import gradio as gr
import pdfplumber
import openai
import re
import unicodedata
import os

# Set up OpenAI API Key (Replace with your actual key)
openai.api_key = "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"


def clean_text(text):
    """Cleans extracted text for better processing by the model."""
    text = unicodedata.normalize("NFKC", text)  # Normalize Unicode characters
    text = re.sub(r'\s+', ' ', text).strip()  # Remove extra spaces and newlines
    text = re.sub(r'[^a-zA-Z0-9.,!?;:\'\"()\-]', ' ', text)  # Keep basic punctuation
    text = re.sub(r'(?i)(page\s*\d+)', '', text)  # Remove page numbers
    return text

def extract_text_from_pdf(pdf_file):
    """Extract and clean text from the uploaded PDF."""
    try:
        with pdfplumber.open(pdf_file) as pdf:
            text = " ".join(clean_text(text) for page in pdf.pages if (text := page.extract_text()))
        return text
    except Exception as e:
        print(f"Error extracting text: {e}")
        return None

def split_text(text, chunk_size=500):
    """Splits text into smaller chunks for faster processing."""
    chunks = []
    for i in range(0, len(text), chunk_size):
        chunks.append(text[i:i+chunk_size])
    return chunks

def chatbot(pdf_file, user_question):
    """Processes the PDF and answers the user's question."""
    
    # Step 1: Extract text from the PDF
    text = extract_text_from_pdf(pdf_file)
    
    # Step 2: Split into chunks
    chunks = split_text(text)
    
    # Step 3: Use only the first chunk for now (to reduce token usage)
    if not chunks:
        return "Could not extract any text from the PDF."

    prompt = f"Based on this document, answer the question:\n\nDocument:\n{chunks[0]}\n\nQuestion: {user_question}"

    # Step 4: Send to OpenAI's GPT-3.5
    response = openai.ChatCompletion.create(
        model="gpt-3.5-turbo",
        messages=[{"role": "user", "content": prompt}]
    )

    # Step 5: Return the chatbot's response
    return response["choices"][0]["message"]["content"]

# Gradio Interface
iface = gr.Interface(
    fn=chatbot,
    inputs=[gr.File(label="Upload PDF"), gr.Textbox(label="Ask a Question")],
    outputs=gr.Textbox(label="Answer"),
    title="PDF Q&A Chatbot"
)

# Launch Gradio app
iface.launch()