app.py

import gradio as gr
import cv2
import numpy as np
from PIL import Image
import pytesseract
from transformers import pipeline

# -------------------------------
# Load Summarization Model
# -------------------------------

def load_model():
    return pipeline("summarization", model="facebook/bart-large-cnn")

summarizer = load_model()

# -------------------------------
# Image Enhancement
# -------------------------------

def enhance_image(image):
    img = np.array(image)
    img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)

    enhanced = cv2.convertScaleAbs(img, alpha=1.5, beta=30)
    blurred = cv2.GaussianBlur(enhanced, (5, 5), 0)
    gray = cv2.cvtColor(blurred, cv2.COLOR_BGR2GRAY)

    thresh = cv2.adaptiveThreshold(
        gray, 255,
        cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
        cv2.THRESH_BINARY, 11, 2
    )
    return thresh

# -------------------------------
# OCR
# -------------------------------

def extract_text(image):
    return pytesseract.image_to_string(image)

# -------------------------------
# Main Pipeline
# -------------------------------

def process_document(input_image):
    if input_image is None:
        return None, "", ""

    enhanced = enhance_image(input_image)
    text = extract_text(enhanced)

    summary_text = "Text too short for summarization."

    if len(text) > 50:
        summary = summarizer(text[:1000], max_length=130,
                             min_length=30, do_sample=False)
        summary_text = summary[0]["summary_text"]

    return enhanced, text, summary_text

# -------------------------------
# Gradio UI
# -------------------------------

with gr.Blocks() as demo:

    gr.Markdown("# SmartDoc AI - Intelligent Document Processor")

    with gr.Row():
        input_img = gr.Image(type="pil", label="Upload Document Image")

    with gr.Row():
        enhanced_output = gr.Image(label="Enhanced Image")

    extracted_text = gr.Textbox(label="Extracted Text", lines=10)
    summary = gr.Textbox(label="Document Summary", lines=6)

    btn = gr.Button("Process Document")

    btn.click(
        fn=process_document,
        inputs=input_img,
        outputs=[enhanced_output, extracted_text, summary]
    )

demo.launch()