import streamlit as st
from llama_index.core import StorageContext, load_index_from_storage, VectorStoreIndex, SimpleDirectoryReader, ChatPromptTemplate
from llama_index.llms.huggingface import HuggingFaceInferenceAPI
from dotenv import load_dotenv
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core import Settings
import os
import base64
import fitz  # PyMuPDF

# Load environment variables
load_dotenv()

# Configure the Llama index settings
Settings.llm = HuggingFaceInferenceAPI(
    model_name="google/gemma-1.1-7b-it",
    tokenizer_name="google/gemma-1.1-7b-it",
    context_window=3000,
    token=os.getenv("HF_TOKEN"),
    max_new_tokens=512,
    generate_kwargs={"temperature": 0.1},
)
Settings.embed_model = HuggingFaceEmbedding(
    model_name="BAAI/bge-large-en-v1.5"
)

# Define the directory for persistent storage and data
PERSIST_DIR = "./db"
DATA_DIR = "data"

# Ensure data directory exists
os.makedirs(DATA_DIR, exist_ok=True)
os.makedirs(PERSIST_DIR, exist_ok=True)


def displayPDF(file):
    # Open the PDF file
    pdf_document = fitz.open(file)
    # Render the first page
    first_page = pdf_document.load_page(0)
    # Convert the page to an image
    pix = first_page.get_pixmap()
    # Save the image
    image_path = "first_page.png"
    pix.save(image_path)
    # Display the image in Streamlit
    st.image(image_path)


def data_ingestion():
    documents = SimpleDirectoryReader(DATA_DIR).load_data()
    storage_context = StorageContext.from_defaults()
    index = VectorStoreIndex.from_documents(documents)
    index.storage_context.persist(persist_dir=PERSIST_DIR)

def handle_query(query):
    storage_context = StorageContext.from_defaults(persist_dir=PERSIST_DIR)
    index = load_index_from_storage(storage_context)
    chat_text_qa_msgs = [
    (
        "user",
        """You are a Q/A Scientific Assistant.Be very careful and answer in detail.
        Context:
        {context_str}
        Question:
        {query_str}
        """
    )
    ]
    text_qa_template = ChatPromptTemplate.from_messages(chat_text_qa_msgs)
    
    query_engine = index.as_query_engine(text_qa_template=text_qa_template)
    answer = query_engine.query(query)
    
    if hasattr(answer, 'response'):
        return answer.response
    elif isinstance(answer, dict) and 'response' in answer:
        return answer['response']
    else:
        return "Sorry, I couldn't find an answer."


# Streamlit app initialization
st.title("RAG Extractor")

if 'messages' not in st.session_state:
    st.session_state.messages = [{'role': 'assistant', "content": 'Hello I am Pingu! Upload a PDF and ask me anything about its content.'}]

# Define your custom icon
custom_icon_url = "custom_icon.jpeg"  # Adjust this to your icon's file path

with st.sidebar:
    st.title("Input")
    uploaded_file = st.file_uploader("Upload your PDF Files and then click on the Submit & Process Button")
    if st.button("Submit & Process"):
        with st.spinner("Loading..."):
            filepath = "data/saved_pdf.pdf"
            with open(filepath, "wb") as f:
                f.write(uploaded_file.getbuffer())  
            data_ingestion()  # Process PDF every time new file is uploaded
            st.success("PDF is ready!")
            displayPDF(filepath)

user_prompt = st.chat_input("Ask me anything about the content of the PDF:")
if user_prompt:
    st.session_state.messages.append({'role': 'user', "content": user_prompt})
    response = handle_query(user_prompt)
    st.session_state.messages.append({'role': 'assistant', "content": response})

for message in st.session_state.messages:
    with st.chat_message(message['role']):
        st.write(message['content'])