app.py

import streamlit as st
import PyPDF2
import openai
import faiss
import os
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity

# Function to extract text from a PDF file
def extract_text_from_pdf(pdf_file):
    reader = PyPDF2.PdfReader(pdf_file)
    text = ""
    for page in reader.pages:
        text += page.extract_text()
    return text

# Function to generate embeddings for a piece of text
def get_embeddings(text, model="text-embedding-ada-002"):
    response = openai.Embedding.create(input=[text], model=model)
    return response['data'][0]['embedding']

# Function to search for similar content
def search_similar(query_embedding, index, stored_texts, top_k=3):
    distances, indices = index.search([query_embedding], top_k)
    results = [(stored_texts[i], distances[0][idx]) for idx, i in enumerate(indices[0])]
    return results

# Streamlit app starts here
st.title("Course Query Assistant")

# Input OpenAI API key
openai_api_key = st.text_input("Enter your OpenAI API key:", type="password")

if openai_api_key:
    openai.api_key = openai_api_key

    # Upload course materials
    uploaded_files = st.file_uploader("Upload Course Materials (PDFs)", type=["pdf"], accept_multiple_files=True)

    if uploaded_files:
        st.write("Processing uploaded course materials...")

        # Extract text and generate embeddings for all uploaded PDFs
        course_texts = []
        for uploaded_file in uploaded_files:
            text = extract_text_from_pdf(uploaded_file)
            course_texts.append(text)

        # Combine all course materials into one large text
        combined_text = " ".join(course_texts)

        # Split combined text into smaller chunks for embedding (max tokens ~1000)
        chunks = [combined_text[i:i+1000] for i in range(0, len(combined_text), 1000)]

        # Generate embeddings for all chunks
        embeddings = [get_embeddings(chunk) for chunk in chunks]

        # Create a FAISS index for similarity search
        index = faiss.IndexFlatL2(len(embeddings[0]))
        index.add(embeddings)

        st.write("Course materials have been processed and indexed.")

        # User query
        query = st.text_input("Enter your question about the course materials:")

        if query:
            # Generate embedding for the query
            query_embedding = get_embeddings(query)

            # Search for similar chunks in the FAISS index
            results = search_similar(query_embedding, index, chunks)

            # Create the context for the GPT prompt
            context = "\n".join([result[0] for result in results])
            modified_prompt = f"Context: {context}\n\nQuestion: {query}\n\nProvide a detailed answer based on the context."

            # Get the GPT-3.5-turbo response
            response = openai.ChatCompletion.create(
                model="gpt-3.5-turbo",
                messages=[{"role": "user", "content": modified_prompt}]
            )

            # Display the response
            st.write("### Intelligent Reply:")
            st.write(response['choices'][0]['message']['content'])