{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://waterpark.dkrz.de/blobmap/manifest-v2.schema.json",
  "title": "blobmap manifest",
  "description": "Blob definitions for one scope. Pure definition: this changes only when the set of blobs changes, never on reads, writes, tiering or restores.",
  "type": "object",
  "required": [
    "schema_version",
    "scope",
    "epoch",
    "hot_always",
    "blobs"
  ],
  "additionalProperties": false,
  "properties": {
    "schema_version": {
      "description": "Format version. Consumers must reject a version they do not know rather than guess.",
      "type": "integer",
      "const": 2
    },
    "scope": {
      "description": "Prefix these definitions apply to, relative to the data store.",
      "type": "string"
    },
    "epoch": {
      "description": "Bumped whenever the definitions change, so a resolver can detect staleness without diffing.",
      "type": "integer",
      "minimum": 1
    },
    "generated_at": {
      "description": "ISO 8601 UTC timestamp of the producing run.",
      "type": "string"
    },
    "generated_by": {
      "description": "Package and version that produced this.",
      "type": "string"
    },
    "policy": {
      "description": "Thresholds used to produce this cut, recorded so a later run can tell whether they changed.",
      "type": "object",
      "additionalProperties": {
        "type": "integer"
      }
    },
    "hot_always": {
      "description": "Glob patterns that are never archivable, whatever the blobs say. Metadata objects and dimension coordinates.",
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "blobs": {
      "description": "Blob definitions. A list of rules, so its length tracks cut decisions rather than object count.",
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "id",
          "prefixes",
          "bucket"
        ],
        "additionalProperties": false,
        "properties": {
          "id": {
            "description": "Join key against tier state, and through it against tape addresses.",
            "type": "string",
            "pattern": "^[a-z0-9_]+$"
          },
          "prefixes": {
            "description": "Key prefixes this blob claims, relative to the scope. More than one when small arrays were coalesced.",
            "type": "array",
            "minItems": 1,
            "items": {
              "type": "string"
            }
          },
          "bucket": {
            "description": "Arithmetic for cutting inside an array. Null means one bucket, so the resolved id is always id_0.",
            "type": [
              "object",
              "null"
            ],
            "required": [
              "index",
              "width",
              "key_encoding"
            ],
            "additionalProperties": false,
            "properties": {
              "index": {
                "description": "Which segment after the prefix holds the chunk index.",
                "type": "integer",
                "minimum": 0
              },
              "width": {
                "description": "Objects per blob along that dimension.",
                "type": "integer",
                "minimum": 1
              },
              "key_encoding": {
                "description": "How to parse the index out of a key.",
                "enum": [
                  "v3_slash",
                  "v2_slash",
                  "v2_flat"
                ]
              }
            }
          }
        }
      }
    },
    "provenance": {
      "description": "Measured numbers, for debugging only. These go stale while the manifest is untouched, so a resolver must never read them.",
      "type": "object"
    }
  }
}