app.py

import os
import streamlit as st
from PyPDF2 import PdfReader
from langchain.text_splitter import CharacterTextSplitter
from langchain.embeddings import HuggingFaceBgeEmbeddings
from langchain.vectorstores import FAISS
from langchain.chat_models import ChatOpenAI
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationalRetrievalChain
from htmlTemplates import css, bot_template, user_template
from langchain.llms import HuggingFaceHub
from deep_translator import GoogleTranslator

# set this key as an environment variable
os.environ["HUGGINGFACEHUB_API_TOKEN"] = st.secrets['huggingface_token']
###########################################################################################
def get_pdf_text(pdf_docs : list) -> str:
    text = ""
    for pdf in pdf_docs:
        pdf_reader = PdfReader(pdf)
        for page in pdf_reader.pages:
            text += page.extract_text()
    return text
#########################################################################
#def load_file(): 
    #loader = TextLoader('en.txt')
  #  documents = loader.load()
  #  return documents 
#####################################################################
def get_text_chunks(text:str) ->list:
    text_splitter = CharacterTextSplitter(
        separator="\n", chunk_size=1500, chunk_overlap=300, length_function=len
    )
    chunks = text_splitter.split_text(text)
    return chunks


def get_vectorstore(text_chunks : list) -> FAISS:
    model = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
    encode_kwargs = {
        "normalize_embeddings": True
    }  # set True to compute cosine similarity
    embeddings = HuggingFaceBgeEmbeddings(
        model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
    )
    vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
    return vectorstore


def get_conversation_chain(vectorstore:FAISS) -> ConversationalRetrievalChain:
    # llm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo-0613")
    llm = HuggingFaceHub(
        repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1",
        #repo_id="TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF"
        model_kwargs={"temperature": 0.5, "max_length": 2048},
    )

    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
    conversation_chain = ConversationalRetrievalChain.from_llm(
        llm=llm, retriever=vectorstore.as_retriever(), memory=memory
    )
    return conversation_chain


def handle_userinput(user_question:str):
    response = st.session_state.conversation({"question": user_question})
    st.session_state.chat_history = response["chat_history"]

    for i, message in enumerate(st.session_state.chat_history):
        if i % 2 == 0:
            text2=message.content 
            translator = GoogleTranslator(source='english', target='persian')
            result = translator.translate(text2)
            st.write("سوال کاربر: "+result)
        else:
            text1=message.content 
            translator = GoogleTranslator(source='english', target='persian')
            result = translator.translate(text1)
            st.write("پاسخ ربات: "+result)

#############################################################################################################
def read_pdf_pr_en(pdf_file_path):
  from deep_translator import GoogleTranslator
  import PyPDF2
  # مسیر فایل PDF را تعیین کنید
  #pdf_file_path = '/content/d2en.pdf'
  # باز کردن فایل PDF
  with open(pdf_file_path, 'rb') as pdf_file:
    pdf_reader = PyPDF2.PdfReader(pdf_file)
    # خواندن محتوای صفحه‌ها
    full_text = ''
    for page in pdf_reader.pages:
        page_pdf=page.extract_text()
        translator = GoogleTranslator(source='persian', target='english')
        result = translator.translate(page_pdf)
        full_text +=result
    st.write(full_text)
    return(full_text)
#################################################################################################################
def get_pdf_text_fa_en(pdf_docs):
    text = "" 
    for pdf in pdf_docs: 
        pdf_reader = PdfReader(pdf) 
    for page in pdf_reader.pages: 
        txt_page=page.extract_text()
        text += txt_page
    return text
########################################################################################################
def main():
   
    st.set_page_config(
        page_title="Chat Bot 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 Bot PDFs :books:")
    user_question1 = st.text_input("Ask a question about your documents:")
    translator = GoogleTranslator(source='persian', target='english')
    user_question = translator.translate(user_question1)
    if st.button("Answer"):
            with st.spinner("Answering"):
              handle_userinput(user_question)
    if st.button("CLEAR"):
            with st.spinner("CLEARING"):
              st.cache_data.clear()

    
    with st.sidebar:
        st.subheader("Your documents")
        pdf_docs = st.file_uploader(
            "Upload your PDFs here and click on 'Process'", accept_multiple_files=True)
        text=get_pdf_text_fa_en(pdf_docs=pdf_docs)
        st.write(text)
        
        if st.button("Process and build Model"):
            with st.spinner("Processing"):
                raw_text=get_pdf_text_fa_en(pdf_docs=pdf_docs)
                # get the text chunks
                text_chunks = get_text_chunks(raw_text)
                # create vector store
                vectorstore = get_vectorstore(text_chunks)
                # create conversation chain
                st.session_state.conversation = get_conversation_chain(vectorstore)
                #compelete build model
                st.write("compelete build model")


if __name__ == "__main__":  
    main()