{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "https://get.aiware.com/schemas/v1.1/master.json",
  "title": "vtn-standard",
  "description": "Standard engine output at Veritone",
  "type": "object",
  "definitions": {
    "header": {
      "type": "object",
      "properties": {
        "schemaId": {
          "type": "string",
          "description": "vtn-standard.master.schema.json or subschema.schema.json (provided by Veritone)",
          "example": "https://docs.veritone.com/schemas/vtn-standard/master.json"
        },
        "sourceEngineId": {
          "type": "string",
          "description": "<GUID> of the engine (provided by Veritone)",
          "example": "fe261482-af8d-4709-aefb-ff1c19b61eb9"
        },
        "sourceEngineName": {
          "type": "string",
          "description": "Engine name used to generate output (provided by Veritone)",
          "example": "capio"
        },
        "taskPayload": {
          "type": "object",
          "description": "JSON key/value task payload describing the tasks used to start the engine run (provided by Veritone). Task payload will be a combination of attributes found on https://docs.veritone.com/#/developer/adapters/quick-start/step-2?id=constructing-a-batch-pull-adapter and any default custom fields (https://docs.veritone.com/#/developer/engines/custom-fields/?id=custom-fields)",
          "example": {
            "detectionThreshold": 0.7,
            "processingFPS": 30
          }
        },
        "taskId": {
          "type": "string",
          "description": "<GUID> of the task (provided by Veritone)",
          "example": "6b9cd86e-de5c-4795-b605-a9b3b29fa3cf"
        },
        "generatedDateUTC": {
          "type": "string",
          "description": "DateTime (ISO8601) of when the engine result was generated (provided by Veritone if not included)",
          "format": "date-time",
          "example": "2017-12-08T17:19:02Z"
        },
        "externalSourceId": {
          "type": "string",
          "description": "Vendor specific referenceId to this engine result (optional)",
          "example": "1"
        },
        "language": {
          "$ref": "#/definitions/language"
        },
        "validationContracts": {
          "type": "array",
          "minItems": 1,
          "items": [
            {
              "$ref": "#/definitions/validationContracts"
            }
          ]
        }
      },
      "required": [
        "validationContracts"
      ]
    },
    "confidence": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Confidence of the result expressed as a range of 0.0 to 1.0"
    },
    "point": {
      "type": "object",
      "properties": {
        "x": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "A floating point describing the location via percentage on the x axis (1.0 would be right of the axis)"
        },
        "y": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "A floating point describing the location via percentage on the y axis (1.0 would be the bottom of the axis)"
        }
      },
      "required": [
        "x",
        "y"
      ],
      "additionalProperties": false,
      "description": "A point describing the 2d coordinates from the top left corner of the source."
    },
    "boundingPoly": {
      "type": "array",
      "items": {
        "$ref": "#/definitions/point"
      }
    },
    "objectCategory": {
      "type": "array",
      "minItems": 1,
      "items": {
        "type": "object",
        "properties": {
          "class": {
            "type": "string",
            "description": "The class of the object recognized",
            "example": "For example, if the label of an object was dog, the class of the object could be animal"
          },
          "@id": {
            "type": "string",
            "description": "Private id variable used by the engine to refer to its internal dataset",
            "example": "1"
          },
          "confidence": {
            "$ref": "#/definitions/confidence"
          }
        },
        "required": [
          "class"
        ],
        "additionalProperties": false
      },
      "description": "Optional categories for object recognition. Used for referencing external taxonomies."
    },
    "objectResultType": {
      "type": "string",
      "enum": [
        "object",
        "face",
        "licensePlate",
        "logo",
        "fingerprint",
        "speaker",
        "sound",
        "concept",
        "keyword",
        "text",
        "namedEntity",
        "barcode",
        "anomaly"
      ]
    },
    "objectResult": {
      "type": "object",
      "properties": {
        "boundingPoly": {
          "$ref": "#/definitions/boundingPoly"
        },
        "confidence": {
          "$ref": "#/definitions/confidence"
        },
        "entityId": {
          "type": "string"
        },
        "entityIdentifierId": {
          "type": "string",
          "description": "The unique ID of recognized entity identifier (for example, a specific image). This information is not always available and is therefore optional."
