Querying the Assistant (Q&A)

Once we've uploaded and processed documents from the knowledge base, we can ask a specific assistant a question and get an accurate answer with source references.

Main URL for API access: https://bot.insightstream.ru

Sending a Question to Get an Answer

GET /{bot_id}/api_v1/answers

After sending the request, the assistant will search the indexed documents and return an answer with source references.

Python

import requests

# Replace with your assistant ID
bot_id = "my_search_bot"

# Replace with your question
search_query = "What are the urgency assessment criteria?"

# Form the URL for the request
url = f"https://bot.insightstream.ru/{bot_id}/api_v1/answers"

# Request parameters
params = {
    "search_query": search_query,
    "sync": 1  # Synchronous request to get an answer immediately
}

# Execute GET request
response = requests.get(url, params=params)

# Check response status
if response.status_code == 200:
    # Get JSON response
    answer_data = response.json()
    
    # Print the answer
    print(f"Answer: {answer_data['summary']}")
    
    # Print sources
    print("\nSources:")
    for doc in answer_data["documents"]:
        print(f"- {doc['title']}: {doc['passage_md']}")
        print(f"  URL: {doc['uri']}")
        
    # Print relevant questions
    if "relevant_questions" in answer_data:
        print("\nSimilar questions:")
        for question in answer_data["relevant_questions"]:
            print(f"- {question}")
else:
    print(f"Error: {response.status_code}")
    print(response.text)

cURL

curl -G "https://bot.insightstream.ru/my_search_bot/api_v1/answers" \
     --data-urlencode "search_query=What are the urgency assessment criteria?" \
     --data "sync=1"

Request Parameters

Parameter	Type	Required	Description
search_query	string	Yes	Question you're asking the assistant
sync	integer	No	If set to 1, the request is executed synchronously and returns an answer immediately.

Response Codes

• 200 OK: Returns an answer and links to relevant documents.
• 400 Bad Request: Invalid input data.
• 404 Not Found: Assistant not found or indexing not completed.
• 429 Too Many Requests: Request limit exceeded.

Example Response

The response is returned in JSON format.

{
  "summary": "Urgency assessment criteria include task priority, impact on business processes, execution deadlines, and resources required for resolution.",
  "documents": [
    {
      "title": "Task_Assessment_Regulation.pdf",
      "uri": "https://bot.insightstream.ru/documents/tmp/Task_Assessment_Regulation.pdf",
      "passage_md": "Task urgency assessment criteria include: 1) Priority - high, medium, low; 2) Impact on business processes - critical, significant, minor; 3) Execution deadlines - immediately, within a day, within a week; 4) Required resources.",
      "doc_type": "application/pdf",
      "loc": {
        "from": 12,
        "to": 13
      }
    }
  ],
  "relevant_questions": [
    "How to properly assess task priority?",
    "Which tasks are considered critical in terms of business impact?",
    "How to allocate resources between tasks of different urgency?"
  ]
}

Response Parameter Descriptions

Parameter	Description
summary	Answer to the asked question, formed based on found documents
documents	List of documents used to form the answer
title	Document name
uri	Link to the source document
passage_md	Relevant passage from the document used for the answer
doc_type	MIME type of the document (e.g., PDF, DOCX, etc.)
relevant_questions	List of similar questions that can be asked to the assistant