{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://sediment.so/schemas/derived-artifacts/attribution/v1.json",
  "title": "Attribution",
  "description": "One completion-to-commit attribution.",
  "type": "object",
  "properties": {
    "org_id": {
      "type": "string",
      "description": "The deployment's tenant id, normalized. Bound to the server."
    },
    "repo": {
      "type": "string",
      "description": "`owner/repo`, lowercased. Empty string when the sender omitted it."
    },
    "commit_sha": {
      "type": "string",
      "description": "Full-length commit sha, lowercased."
    },
    "file_path": {
      "type": "string",
      "description": "The changed file this attribution is keyed on."
    },
    "inference_call_id": {
      "type": "string",
      "description": "Id of the inference call this record belongs to."
    },
    "session_id": {
      "type": "string",
      "description": "The coding-agent session this belongs to (ADR 0002)."
    },
    "similarity_score": {
      "type": "number",
      "description": "Token-overlap similarity in [0, 1]. Under `git_notes` attribution it only ranks completions within a proven session; under `jaccard` it is the evidence, and it discounts confidence."
    },
    "attribution_source": {
      "enum": [
        "git_notes",
        "jaccard"
      ],
      "type": "string",
      "description": "How the underlying attribution was established."
    },
    "provenance": {
      "$ref": "#/$defs/Provenance",
      "description": "The structured provenance for this artifact."
    }
  },
  "additionalProperties": false,
  "required": [
    "org_id",
    "repo",
    "commit_sha",
    "file_path",
    "inference_call_id",
    "session_id",
    "similarity_score",
    "attribution_source",
    "provenance"
  ],
  "$defs": {
    "Provenance": {
      "type": "object",
      "properties": {
        "policy_version": {
          "type": "string",
          "description": "The derivation implementation policy version."
        },
        "quarantine_revision": {
          "type": "integer",
          "description": "The integer quarantine-log high-water mark used for the derivation."
        },
        "policy_digest": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ],
          "description": "SHA-256 of the fully resolved derivation policy when available."
        }
      },
      "additionalProperties": false,
      "required": [
        "policy_version",
        "quarantine_revision",
        "policy_digest"
      ]
    }
  }
}