Custom RAG Filter pipeline hang forever

[DuyTa506]

Hey guys, i;m very love this repo, and webui, too
I'm trying to test the pipline as a rag filter to process the user query to add context in it.
Here's the code
When i first time test the question is smooth and okay. But after that, all the UI hang, and this issue raise.

File "E:\AI_Project\NCKH\Graduation\open-webui\backend\main.py", line 1177, in chat_completed
"chat_id": data["chat_id"],
~~~~^^^^^^^^^^^
TypeError: 'NoneType' object is not subscriptable

What did i missed ?

from typing import List, Optional
from pydantic import BaseModel
from schemas import OpenAIChatMessage
from utils.pipelines.main import get_last_user_message, get_last_assistant_message
from blueprints.rag_utils import format_docs
from blueprints.prompts import accurate_rag_prompt
class Pipeline:
    class Valves(BaseModel):
        # List target pipeline ids (models) that this filter will be connected to.
        # If you want to connect this filter to all pipelines, you can set pipelines to ["*"]
        pipelines: List[str] = []

        # Assign a priority level to the filter pipeline.
        # The priority level determines the order in which the filter pipelines are executed.
        # The lower the number, the higher the priority.
        priority: int = 0

        # Add your custom parameters/configuration here e.g. API_KEY that you want user to configure etc.
        pass

    def __init__(self):
        self.documents = None
        self.index = None
        self.chain = None
        self.type = "filter"
        self.name = "Filter"
        self.valves = self.Valves(**{"pipelines": ["*"]})

        pass

    async def on_startup(self):
        # This function is called when the server is started.
        print(f"on_startup:{__name__}")
        from typing import List
        import bs4
        from langchain.text_splitter import RecursiveCharacterTextSplitter
        from langchain_community.document_loaders import WebBaseLoader
        from langchain_community.vectorstores import Chroma
        from langchain_core.runnables import RunnablePassthrough
        from langchain_community.embeddings import HuggingFaceEmbeddings
        loader = WebBaseLoader(
        web_paths=("https://viblo.asia/p/graphrag-mot-su-nang-cap-moi-cua-rag-truyen-thong-chang-EoW4oXRBJml",),
        bs_kwargs=dict(
            parse_only=bs4.SoupStrainer(
                class_=("post-content", "post-title", "post-header")
            )
        ),
    )

        docs = loader.load()

        text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
        splits = text_splitter.split_documents(docs)

        embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")

        # Embed
        vectorstore = Chroma.from_documents(documents=splits, 
                                            embedding=embeddings) ##Replace the real Embedding in here

        retriever = vectorstore.as_retriever()
        
        self.chain = (
            {"context": retriever | format_docs, "question": RunnablePassthrough()}
            | accurate_rag_prompt
        )
        
        pass

    async def on_shutdown(self):
        # This function is called when the server is stopped.
        print(f"on_shutdown:{__name__}")
        pass
    
    def add_or_update_user_message(self, content: str, messages: List[dict]):
        """
        Adds a new system message at the beginning of the messages list
        or updates the existing system message at the beginning.

        :param msg: The message to be added or appended.
        :param messages: The list of message dictionaries.
        :return: The updated list of message dictionaries.
        """

        if messages and messages[0].get("role") == "user":
            messages[0]["content"] += f"{content}\n{messages[0]['content']}"
        else:
            # Insert at the beginning
            messages.insert(0, {"role": "user", "content": content})

        return messages

    async def inlet(self, body: dict, user: Optional[dict] = None) -> dict:
        messages = body.get("messages", [])
        user_message = get_last_user_message(messages)
        if user_message is not None:
            rag_prompt = self.chain.invoke(user_message).text
            for message in reversed(messages):
                if message["role"] == "user":
                    message["content"] = rag_prompt
                break
        print(body)
        body = {**body, "messages": messages}
        return body
        
    async def outlet(self, body: dict, user: Optional[dict] = None) -> dict:
            pass
            # messages = body["messages"]
            # assistant_message = get_last_assistant_message(messages)

            # if assistant_message is not None:
            #         # Do something

            #         for message in reversed(messages):
            #             if message["role"] == "assistant":
            #                 message["content"] = "UPDATED CORRESPONDING CONTENT THAT USER WILL SEE"
            #                 break

            # body = {**body, "messages": messages}
            # return body

[DuyTa506]

The logic changed , i'm just swap to chunk the output of retrieval chain to separate system prompt and update only system prompt.
And it work
Still don't know what wrong with the old code