----------------------
Yes, you can call the query asynchronously using the
arun
method in the
LlamaIndexTool
class. Here's an example from the
source (
https://gpt-index.readthedocs.io/en/latest/reference/langchain_integrations/base.html):
async def arun(tool_input: Union[str, Dict], verbose: Optional[bool] = None, start_color: Optional[str] = 'green', color: Optional[str] = 'green', callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) β Any
This method allows you to run the tool asynchronously, which can be useful when querying a LlamaIndex.
----------------------
Note: you can ask me a follow up question by @ mentioning me again
:speech_balloon:
----------------------