app.py

import os
import streamlit as st
from PyPDF2 import PdfReader
from sentence_transformers import SentenceTransformer
import faiss
import numpy as np
import matplotlib.pyplot as plt
from groq import Groq

GROQ_API_KEY = "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"
client = Groq(api_key=GROQ_API_KEY)

# Initialize embedding model
embedding_model = SentenceTransformer('all-MiniLM-L6-v2')  # Open-source model

# Initialize FAISS index
dimension = 384  # Embedding size of the model
index = faiss.IndexFlatL2(dimension)

# Helper Functions
def extract_text_from_pdfs(files):
    """Extract text from multiple PDF files."""
    all_text = ""
    for file in files:
        reader = PdfReader(file)
        for page in reader.pages:
            all_text += page.extract_text()
    return all_text

def create_chunks(text, chunk_size=500):
    """Split text into chunks of a specified size."""
    words = text.split()
    return [' '.join(words[i:i + chunk_size]) for i in range(0, len(words), chunk_size)]

def generate_embeddings(chunks):
    """Generate embeddings for the given chunks."""
    return embedding_model.encode(chunks, convert_to_numpy=True)

def query_groq(prompt):
    """Query the Groq model for a response."""
    response = client.chat.completions.create(
        messages=[{"role": "user", "content": prompt}],
        model="llama3-8b-8192"
    )
    return response.choices[0].message.content

def visualize_comparative_insights(methods, results):
    """Generate a scatter plot of methods vs. results."""
    plt.figure(figsize=(8, 6))
    plt.scatter(methods, results, c='blue', alpha=0.7)
    plt.title("Methods vs. Results")
    plt.xlabel("Methods")
    plt.ylabel("Results")
    st.pyplot(plt)

# Streamlit App
st.title("Research Paper Analyzer")
st.write("Upload research papers, ask questions, and gain AI-driven insights!")

# Step 1: PDF Upload
uploaded_files = st.file_uploader("Upload your PDF files", type=["pdf"], accept_multiple_files=True)
if uploaded_files:
    with st.spinner("Processing the PDFs..."):
        # Extract text
        pdf_text = extract_text_from_pdfs(uploaded_files)
        st.success("PDFs processed successfully!")

        # Step 2: Chunking
        chunks = create_chunks(pdf_text)
        st.info(f"Documents split into {len(chunks)} chunks.")

        # Step 3: Embedding Creation
        embeddings = generate_embeddings(chunks)
        index.add(np.array(embeddings))
        st.success("Embeddings stored in FAISS database.")

        # Step 4: Query
        user_query = st.text_input("Ask a question:")
        if user_query:
            with st.spinner("Searching and generating a response..."):
                # Embed user query
                query_embedding = embedding_model.encode([user_query], convert_to_numpy=True)

                # Search in FAISS
                distances, indices = index.search(np.array(query_embedding), k=5)
                relevant_chunks = [chunks[i] for i in indices[0]]

                # Combine retrieved chunks as context
                context = " ".join(relevant_chunks)

                # Query Groq model
                prompt = f"Context: {context}\n\nQuestion: {user_query}\n\nAnswer:"
                answer = query_groq(prompt)

                # Display response
                st.write("### Answer:")
                st.write(answer)

        # Step 5: Comparative Insights
        if st.button("Generate Comparative Insights"):
            # Example data for visualization
            methods = [1, 2, 3, 4, 5]  # Replace with actual methods data
            results = [3.2, 4.1, 5.6, 4.8, 6.0]  # Replace with actual results data
            visualize_comparative_insights(methods, results)

        # Step 6: Bibliography Suggestions
        if st.button("Suggest Related Papers"):
            related_papers = ["Paper A", "Paper B", "Paper C"]  # Replace with actual suggestions
            st.write("### Suggested Papers:")
            for paper in related_papers:
                st.write(f"- {paper}")