semantic_router.encoders.litellm

litellm_to_list

def litellm_to_list(embeds: "litellm.EmbeddingResponse") -> list[list[float]]

Convert a LiteLLM embedding response to a list of embeddings.

Arguments:

• embeds: The LiteLLM embedding response.

Returns:

A list of embeddings.

LiteLLMEncoder Objects

class LiteLLMEncoder(DenseEncoder, AsymmetricDenseMixin)

LiteLLM encoder class for generating embeddings using LiteLLM.

The LiteLLMEncoder class is a subclass of DenseEncoder and utilizes the LiteLLM SDK to generate embeddings for given documents. It supports all encoders supported by LiteLLM and supports customization of the score threshold for filtering or processing the embeddings.

litellm is an optional dependency: install it with pip install "semantic-router[litellm]".

__init__

def __init__(name: str | None = None,
             score_threshold: float | None = None,
             api_key: str | None = None)

Initialize the LiteLLMEncoder.

Arguments:

• name (str): The name of the embedding model to use. Must use LiteLLM naming convention (e.g. "openai/text-embedding-3-small" or "mistral/mistral-embed").
• score_threshold (float): The score threshold for the embeddings.

Raises:

• ImportError: If litellm is not installed.

__call__

def __call__(docs: list[Any], **kwargs) -> list[list[float]]

Encode a list of text documents into embeddings using LiteLLM.

Arguments:

• docs: List of text documents to encode.

Returns:

List of embeddings for each document.

acall

async def acall(docs: list[Any], **kwargs) -> list[list[float]]

Encode a list of documents into embeddings using LiteLLM asynchronously.

Arguments:

• docs: List of documents to encode.

Returns:

List of embeddings for each document.