app.py

# Highlight changes and categorize small and large differences
def highlight_changes(img, mask):
    overlay = img.copy()
    contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    summary = []
    for cnt in contours:
        area = cv2.contourArea(cnt)
        x, y, w, h = cv2.boundingRect(cnt)
        if area > 500:  # Major differences
            cv2.rectangle(overlay, (x, y), (x + w, y + h), (0, 0, 255), 2)  # Red
            summary.append(f"Major Difference: Location=({x}, {y}), Size=({w}x{h}), Area={area}")
        elif 100 < area <= 500:  # Small differences
            cv2.rectangle(overlay, (x, y), (x + w, y + h), (255, 0, 0), 2)  # Blue
            summary.append(f"Small Difference: Location=({x}, {y}), Size=({w}x{h}), Area={area}")
    return overlay, summary

# Generate comparison PDF with detailed summary
def generate_comparison_pdf(original_pdf, edited_pdf):
    original_images = convert_pdf_to_images(original_pdf)
    edited_images = convert_pdf_to_images(edited_pdf)
    combined_images = []
    all_summaries = []

    for page_num, (orig_img, edit_img) in enumerate(zip(original_images, edited_images), start=1):
        aligned_img = align_images(orig_img, edit_img)
        diff_mask = compare_images(orig_img, aligned_img)
        highlighted_img, summary = highlight_changes(edit_img, diff_mask)

        # Add page number to summary
        page_summary = [f"Page {page_num}:"]
        page_summary.extend(summary)
        all_summaries.extend(page_summary)

        # Ensure dimensions match
        height = min(orig_img.shape[0], highlighted_img.shape[0])
        orig_img_resized = orig_img[:height]
        highlighted_img_resized = highlighted_img[:height]
        combined_images.append(np.hstack((orig_img_resized, highlighted_img_resized)))

    # Generate the PDF
    output_path = "outputs/comparison_result.pdf"
    pdf = FPDF()
    for img in combined_images:
        temp_path = "temp_image.png"
        cv2.imwrite(temp_path, img)
        pdf.add_page()
        pdf.image(temp_path, x=10, y=10, w=190)
        os.remove(temp_path)

    # Add detailed summary to the PDF
    summary_path = "outputs/summary.txt"
    with open(summary_path, "w") as f:
        f.write("\n".join(all_summaries))
    pdf.add_page()
    pdf.set_font("Arial", size=12)
    pdf.multi_cell(0, 10, "\n".join(all_summaries))

    pdf.output(output_path)
    return output_path, summary_path

# Gradio interface function
def pdf_comparison(original_pdf, edited_pdf):
    # Get the file size in bytes
    original_file_size = os.path.getsize(original_pdf.name)
    edited_file_size = os.path.getsize(edited_pdf.name)

    # Check if either file exceeds 50 MB (50 * 1024 * 1024 bytes)
    if original_file_size > 50 * 1024 * 1024 or edited_file_size > 50 * 1024 * 1024:
        return "Error: File size exceeds 50 MB. Please upload smaller files."

    # Proceed with PDF comparison
    result_path, summary_path = generate_comparison_pdf(original_pdf.name, edited_pdf.name)
    return result_path, summary_path

# Gradio interface
interface = gr.Interface(
    fn=pdf_comparison,
    inputs=[
        gr.File(label="Upload Original PDF", file_types=[".pdf"]),
        gr.File(label="Upload Edited PDF", file_types=[".pdf"])
    ],
    outputs=[
        gr.File(label="Download Comparison Report"),
        gr.File(label="Download Detailed Summary")
    ],
)

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