Text analysis

POST /nlu/v1/agents/{agent_id}/processing/conversation_analysis

Analyzing user sentences in a conversation using the editing model.

Sample input

{
  "text": "Quero retirar R$10 da poupança"
}

Sample Response

{
  "input_text": "Quero retirar R$10 da poupança",
  "intents": [
    {
      "name": "Sacar",
      "score": 0.56
    }
  ],
  "entities": [
    {
      "name": "sys.currency",
      "value": "10",
      "unit":"BRL",
      "text": "r$ 10",
      "start_pos": 14,
      "end_pos": 19,
      "score": 1.0
    }
  ],
  "metadata": {
    "pipeline": "pt-br/conversation-1.0",
    "worker_id": "06023f4e-c4a6-4b8e-aa5d-be0ba0dfa09b",
    "transation_id": "6122d3f3-f7d2-476d-b8fb-250600c2d7e7",
    "response_time": 12.190378904342651
  }
}

Parameters

• agent_id – Agent ID

• name – Published model’s name

Status Codes

• 200 OK – Success

• 404 Not Found – Agent or version not found

• 500 Internal Server Error – Internal error

Response JSON Object

• input_text (string) – Sentence sent.

• intents (List[Intent]) – Detected intentions.

• entities (List[Entity]) – List of entities found.

• metadata (Dict[Metadata]) – Metadatas.

Intent Object

Response JSON Object

• name (string) – Intent name.

• score (float) – Confidence score (0.0 - 1.0).

Entity Object

Response JSON Object

• name (string) – Entity type name.

• score (float) – Confidence score (0.0 - 1.0).

• text (string) – Entity text.

• value (string) – Normalized value.

• unit (string) – Entity measurement unit.

• start_pos (integer) – Starting position (characters).

• end_pos (integer) – Ending position (characters).

Metadata Object

Response JSON Object

• pipeline (string) – Type of pipeline used.

• response_time (string) – Analysis time.