Einstein Dataset

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://github.com/api-search/salesforce-einstein/json-schema/salesforce-einstein-dataset-schema.json",
  "title": "Einstein Dataset",
  "description": "Represents a dataset used for training Einstein Vision or Language models, including labels, examples, and metadata.",
  "type": "object",
  "properties": {
    "id": {
      "type": "integer",
      "description": "Unique dataset identifier."
    },
    "name": {
      "type": "string",
      "description": "Name of the dataset."
    },
    "type": {
      "type": "string",
      "description": "Type of the dataset.",
      "enum": [
        "image",
        "image-multi-label",
        "image-detection",
        "text-sentiment",
        "text-intent"
      ]
    },
    "createdAt": {
      "type": "string",
      "format": "date-time",
      "description": "Timestamp when the dataset was created."
    },
    "updatedAt": {
      "type": "string",
      "format": "date-time",
      "description": "Timestamp when the dataset was last updated."
    },
    "totalExamples": {
      "type": "integer",
      "description": "Total number of examples in the dataset.",
      "minimum": 0
    },
    "totalLabels": {
      "type": "integer",
      "description": "Total number of labels in the dataset.",
      "minimum": 0
    },
    "available": {
      "type": "boolean",
      "description": "Whether the dataset is available for training and predictions."
    },
    "statusMsg": {
      "type": "string",
      "description": "Status message describing the dataset's current state."
    },
    "labelSummary": {
      "type": "object",
      "description": "Summary of labels in the dataset.",
      "properties": {
        "labels": {
          "type": "array",
          "items": {
            "$ref": "#/$defs/Label"
          }
        }
      }
    },
    "object": {
      "type": "string",
      "description": "Object type identifier.",
      "const": "dataset"
    }
  },
  "required": ["id", "name", "type"],
  "$defs": {
    "Label": {
      "type": "object",
      "description": "A label within a dataset used to classify examples.",
      "properties": {
        "id": {
          "type": "integer",
          "description": "Unique label identifier."
        },
        "datasetId": {
          "type": "integer",
          "description": "ID of the dataset this label belongs to."
        },
        "name": {
          "type": "string",
          "description": "Name of the label."
        },
        "numExamples": {
          "type": "integer",
          "description": "Number of examples with this label.",
          "minimum": 0
        }
      },
      "required": ["id", "name"]
    }
  }
}