Streaming responce in fast api using langchain
async def get_streaming_response(self, query): if self.retriever is not None: print(“Preparing Streaming…!!!”) callback = self.window_memory.callbacks[0] await self.window_memory.arun(query, callback=callback) print(“nStreaming done..!!”) else: print(“No retriever found to process query.”) yield “No retriever available” I am using AsyncIteratorCallbackHandler class AsyncIteratorCallbackHandler(BaseCallbackHandler): def init(self): self.stdout_handler = StreamingStdOutCallbackHandler() async def on_llm_new_token(self, token: str, **kwargs: Any) -> Any: await self.stdout_handler.on_llm_new_token(token, **kwargs) I am […]