Skip to content

SearchString API

The SearchString API transforms and optimizes user queries for better search results by adding conversational context and translating to English.

Overview

User messages are often suboptimal for vector or full-text search, especially in ongoing conversations where context is implied. This API enhances queries by:

  • Adding conversational context from chat history
  • Translating non-English queries to English
  • Expanding vague references with specific details
  • Creating optimal search strings for knowledge base queries

Methods

unique_sdk.SearchString.create - Transform and optimize queries

Transform a user prompt into an optimized search string.

Parameters:

  • prompt (str, required) - User's original query text
  • chatId (str, optional) - Chat ID to include conversation context for better results
  • messages (List[HistoryMessage], optional) - Explicit message history for context. See HistoryMessage for structure.
  • languageModel (Literal["AZURE_GPT_4_0613", "AZURE_GPT_4_32K_0613", "AZURE_GPT_4_TURBO_1106"], optional) - Language model to use for transformation

Returns:

Returns a SearchString object.

Example - Basic Usage:

1
2
3
4
5
search_string = unique_sdk.SearchString.create(
    user_id=user_id,
    company_id=company_id,
    prompt="Was ist der Sinn des Lebens?",  # German query
)

Example - With Chat Context:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
# Conversation context:
# User: "Tell me about The Hitchhiker's Guide to the Galaxy"
# Assistant: "It's a science fiction comedy series..."
# User: "Who is the author?"

search_string = unique_sdk.SearchString.create(
    user_id=user_id,
    company_id=company_id,
    prompt="Who is the author?",  # Vague without context
    chatId=chat_id
)

Example - With Explicit Message History:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
search_string = unique_sdk.SearchString.create(
    user_id=user_id,
    company_id=company_id,
    prompt="What are the main themes?",
    messages=[
        {"role": "user", "content": "Tell me about 1984"},
        {"role": "assistant", "content": "1984 is a dystopian novel by George Orwell..."},
        {"role": "user", "content": "What are the main themes?"}
    ]
)

Input Types

HistoryMessage

The HistoryMessage type defines a message in the conversation history

Fields:

  • role (Literal["system", "user", "assistant"], required) - Message role
  • text (str, required) - Message text content

Used in: SearchString.create() messages parameter