app.py

from langchain_core.prompts import PromptTemplate
from langchain_community.embeddings import HuggingFaceBgeEmbeddings
from langchain_community.vectorstores import FAISS
from langchain_community.llms.ctransformers import CTransformers
from langchain.chains.retrieval_qa.base import RetrievalQA
from langchain_community.llms import HuggingFaceHub

from langchain.document_loaders import PyPDFLoader
from langchain.document_loaders import PyPDFDirectoryLoader

from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import FAISS

from langchain_community.embeddings import HuggingFaceBgeEmbeddings

from langchain.prompts import PromptTemplate

from langchain.chains import create_retrieval_chain
from langchain.chains import RetrievalQA

from langchain.chains.combine_documents import create_stuff_documents_chain

import os 
import streamlit as st
import fitz  # PyMuPDF
from PIL import Image
import io

DB_FAISS_PATH = 'vectorstores/'
pdf_path = 'Oxford/Oxford-psychiatric-handbook-1-760.pdf'

# custom_prompt_template = '''use the following pieces of information to answer the user's questions.
# If you don't know the answer, please just say that don't know the answer, don't try to make uo an answer.
# Context : {context}
# Question : {question}
# only return the helpful answer below and nothing else.
# '''
custom_prompt_template =  prompt_template="""
Use the following piece of context to answer the question asked.
Please try to provide the answer only based on the context
{context}
Question:{question}
 """
def set_custom_prompt():
    """
    Prompt template for QA retrieval for vector stores
    """
    prompt = PromptTemplate(template = custom_prompt_template,
                            input_variables = ['context','question'])
    
    return prompt
    
os.environ['HUGGINGFACEHUB_API_TOKEN'] = os.getenv('HUGGINGFACEHUB_API_TOKEN')
def load_llm():
    # llm = CTransformers(
    #     model = 'TheBloke/Llama-2-7B-Chat-GGML',
    #     model_type = 'llama',
    #     max_new_token = 512,
    #     temperature = 0.5
    # )
    llm = HuggingFaceHub(
    repo_id = "mistralai/Mistral-7B-v0.1",
        huggingfacehub_api_token=os.getenv('HUGGINGFACEHUB_API_TOKEN'),
    model_kwargs = {'temperature': 0.1, "max_length": 500}
)
    return llm

def retrieval_qa_chain(llm,prompt,db):
    qa_chain = RetrievalQA.from_chain_type(
        llm = llm,
        chain_type = 'stuff',
        retriever = db.as_retriever(search_type = 'similarity',search_kwargs = {'k': 3}),
        return_source_documents = True,
        chain_type_kwargs = {'prompt': prompt}
    )

    return qa_chain

def qa_bot():
    embeddings = HuggingFaceBgeEmbeddings(model_name = 'BAAI/bge-small-en-v1.5',#'sentence-transformers/all-MiniLM-L6-v2',
                                          model_kwargs = {'device':'cpu'},
                                            encode_kwargs = {'normalize_embeddings': True})
    
    
    db = FAISS.load_local(DB_FAISS_PATH, embeddings, allow_dangerous_deserialization=True)
    llm = load_llm()
    qa_prompt = set_custom_prompt()
    qa = retrieval_qa_chain(llm,qa_prompt, db)

    return qa

def final_result(query):
    qa_result = qa_bot()
    response = qa_result({'query' : query})

    return response

def get_pdf_page_as_image(pdf_path, page_number):
    document = fitz.open(pdf_path)
    page = document.load_page(page_number)  
    pix = page.get_pixmap()
    img = Image.open(io.BytesIO(pix.tobytes()))
    return img

# Streamlit webpage title
st.title('Medical Chatbot')

# User input
user_query = st.text_input("Please enter your question:")

# Button to get answer
if st.button('Get Answer'):
    if user_query:
        # Call the function from your chatbot script
        response = final_result(user_query)
        if response:
            # Displaying the response
            st.write("### Answer")
            st.write(response['result'])

            # Displaying source document details if available
            if 'source_documents' in response:
                st.write("### Source Document Information")
                for doc in response['source_documents']:
                    # Retrieve and format page content by replacing '\n' with new line
                    formatted_content = doc.page_content.replace("\\n", "\n")
                    st.write("#### Document Content")
                    st.text_area(label="Page Content", value=formatted_content, height=300)

                    # Retrieve source and page from metadata
                    source = doc.metadata['source']
                    page = doc.metadata['page']
                    st.write(f"Source: {source}")
                    st.write(f"Page Number: {page+1}")
                    
                    # Display the PDF page as an image
                    #source = r"{source}"
                    pdf_page_image = get_pdf_page_as_image(pdf_path, page)
                    st.image(pdf_page_image, caption=f"Page {page+1} from {source}")
                    
        else:
            st.write("Sorry, I couldn't find an answer to your question.")
    else:
        st.write("Please enter a question to get an answer.")