{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://agentpaywire.com/schemas/dataset-card.schema.json",
  "title": "Agent Paywire dataset card",
  "description": "One training or evaluation dataset. English only. Our metadata is CC BY 4.0. The dataset itself keeps its own license.",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "id",
    "name",
    "publisher",
    "url",
    "license",
    "license_verified",
    "license_url",
    "approx_size",
    "size_bytes",
    "format",
    "access",
    "relevance",
    "mirror_allowed",
    "verified_at",
    "http_status",
    "archive_url",
    "data_period",
    "sample_url",
    "notes"
  ],
  "properties": {
    "id": {
      "type": "string",
      "minLength": 1
    },
    "name": {
      "type": "string",
      "minLength": 1
    },
    "publisher": {
      "type": "string",
      "minLength": 1
    },
    "url": {
      "type": "string",
      "format": "uri"
    },
    "license": {
      "anyOf": [
        {
          "type": "string",
          "minLength": 1
        },
        {
          "type": "null"
        }
      ]
    },
    "license_verified": {
      "type": "boolean"
    },
    "license_url": {
      "anyOf": [
        {
          "type": "string",
          "format": "uri"
        },
        {
          "type": "null"
        }
      ]
    },
    "approx_size": {
      "type": "string",
      "minLength": 1
    },
    "size_bytes": {
      "anyOf": [
        {
          "type": "integer",
          "minimum": 0
        },
        {
          "type": "null"
        }
      ]
    },
    "format": {
      "type": "string",
      "minLength": 1
    },
    "access": {
      "enum": [
        "open",
        "login",
        "terms-acceptance"
      ]
    },
    "relevance": {
      "type": "array",
      "minItems": 1,
      "items": {
        "type": "string",
        "minLength": 1
      }
    },
    "mirror_allowed": {
      "type": "boolean"
    },
    "verified_at": {
      "type": "string",
      "pattern": "^\\d{4}-\\d{2}-\\d{2}$"
    },
    "http_status": {
      "anyOf": [
        {
          "type": "integer",
          "minimum": 100,
          "maximum": 599
        },
        {
          "type": "null"
        }
      ]
    },
    "archive_url": {
      "anyOf": [
        {
          "type": "string",
          "format": "uri"
        },
        {
          "type": "null"
        }
      ]
    },
    "data_period": {
      "anyOf": [
        {
          "type": "string",
          "minLength": 1
        },
        {
          "type": "null"
        }
      ]
    },
    "sample_url": {
      "anyOf": [
        {
          "type": "string",
          "format": "uri"
        },
        {
          "type": "null"
        }
      ]
    },
    "notes": {
      "type": "string",
      "minLength": 1
    }
  }
}