Module query_engine

Class QueryEngine

Manages interactions with different LLM and embedding services for
querying and indexing, supporting dynamic service initialization.

Attributes:

Methods:

• init(self, api_key, model_name, embedding_name, llm_type='azure', embedding_type='azure', endpoint=None, api_version=None)

  Initializes QueryEngine with specific service configurations.

  Args: api_key (str): API key for the LLM and embedding services. model_name (str): Model name for the LLM service. embedding_name (str): Model name for the embedding service. llm_type (str): Type of LLM service ('azure' or 'together'). embedding_type (str): Type of embedding service ('azure' or 'together'). endpoint (str, optional): Endpoint URL for the service, for Azure. api_version (str, optional): API version for the service, for Azure.

• create_vector_store_index(self, directory_path)

  Creates a vector index from documents in a specified directory.

• initialize_services(self)

  Initializes the specified LLM and embedding services based on config.

• query(self, index, query_text)

  Queries an index with text and returns the results.