Issue on docs about docs

[xg35]

that indeed did the job. now looks like this

[willydouhard]

Nice catch, what version of langflow are you using? Can you try upgrading it? Or maybe they released a breaking change.

[xg35]

Langflow v0.2.7

[willydouhard]

This seems pretty outdated, can you run pip -U langflow and try again?

[xg35]

yeah updated it to 0.4.16 and similar error:

TypeError: load_flow_from_json() got an unexpected keyword argument 'flow'
2023-09-13 16:32:38 - 'NoneType' object has no attribute 'agent'
Traceback (most recent call last):
File "/Users/simonau/Documents/Lending/env/lib/python3.11/site-packages/chainlit/utils.py", line 40, in wrapper
return await user_function(**params_values)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/simonau/Documents/Lending/main_tool_langflow.py", line 29, in main
flow.agent.llm_chain.llm.streaming = True
^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'agent'

[willydouhard]

So this error is a bit different, the flow was successfully loaded but it no longer have an agent attribute. But at least this is reaching the code of your app file.

[xg35]

import json

from langchain.agents import AgentExecutor
from langflow import load_flow_from_json
from chainlit.langflow import load_flow
import chainlit as cl

#import os

#os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY")
#os.environ["SERPAPI_API_KEY"] = os.getenv("SERPAPI_API_KEY")

with open("./schema.json", "r") as f:
schema = json.load(f)

@cl.on_chat_start
async def start():
flow = await load_flow(schema=schema)
cl.user_session.set("flow", flow)

@cl.on_message
async def main(message):
# Load the flow from the user session
flow = cl.user_session.get("flow") # type: AgentExecutor

# Enable streaming
flow.agent.llm_chain.llm.streaming = True

# Run the flow
res = await cl.make_async(flow.run)(
    message, callbacks=[cl.LangchainCallbackHandler()]
)

# Send the response
await cl.Message(content=res).send()

This is the exact code.

[willydouhard]

maybe update the main function like so:

@cl.on_message
async def main(message):
# Load the flow from the user session
flow = cl.user_session.get("flow") # type: AgentExecutor

# Run the flow
res = await cl.make_async(flow)(
    message, callbacks=[cl.LangchainCallbackHandler()]
)

# Send the response
await cl.Message(content=res).send()

[xg35]

Any idea how i can get it parse it correctly?

[willydouhard]

What do you mean by "parsing"?