{
  "schema_version": "1.3.0",
  "slug": "what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices",
  "title": "What embedding metadata invariants must survive schema evolution in vector indices?",
  "date_of_record": "2026 October 20",
  "scope": "* **Applies when:** The vector index contains vectors from multiple metadata schema versions, and the session's decisions depend on accurate metadata interpretation. * **Does not apply when:** The metadata schema is immutable, or the index ",
  "projections": {
    "unified_markdown_uri": "https://corpusfetch.com/c/what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices.md",
    "method_python_slice_uri": "https://corpusfetch.com/c/what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices.py",
    "remnant_warrant_slice_uri": "https://corpusfetch.com/c/what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices.warrant.json"
  },
  "token_profile": {
    "full_markdown_tokens": 1058,
    "python_slice_tokens": 279,
    "warrant_slice_tokens": 220
  },
  "remnant_summary": {
    "core_distinction": "The separation between **vector similarity** (the vector's mathematical proximity to the query) and **metadata interpretation** (the meaning of the fields attached to the vector under its schema version). Relying on similarity creates the **Metadata Version Confusion**: the session retrieves a vector whose metadata was stored under schema v1, but interprets ...",
    "what_was_observed": "In evolving vector index architectures, failing to account for metadata schema versions produces three primary failure modes: 1. **Type Confusion:** Schema v1 stored `timestamp` as a Unix epoch integer. Schema v2 stores it as an ISO 8601 string. The session reads an integer, treats it as a string, and downstream parsing fails silently. 2. **Semantic Shift:** Schema v1's `priority` field was 1-5. Schema v2 redefines i...",
    "condition_of_validity": "* **Applies when:** The vector index contains vectors from multiple metadata schema versions, and the session's decisions depend on accurate metadata interpretation. * **Does not apply when:** The metadata schema is immutable, or the index supports only one schema version at a time (via atomic migrations).",
    "decision_boundary": "1. **Does the index contain multiple metadata schema versions?** If no, standard interpretation applies. If yes, proceed. 2. **Is the `metadata_schema_version_signature` present for each vector?** If no, the session cannot determine which schema to apply. It must treat the metadata as `UNVERIFIED`. 3. **Is there a `field_mapping_matrix` that allows translati...",
    "operational_effect": "1. **Version-Aware Interpretation:** Prevents silent metadata misinterpretation across schema versions. 2. **Consistent Cross-Version Queries:** Enables queries that span versions by translating to a common semantics. 3. **Field Absence Discipline:** Uses signed defaults for missing fields, preventing cross-tenant contamination."
  },
  "matched_failure_symptoms": [
    "SCHEMA_DRIFT_OR_COERCION_CRASH"
  ],
  "composition_edges": {
    "primary_domain": "epistemic_provenance_and_federation_routing",
    "upstream_precondition_residuals": [
      "https://corpusfetch.com/c/according-to-what-precedence-hierarchy-must-an-instance-resolve-contradictory-directives-across-system-developer-and-user-tiers.md",
      "https://corpusfetch.com/c/how-should-an-instance-decide-that-a-stored-conclusion-has-gone-stale-without-a-new-observation.md"
    ],
    "downstream_consequence_residuals": [
      "https://corpusfetch.com/c/by-what-checkpoint-and-compaction-boundaries-must-an-instance-externalize-working-state-to-prevent-context-window-drift-and-truncation-amnesia.md",
      "https://corpusfetch.com/c/across-what-additive-schema-and-unknown-field-preservation-boundaries-must-an-instance-evolve-wire-contracts-to-prevent-rolling-deploy-deserialization-crashes.md"
    ],
    "co_occurring_sibling_residuals": [
      "https://corpusfetch.com/c/how-can-an-instance-capture-implicit-user-preferences-without-storing-private-data.md",
      "https://corpusfetch.com/c/how-can-an-instance-compress-long-reasoning-traces-into-binding-conclusions.md"
    ]
  },
  "empirical_verification_warrant": {
    "status": "AST_AND_INVARIANT_VERIFIED",
    "syntax_valid": true,
    "code_sha256": "413e373313dbda362cd9003104c80a73edff01dc84677b335d478a42ca265669",
    "verified_classes": [
      "MetadataVerdict",
      "VectorMetadata",
      "SchemaContext",
      "MetadataPolicy",
      "MetadataManifest",
      "EmbeddingMetadataGuard"
    ],
    "verified_functions": [
      "_hash",
      "evaluate_metadata"
    ],
    "verified_functions_count": 2,
    "defensive_guards_count": 0,
    "raised_exceptions": [],
    "sample_fail_closed_invariants": [],
    "stdlib_dependencies": [
      "dataclasses",
      "enum",
      "hashlib",
      "json"
    ],
    "observation_mode": "EMPIRICAL_BOUNDARY_CAPTURE",
    "zero_external_deps": true
  }
}