app.py

# app.py

from pypdf import PdfReader
import gradio as gr
from sentence_transformers import SentenceTransformer
import faiss
import numpy as np

# Load embedding model
model = SentenceTransformer('all-MiniLM-L6-v2')

# Global state to persist embeddings and chunks
index = None
chunks = []

# Step 1: Extract text from uploaded PDFs
def extract_text_from_pdfs(files):
    all_text = ""
    for file in files:
        reader = PdfReader(file.name)
        for page in reader.pages:
            text = page.extract_text()
            if text:
                all_text += text + "\n"
    return all_text

# Step 2: Chunk text
def chunk_text(text, chunk_size=500, overlap=50):
    words = text.split()
    result = []
    for i in range(0, len(words), chunk_size - overlap):
        chunk = " ".join(words[i:i + chunk_size])
        result.append(chunk)
    return result

# Step 3: Embed and store chunks
def create_index(text_chunks):
    global index, chunks
    chunks = text_chunks
    embeddings = model.encode(chunks)
    index = faiss.IndexFlatL2(len(embeddings[0]))
    index.add(np.array(embeddings))

# Step 4: Retrieve top relevant chunks
def get_top_chunks(query, k=3):
    query_vec = model.encode([query])
    D, I = index.search(np.array(query_vec), k)
    return [chunks[i] for i in I[0]]

# Step 5: Fake LLM response (replace with real API call if needed)
def call_llm(context, question):
    return f"Answer (simulated): Based on context:\n\n{context}\n\nQuestion: {question}"

# Step 6: Gradio main function
def rag_pipeline(files, question):
    text = extract_text_from_pdfs(files)
    text_chunks = chunk_text(text)
    create_index(text_chunks)
    top_chunks = get_top_chunks(question)
    context = "\n".join(top_chunks)
    answer = call_llm(context, question)
    return answer

# Step 7: Gradio UI
demo = gr.Interface(
    fn=rag_pipeline,
    inputs=[
        gr.File(file_types=[".pdf"], file_count="multiple", label="Upload PDFs"),
        gr.Textbox(lines=2, label="Ask a question")
    ],
    outputs="text",
    title="RAG PDF Chatbot",
    description="Upload PDFs and ask questions based on their content"
)

if __name__ == "__main__":
    demo.launch()