v11

latestSwagger 2.0raw.githubusercontent.com2022-10-03168148314.0 KB

Suggests example utterances that would improve the accuracy of the intent model in a version of the application.

get/apps/{appId}/versions/{versionId}/intents/{intentId}/suggest

Path parameters

appIdstring uuid required

The application ID.

versionIdstring required

The version ID.

intentIdstring uuid required

The intent classifier ID.

Query parameters

takeinteger

The number of entries to return. Maximum page size is 500. Default is 100.

enableNestedChildrenboolean

Toggles nested/flat format

Response

If there's no trained intent model, nothing is returned in the response. If there's a trained model and active learning finds any relevant queries, they are returned with the intent model predictions. If there's a trained model but active learning didn't find any relevant queries, an empty list is returned in the response.

textstring

The utterance. For example, "What's the weather like in seattle?"

tokenizedTextstring[]

The tokenized utterance.

Example response

[
  {
    "entityPredictions": [
      {
        "children": [
          {
            "entityName": "child",
            "startTokenIndex": 0,
            "endTokenIndex": 0,
            "phrase": "string"
          }
        ]
      }
    ]
  }
]