{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "QuasiLeak-Bench prediction row",
  "description": "One row per document. A span is a verbatim string, located at its first exact occurrence, or a half-open character interval. `predictions` may replace `evidence`.",
  "type": "object",
  "required": ["sample_id"],
  "anyOf": [{"required": ["evidence"]}, {"required": ["predictions"]}],
  "properties": {
    "sample_id": {"type": "string", "minLength": 1},
    "predictions": {"$ref": "#/$defs/groups"},
    "evidence": {"$ref": "#/$defs/groups"}
  },
  "$defs": {
    "groups": {
      "type": "array",
      "items": {
        "type": "object",
        "required": ["attribute", "spans"],
        "properties": {
          "attribute": {
            "enum": ["age", "gender", "nationality", "education", "occupation", "income", "relationship", "health"]
          },
          "spans": {
            "type": "array",
            "minItems": 2,
            "maxItems": 4,
            "items": {
              "oneOf": [
                {"type": "string", "minLength": 1},
                {
                  "type": "object",
                  "required": ["start", "end"],
                  "properties": {
                    "start": {"type": "integer", "minimum": 0},
                    "end": {"type": "integer", "minimum": 1}
                  }
                }
              ]
            }
          },
          "rationale": {"type": "string"}
        }
      }
    }
  }
}