Tovie Data Agent public API (1.0.0)

Download OpenAPI specification:Download

ProjectsPublic

Data Agent project operations API.

Get project info

Obtain information on the current knowledge base project.

Authorizations:

bearerAuth

Responses

Response samples

200

Content type

application/json

{"id": 0,
"name": "string",
"status": "CREATED",
"resources": {"llmModels": ["string"
]
},
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z"
}

QueriesPublic

Send requests to search for chunks and generate responses (without considering chat history).

Retrieve chunks

Retrieve chunks relevant to the user’s query from the knowledge base.

Authorizations:

bearerAuth

Request Body schema: application/json
required

query required	string Text of the user’s query.
	Array of objects (HistoryRecord) Dialogue history. Entries are displayed in reverse chronological order (from latest to earliest).
	object (RetrievingSettings) Chunk search settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"history": [{"content": "string",
"role": "user"
}
],
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
}
}
}

Response samples

200
400

Content type

application/json

{"chunks": [{"score": 0.1,
"content": "string",
"docId": "string",
"metadata": {"sourcePath": "string",
"sourceUrl": "string",
"segment": "string"
}
}
]
}

Generate response

Synchronous request to generate a response to the user’s query.

Please note that request processing may take a significant amount of time. Ensure that the connection timeout set in your HTTP client is more than 1 minute.

Authorizations:

bearerAuth

Request Body schema: application/json
required

query required	string Text of the user’s query.
	Array of objects (HistoryRecord) Dialogue history. Entries are displayed in reverse chronological order (from latest to earliest).
	object (RagSettings) Query processing settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"history": [{"content": "string",
"role": "user"
}
],
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Response samples

200
400

Content type

application/json

{"id": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Generate response (asynchronous request)

Asynchronous request to generate a response to the user’s query.

The result can be obtained via GET /api/knowledge-hub/query/{queryId} endpoint, where queryId is the request identifier received in the current response.

Authorizations:

bearerAuth

Request Body schema: application/json
required

query required	string Text of the user’s query.
	Array of objects (HistoryRecord) Dialogue history. Entries are displayed in reverse chronological order (from latest to earliest).
	object (RagSettings) Query processing settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"history": [{"content": "string",
"role": "user"
}
],
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Response samples

200
400

Content type

application/json

{"id": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Processing status of the response generation request

Get the current processing status of the response generation request. Long polling is used if the waitTimeSeconds parameter is specified.

Authorizations:

bearerAuth

path Parameters

queryId

required

integer <int64> (LongId)

Identifier of the response generation request.

query Parameters

waitTimeSeconds

integer <int32> [ 0 .. 30 ]

Default: 3

HTTP request timeout. Used in long polling.

Responses

Response samples

200

Content type

application/json

{"id": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Cancel response generation request

Authorizations:

bearerAuth

path Parameters

queryId

required

integer <int64> (LongId)

Identifier of the response generation request.

Responses

Response samples

200

Content type

application/json

{"id": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

ChatsPublic

Send requests to search for chunks and generate responses (within a user chat).

Chat creation

Authorizations:

bearerAuth

Request Body schema: application/json
required

name	string Name of the user’s chat.
	object (RagSettings) Query processing settings.

Responses

Request samples

Payload

Content type

application/json

{"name": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Response samples

200
400

Content type

application/json

{"id": 0,
"name": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

User chat information

Authorizations:

bearerAuth

path Parameters

chatId

required

integer <int64> (LongId)

Identifier of the chat in the knowledge base project.

Responses

Response samples

200
400

Content type

application/json

{"id": 0,
"name": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Retrieve chunks

Retrieve chunks from the knowledge base that are relevant to the user’s query within the chat.

Authorizations:

bearerAuth

path Parameters

chatId

required

integer <int64> (LongId)

Identifier of the chat in the knowledge base project.

Request Body schema: application/json
required

query required	string Text of the user’s query.
	object (RetrievingSettings) Chunk search settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
}
}
}

Response samples

200
400

Content type

application/json

{"chunks": [{"score": 0.1,
"content": "string",
"docId": "string",
"metadata": {"sourcePath": "string",
"sourceUrl": "string",
"segment": "string"
}
}
]
}

Generate response

Synchronous request to generate a response to the user’s query. The chat message history is taken into account.