        },
        "libraryId": {
          "type": "string"
        },
        "label": {
          "type": "string"
        },
        "objectCategory": {
          "$ref": "#/definitions/objectCategory"
        },
        "uri": {
          "type": "string"
        },
        "type": {
          "$ref": "#/definitions/objectResultType"
        },
        "sentiment": {
          "$ref": "#/definitions/sentiment"
        },
        "emotions": {
          "$ref": "#/definitions/emotions"
        },
        "text": {
          "$ref": "#/definitions/objectText"
        },
        "page": {
          "$ref": "#/definitions/page"
        },
        "paragraph": {
          "$ref": "#/definitions/paragraph"
        },
        "sentence": {
          "$ref": "#/definitions/sentence"
        }
      },
      "required": [
        "type"
      ],
      "additionalProperties": false
    },
    "objectText": {
      "description": "The extracted textual content of a found object.",
      "type": "string"
    },
    "word": {
      "description": "Transcript. JSON utterance (all word edges between 2 time nodes). Array of n objects describing each alternative word",
      "type": "string"
    },
    "words": {
      "allOf": [
        {
          "type": "array",
          "minItems": 1,
          "items": {
            "type": "object",
            "properties": {
              "word": {
                "$ref": "#/definitions/word"
              },
              "confidence": {
                "$ref": "#/definitions/confidence"
              },
              "utteranceLength": {
                "$ref": "#/definitions/utteranceLength"
              },
              "bestPath": {
                "type": "boolean"
              }
            },
            "required": [
              "word"
            ],
            "additionalProperties": false
          }
        },
        {
          "type": "array",
          "if": {
            "minItems": 2
          },
          "then": {
            "items": {
              "required": [
                "confidence",
                "utteranceLength"
              ]
            },
            "uniqueItems": true,
            "requireBestPath": true
          }
        }
      ]
    },
    "utteranceLength": {
      "description": "Number of consecutive time-slices the utterance spans. example: of->thrones----->of->their-->own-> utteranceLength: thrones: 2; their,own: 1",
      "type": "integer",
      "minimum": 1,
      "example": 2
    },
    "language": {
      "description": "Language Identification. Format: BCP-47 https://tools.ietf.org/rfc/bcp/bcp47.txt",
      "type": "string",
      "examples": [
        "en",
        "en-US"
      ]
    },
    "seriesItem": {
      "type": "object",
      "properties": {
        "startTimeMs": {
          "type": "integer",
          "minimum": 0,
          "description": "The start time relative to the start of the media asset in milliseconds.",
          "example": 0
        },
        "stopTimeMs": {
          "type": "integer",
          "minimum": 0,
          "description": "The stop time relative to the start of the media asset in milliseconds.",
          "example": 3000
        },
        "language": {
          "$ref": "#/definitions/language"
        },
        "object": {
          "$ref": "#/definitions/objectResult"
        },
        "words": {
          "$ref": "#/definitions/words"
        }
      },
      "required": [
        "startTimeMs",
        "stopTimeMs"
      ],
      "additionalProperties": false
    },
    "series": {
      "type": "array",
      "items": {
        "$ref": "#/definitions/seriesItem"
      }
    },
    "media": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "assetId": {
            "type": "string",
            "description": "The ID of the asset output by the engine"
          },
          "language": {
            "$ref": "#/definitions/language"
          }
        },
        "required": [
          "assetId"
        ]
      }
    },
    "validationContracts": {
      "type": "string",
      "enum": [
        "concept",
        "entity",
        "keyword",
        "language",
        "media-translated",
        "object",
        "sentiment",
        "summary",
        "text",
        "transcript",
        "anomaly"
      ]
    },
    "page": {
      "type": "integer",
      "description": "The page on which the text or insight appears"
    },
    "paragraph": {
      "type": "integer",
      "description": "The paragraph in which the text or insight appears. If page is provided, it should be the paragraph number on the page (restart at 1 for each new page). Otherwise it may be an increasing number for the whole document."
    },
    "sentence": {
      "type": "integer",
      "description": "The sentence in which the text or insight appears. If paragraph is provided, it should be the sentence number within the paragraph (restart at 1 for each new paragraph). Otherwise it may be an increasing number for the whole document."
    },
    "noDocumentIndexing": {
      "description": "Disable document indexing (page/paragraph/sentence) for the object",
      "properties": {
        "page": false,
        "paragraph": false,
        "sentence": false
      }
    },
    "sentiment": {
      "type": "object",
      "properties": {
        "positiveValue": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "Positive sentiment rating"
        },
        "positiveConfidence": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "Confidence of positive sentiment rating"
        },
        "negativeValue": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "Negative sentiment rating"
        },
        "negativeConfidence": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "description": "Confidence of negative sentiment rating"
        }
      },
      "dependencies": {
        "positiveConfidence": {
          "required": [
            "positiveValue"
          ]
        },
        "negativeConfidence": {
          "required": [
            "negativeValue"
          ]
        }
      }
    },
    "emotions": {
      "type": "array",
      "items": {
        "properties": {
          "emotion": {
            "type": "string"
          },
          "emotionValue": {
            "type": "number",
            "minimum": 0,
            "maximum": 1
          },
          "emotionConfidence": {
            "type": "number",
            "minimum": 0,
            "maximum": 1
          }
        },
        "required": [
          "emotion"
        ]
      }
    }
  }
}