app.py

import os
from dotenv import load_dotenv

from langchain.vectorstores import FAISS
from langchain.embeddings import OpenAIEmbeddings
from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import CharacterTextSplitter
from langchain.chains.question_answering import load_qa_chain
from langchain_groq import ChatGroq  # ✅ Correct import

# Load environment variables (make sure you have a .env file)
load_dotenv()

# Get keys from environment
groq_api_key = os.getenv("GROQ_API_KEY")
openai_api_key = os.getenv("OPENAI_API_KEY")

# Load and split PDF
loader = PyPDFLoader("university_karachi_semester_fees.pdf")
pages = loader.load_and_split()

text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
docs = text_splitter.split_documents(pages)

# Embed and create vector store
embedding = OpenAIEmbeddings(openai_api_key=openai_api_key)  # ✅ Pass the OpenAI API key
vectorstore = FAISS.from_documents(docs, embedding)

# Set up LLM with GROQ
llm = ChatGroq(
    model_name="mixtral-8x7b-32768",  # ✅ Correct parameter name
    groq_api_key=groq_api_key
)

# Question-answering chain
qa_chain = load_qa_chain(llm, chain_type="stuff")

# Simple user loop
while True:
    query = input("\nAsk something about KU: ")
    if query.lower() in ["exit", "quit"]:
        break
    matched_docs = vectorstore.similarity_search(query)
    response = qa_chain.run(input_documents=matched_docs, question=query)
    print("\nAnswer:", response)