Please note that request processing may take a significant amount of time. Ensure that the connection timeout set in your HTTP client is more than 1 minute.

Authorizations:

bearerAuth

path Parameters

chatId

required

integer <int64> (LongId)

Identifier of the chat in the knowledge base project.

Request Body schema: application/json
required

query required	string Text of the user’s query.
	object (RagSettings) Query processing settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Response samples

200
400

Content type

application/json

{"id": 0,
"chatId": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Generate response (asynchronous request)

Asynchronous request to generate a response to the user’s query. The chat message history is taken into account.

The result can be obtained via GET /api/knowledge-hub/chat/{chatId}/query/{queryId} endpoint, where queryId is the request identifier received in the current response.

Authorizations:

bearerAuth

path Parameters

chatId

required

integer <int64> (LongId)

Identifier of the chat in the knowledge base project.

Request Body schema: application/json
required

query required	string Text of the user’s query.
	object (RagSettings) Query processing settings.

Responses

Request samples

Payload

Content type

application/json

{"query": "string",
"settings": {"pipeline": "semantic",
"search": {"similarityTopK": 5,
"candidateRadius": 10,
"reranker": {"type": "manual",
"minScore": -10,
"maxChunksPerDocument": 1,
"maxChunks": 1,
"scoreReductionLimit": 1
},
"fullTextSearch": {"strategy": "hybrid",
"semanticPortion": 10,
"ftsPortion": 1,
"threshold": 1
},
"rephraseUserQuery": {"prompt": "Read the dialogue history, rephrase the current user question considering the history by adding it as context. Make the question more understandable, clear, and structured. Add similar queries and a title to the question, and return the text with the title."
},
"segment": "FAQ"
},
"llm": {"model": "string",
"contextWindow": 4000,
"maxTokens": 500,
"temperature": 1,
"topP": 1,
"frequencyPenalty": -2,
"presencePenalty": -2
},
"responseGeneration": {"prompt": "string"
}
}
}

Response samples

200
400

Content type

application/json

{"id": 0,
"chatId": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Processing status of the response generation request

Get the current processing status of the response generation request within a user chat. Long polling is used if the waitTimeSeconds parameter is specified.

Authorizations:

bearerAuth

path Parameters

chatId required	integer <int64> (LongId) Identifier of the chat in the knowledge base project.
queryId required	integer <int64> (LongId) Identifier of the response generation request.

query Parameters

waitTimeSeconds

integer <int32> [ 0 .. 30 ]

Default: 3

HTTP request timeout. Used in long polling.

Responses

Response samples

200

Content type

application/json

{"id": 0,
"chatId": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Cancel response generation request within a user chat

Authorizations:

bearerAuth

path Parameters

chatId required	integer <int64> (LongId) Identifier of the chat in the knowledge base project.
queryId required	integer <int64> (LongId) Identifier of the response generation request.

Responses

Response samples

200

Content type

application/json

{"id": 0,
"chatId": 0,
"request": "string",
"response": "string",
"status": "READY_TO_PROCESS",
"createdAt": "2019-08-24T14:15:22Z",
"updatedAt": "2019-08-24T14:15:22Z",
"comment": "string"
}

Tovie Data Agent public API (1.0.0)

ProjectsPublic

Get project info

Authorizations:

Responses

Response samples

QueriesPublic

Retrieve chunks

Authorizations:

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Generate response

Authorizations:

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Generate response (asynchronous request)

Authorizations:

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Processing status of the response generation request

Authorizations:

path Parameters

query Parameters

Responses

Response samples

Cancel response generation request

Authorizations:

path Parameters

Responses

Response samples

ChatsPublic

Chat creation

Authorizations:

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

User chat information

Authorizations:

path Parameters

Responses

Response samples

Retrieve chunks

Authorizations:

path Parameters

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Generate response

Authorizations:

path Parameters

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Generate response (asynchronous request)

Authorizations:

path Parameters

Request Body schema: application/jsonrequired

Responses

Request samples

Response samples

Processing status of the response generation request

Authorizations:

path Parameters

query Parameters

Responses

Response samples

Cancel response generation request within a user chat

Authorizations:

path Parameters

Responses

Response samples

Request Body schema: application/json
required

Request Body schema: application/json
required

Request Body schema: application/json
required

Request Body schema: application/json
required

Request Body schema: application/json
required

Request Body schema: application/json
required

Request Body schema: application/json
required