how to stream output messages only

[navaneethstabilix]

config = {"configurable": {"thread_id": "384156"}}

async def generate_chat_responses(message):
full_response = " "
async for chunk in react_graph_memory.astream(
{"messages": [("user", message)]}
,config=config,stream_mode="messages"
):
print("haha",chunk)
# yield f"{chunk}\n\n"
yield chunk

To consume this asynchronous generator

async def main():
async for response in generate_chat_responses("what is the weather here"):
print(response)

Run in an async environment (e.g., Jupyter or Python script)

import asyncio
await main()

how do i stream outpur message the rresponse from model onyl for this case

[dosubot[bot]]

To stream only the output messages from the model using your asynchronous generator code, you can modify the generate_chat_responses function to extract and yield only the message content from each chunk. Here's how you can do it:

async def generate_chat_responses(message):
    async for chunk in react_graph_memory.astream(
        {"messages": [("user", message)]},
        config=config,
        stream_mode="messages"
    ):
        # Assuming chunk is a dictionary with a key "content" for the message text
        message_content = chunk["message"]["content"]
        yield message_content

# To consume this asynchronous generator
async def main():
    async for response in generate_chat_responses("what is the weather here"):
        print(response)

# Run in an async environment (e.g., Jupyter or Python script)
import asyncio
await main()

This code assumes that each chunk object has a structure where the message content is accessible via chunk["message"]["content"]. Adjust the key access as necessary based on the actual structure of the chunk object in your implementation.

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}