app.py

import os
from langchain_community.document_loaders import PyPDFLoader
from langchain_community.vectorstores import faiss
from langchain.memory import ConversationBufferMemory
from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
from tempfile import NamedTemporaryFile
from dotenv import load_dotenv
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain.chains import ConversationalRetrievalChain
import streamlit as st
import nest_asyncio

nest_asyncio.apply()
load_dotenv()

# Initialize app resources
st.set_page_config(page_title="StudyAssist", page_icon=":book:")
st.title("StudyAssist(pharmassist-v0)")
st.write(
    "An AI/RAG application to aid students in their studies, specially optimized for the pharm 028 students. In simpler terms, chat with your pdf"
)


@st.cache_resource
def initialize_resources():
    llm_gemini = ChatGoogleGenerativeAI(
        model="gemini-1.5-flash-latest", google_api_key=os.getenv("GOOGLE_API_KEY")
    )
    return llm_gemini


def get_retriever(pdf_file):
    with NamedTemporaryFile(suffix="pdf") as temp:
        temp.write(pdf_file.getvalue())
        pdf_loader = PyPDFLoader(temp.name, extract_images=True)
        pages = pdf_loader.load()

    # st.write(f"AI Chatbot for {course_material}")

    underlying_embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
    text_splitter = RecursiveCharacterTextSplitter(
        chunk_size=1000,
        chunk_overlap=20,
        length_function=len,
        is_separator_regex=False,
        separators="\n",
    )
    documents = text_splitter.split_documents(pages)
    vectorstore = faiss.FAISS.from_documents(documents, underlying_embeddings)
    doc_retiever = vectorstore.as_retriever(
        search_type="mmr", search_kwargs={"k": 5, "fetch_k": 10}
    )

    return doc_retiever


chat_model = initialize_resources()

# Streamlit UI
# Course list and pdf retrieval

courses = ["PMB", "PCL", "Kelechi_research"]  # "GSP", "CPM", "PCG",  "PCH"
course_pdfs = None
doc_retriever = None
conversational_chain = None

# course = st.sidebar.selectbox("Choose course", (courses))
# docs_path = f"pdfs/{course}"
# course_pdfs = os.listdir(docs_path)
# pdfs = [os.path.join(docs_path, pdf) for pdf in course_pdfs]

course_material = "{Not selected}"


# @st.cache_resource
def query_response(query, _retriever):
    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
    conversational_chain = ConversationalRetrievalChain.from_llm(
        llm=chat_model, retriever=_retriever, memory=memory, verbose=False
    )
    response = conversational_chain.run(query)

    return response


if "doc" not in st.session_state:
    st.session_state.doc = ""

course_material = st.file_uploader("or Upload your own pdf", type="pdf")

if st.session_state != "":
    try:
        with st.spinner("loading document.."):
            doc_retriever = get_retriever(course_material)
        st.success("File loading successful, vector db initialize")
    except Exception as e:
        st.error(e)

    # We store the conversation in the session state.
    # This will be use to render the chat conversation.
    # We initialize it with the first message we want to be greeted with.
    if "messages" not in st.session_state:
        st.session_state.messages = [
            {"role": "assistant", "content": "Yoo, How far boss?"}
        ]

    if "current_response" not in st.session_state:
        st.session_state.current_response = ""

    # We loop through each message in the session state and render it as
    # a chat message.
    for message in st.session_state.messages:
        with st.chat_message(message["role"]):
            st.markdown(message["content"])

    # We take questions/instructions from the chat input to pass to the LLM
    if user_prompt := st.chat_input("Ask...", key="user_input"):
        # Add our input to the session state
        st.session_state.messages.append({"role": "user", "content": user_prompt})

        # Add our input to the chat window
        with st.chat_message("user"):
            st.markdown(user_prompt)

        # Pass our input to the llm chain and capture the final responses.
        # here once the llm has finished generating the complete response.
        response = query_response(user_prompt, doc_retriever)
        # Add the response to the session state
        st.session_state.messages.append({"role": "assistant", "content": response})

        # Add the response to the chat window
        with st.chat_message("assistant"):
            st.markdown(response)
#
st.write("")
st.write("")


st.markdown(
    """
    <div style="text-align: center; padding: 1rem;">
        Project by <a href="https://github.com/kelechi-c" target="_blank" style="color: white; font-weight: bold; text-decoration: none;">
         kelechi(tensor)</a>
    </div>
""",
    unsafe_allow_html=True,
)