app.py

import os
import streamlit as st
from dotenv import load_dotenv
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.llms import llamacpp
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.callbacks import CallbackManager, StreamingStdOutCallbackHandler
from langchain.chains import create_history_aware_retriever, create_retrieval_chain, ConversationalRetrievalChain
from langchain.document_loaders import TextLoader
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain_community.chat_message_histories.streamlit import StreamlitChatMessageHistory
from langchain.prompts import PromptTemplate
from langchain.vectorstores import Chroma
from utills import load_txt_documents, split_docs, load_uploaded_documents, retriever_from_chroma
from langchain.text_splitter import TokenTextSplitter, RecursiveCharacterTextSplitter
from langchain_community.document_loaders.directory import DirectoryLoader
from HTML_templates import css, bot_template, user_template




def get_vectorstore(text_chunks):
    model_name = "sentence-transformers/all-mpnet-base-v2"
    model_kwargs = {'device': 'cpu'}
    encode_kwargs = {'normalize_embeddings': True}
    embeddings = HuggingFaceEmbeddings(
        model_name=model_name,
        model_kwargs=model_kwargs,
        encode_kwargs=encode_kwargs
    )
    vectorstore_path = "docs/chroma/"
    if not os.path.exists(vectorstore_path):
        os.makedirs(vectorstore_path)
    vectorstore = Chroma.from_documents(
        documents=text_chunks, embedding=embeddings, persist_directory="docs/chroma/")
    return vectorstore    



data_path = "data"


    
documents = []
    
for filename in os.listdir(data_path):

    if filename.endswith('.txt'):

        file_path = os.path.join(data_path, filename)

        documents = TextLoader(file_path).load()

        documents.extend(documents)
    



docs = split_docs(documents, 350, 40)

vectorstore = get_vectorstore(docs)






def main(vectorstore):

    st.set_page_config(page_title="Chat with multiple PDFs",
                       page_icon=":books:")
    st.write(css, unsafe_allow_html=True)

    if "conversation" not in st.session_state:
        st.session_state.conversation = None
    if "chat_history" not in st.session_state:
        st.session_state.chat_history = None

    st.header("Chat with multiple PDFs :books:")
    user_question = st.text_input("Ask a question about your documents:")
    if user_question:
        handle_userinput(user_question,vectorstore)


def handle_userinput(user_question,vectorstore):
    Rag_chain = create_conversational_rag_chain(vectorstore)
    msgs = StreamlitChatMessageHistory(key="special_app_key")
    response = Rag_chain.invoke(
    {"input": user_question},
    
    config={
        "configurable": {"session_id": "k"}
    }, 
    )["answer"]
    
    st.session_state.chat_history = response['chat_history']

    # Display messages in chat history
    for i, message in enumerate(st.session_state.chat_history):
        if i % 2 == 0:
            st.write(user_template.replace("{{MSG}}", message.content), unsafe_allow_html=True)
        else:
            st.write(bot_template.replace("{{MSG}}", message.content), unsafe_allow_html=True)

    if 'retrieved_documents' in response:
        st.subheader("Retrieved Documents")
        for doc in response['source_documents']:
            st.write(f"Document: {doc.metadata['source']}")
            st.write(doc.page_content)       



def create_conversational_rag_chain(vectorstore):
    
    model_path = ('qwen2-0_5b-instruct-q4_0.gguf')

    callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])

    llm = llamacpp.LlamaCpp(
        model_path=model_path,
        n_gpu_layers=1,
        temperature=0.1,
        top_p=0.9,
        n_ctx=22000,
        max_tokens=200,
        repeat_penalty=1.7,
        callback_manager=callback_manager,
        verbose=False,
    )

    contextualize_q_system_prompt = """Given a context, chat history and the latest user question
    which maybe reference context in the chat history, formulate a standalone question
    which can be understood without the chat history. Do NOT answer the question,
    just reformulate it if needed and otherwise return it as is."""

    contextualize_q_prompt = ChatPromptTemplate.from_messages(
        [
            ("system", contextualize_q_system_prompt),
            MessagesPlaceholder("chat_history"),
            ("human", "{input}"),
        ]
    )

    ha_retriever = create_history_aware_retriever(llm, vectorstore.as_retriever(), contextualize_q_prompt)

    qa_system_prompt = """You are an assistant for question-answering tasks. 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. Be as informative as possible, be polite and formal.\n{context}"""

    qa_prompt = ChatPromptTemplate.from_messages(
        [
            ("system", qa_system_prompt),
            MessagesPlaceholder("chat_history"),
            ("human", "{input}"),
        ]
    )

    question_answer_chain = create_stuff_documents_chain(llm, qa_prompt)
    rag_chain = create_retrieval_chain(ha_retriever, question_answer_chain)
    msgs = StreamlitChatMessageHistory(key="special_app_key")

    conversational_rag_chain = RunnableWithMessageHistory(
        rag_chain,
        lambda session_id: msgs,
        input_messages_key="input",
        history_messages_key="chat_history",
        output_messages_key="answer",
        return_source_documents=True
    )

    return conversational_rag_chain


 

if __name__ == "__main__":
    main(vectorstore)