start.py

try: 
    st.session_state.retriever = ""
    st.session_state.rag_chain = "" 
    global rag_chain, retriever, em
    import requests
    from langchain.document_loaders import TextLoader
    from langchain.text_splitter import CharacterTextSplitter
    import re

    import requests
    #import fitz  # PyMuPDF
    
    import inspect

    def get_LN():
        """
        Returns the line number in the caller's source code.
        """
        # Get the current frame, then the caller's frame, and finally the line number
        frame = inspect.currentframe()
        caller_frame = frame.f_back
        line_number = caller_frame.f_lineno
        return str(line_number)


    st.session_state.ln = get_LN()

    os.environ["OPENAI_API_KEY"]  = os.getenv('openkey')
    org = os.getenv('openorg')


    from langchain.embeddings import OpenAIEmbeddings
    from langchain.vectorstores import Weaviate
    import weaviate
    from weaviate.embedded import EmbeddedOptions
    #from dotenv import load_dotenv,find_dotenv
    
    # Load OpenAI API key from .env file
    #load_dotenv(find_dotenv())
    st.session_state.ln = get_LN()
    
    # Setup vector database|
    client = weaviate.Client(
    embedded_options = EmbeddedOptions()
    )
    ############################################################################
    st.session_state.ln = get_LN()
    
    pickle_file_path = 'vectorstore.pkl'
    import pickle
    
    with open(pickle_file_path, 'rb') as file:
        docs = pickle.load(file)


    st.session_state.ln = get_LN()
    vectorstore = Weaviate.from_documents(
        client = client,    
        documents = docs,
        embedding = OpenAIEmbeddings(),
        by_text = False
    )
    st.session_state.ln = get_LN()
    # Define vectorstore as retriever to enable semantic search
    retriever = vectorstore.as_retriever()

    st.session_state.ln = get_LN()

############################################################################

    
    # Create RAG chain
    
    #from langchain.chat_models import ChatOpenAI
    
    from langchain_openai import ChatOpenAI
    
    from langchain.prompts import ChatPromptTemplate
    from langchain.schema.runnable import RunnablePassthrough
    from langchain.schema.output_parser import StrOutputParser
    
    # Define LLM
    llm = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0)
    
    #llm = ChatOpenAI(model_name="gpt-4", temperature=0.15)
    
    st.session_state.ln = get_LN()
    
    # Define prompt template / context
    template = """You are a lawyer responding to creditors questions. 
    Use the following pieces of retrieved context to answer the question. 
    If you don't know the answer, just say that you don't know and add a funny joke. 
    Keep the answer concise.
    Question: {question} 
    Context: {context} 
    Answer:
    """
    
    prompt = ChatPromptTemplate.from_template(template)
    
    # Setup RAG pipeline
    rag_chain = (
        {"context": retriever,  "question": RunnablePassthrough()} 
        | prompt 
        | llm
        | StrOutputParser() 
    )


    st.session_state.retriever = retriever
    st.session_state.rag_chain = rag_chain



    st.session_state.ln = get_LN()



    em = "aaa"
    st.session_state.started = True
except Exception as e:
    em = "bbb  " + str(e)



st.session_state.em = em