates
Closed Surfaces
- source text
- translation text
- morphology claims
- etymology claims
- lexical range claims
- usage and frequency claims
- equivalence claims
- generated answers
- public receipt exports
- runtime tool execution
UI Read Surfaces
Visible HTML evidence surface
Requires source, evidence, receipt, page_registry.
JSON-LD evidence mirror
Requires source, evidence, page_registry.
Receipt pointer drawer
Requires receipt, source, evidence.