app.py

import streamlit as st
from dotenv import load_dotenv
from streamlit_extras.add_vertical_space import add_vertical_space
from langchain.chains import RetrievalQA
from langchain.vectorstores import Chroma
from langchain.prompts import PromptTemplate
from langchain.embeddings import GooglePalmEmbeddings
from langchain.llms import GooglePalm
import os, glob
from design import image, link, layout, footer

__import__('pysqlite3')
import sys

sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')

st.set_page_config(page_title="ZeroRead", layout="wide",page_icon="📚",initial_sidebar_state="expanded")

prompt_template  = """
Use the following piece of context to answer the question. Please provide a detailed response, that should be long for each of the question. Provide formula or python code from the context for whichever question possible.

{context}

Question: {question}
"""

prompt = PromptTemplate(template = prompt_template , input_variables=["context", "question"])

load_dotenv()

def main():
    st.header("ZeroRead 📚💬")
    st.write("Books that are used are exactly from our syllabus and are available [here](https://drive.google.com/drive/folders/1fbdrnSsf4zO2cI7R8mi57YuWVstkoIK5?usp=share_link)")
    os.environ['GOOGLE_API_KEY'] = st.secrets['GOOGLE_API_KEY']

    select_subject = st.selectbox("Pick a Subject", ["18AI744-Business Intelligence", "18AI734-Cloud Computing", "18AI72-Machine-Learning (Mod-1,3,4,5)"])
    
    embeddings=GooglePalmEmbeddings()

    vectordb = Chroma(persist_directory=select_subject, embedding_function=embeddings)

    retriever = vectordb.as_retriever(search_kwargs={"k": 3,'include_metadata': True})

    llm = GooglePalm(temperature=0.5)  
    chain = RetrievalQA.from_chain_type(
        llm=llm,
        chain_type="stuff",
        retriever=retriever,
        return_source_documents=True,
        chain_type_kwargs={"prompt":prompt}
    )

    query = st.text_input("Ask questions from the book:")
   
    if query:
        response = chain(query)
        st.write(response["result"])
        with st.expander("TextBook -  Context"):
            for doc in response["source_documents"]:
                st.write(f"{doc.metadata['source']} \n {doc.page_content}")
                



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