PDFuzz / gradio_demo.py
acmc's picture
Create gradio_demo.py
d485cda verified
raw
history blame
3.82 kB
#!/usr/bin/env python3
"""Simple Gradio demo for the PDF attacker tools
Allows entering text, choosing attack type, and downloading the generated PDF.
"""
import os
import time
from typing import Tuple
import PyPDF2
import gradio as gr
from pdf_attacker import PDFAttacker
attacker = PDFAttacker()
def _ensure_tmp_dir() -> str:
"""Ensure tmp dir exists and return its path"""
path = os.path.join(os.getcwd(), "tmp")
os.makedirs(path, exist_ok=True)
return path
def _extract_text_from_pdf(pdf_path: str) -> str:
"""Extract text from a PDF file for preview"""
try:
with open(pdf_path, 'rb') as f:
reader = PyPDF2.PdfReader(f)
text = ""
for page in reader.pages:
page_text = page.extract_text()
if page_text:
text += page_text
return text.strip()
except Exception as e:
return f"Error extracting text: {e}"
def generate_pdf(
text: str,
mode: str,
attack_factor: float = 0.7,
target_text: str = "",
) -> Tuple[str, str]:
"""Generate selected PDF and return (pdf_path, extracted_text)
Inputs: text, mode: 'normal'|'attacked'|'targeted', attack_factor, target_text
Outputs: path to generated PDF, extracted text preview
"""
tmp_dir = _ensure_tmp_dir()
timestamp = int(time.time() * 1000)
filename = f"{mode}_{timestamp}.pdf"
output_path = os.path.join(tmp_dir, filename)
# Clean input text
clean_text = " ".join(text.split())
try:
if mode == 'normal':
attacker.create_normal_pdf(text=clean_text, output_path=output_path)
elif mode == 'attacked':
attacker.create_attacked_pdf(text=clean_text, output_path=output_path, attack_factor=attack_factor)
elif mode == 'targeted':
# Targeted may raise ValueError if not feasible
attacker.create_targeted_pdf(text=clean_text, target_text=target_text, output_path=output_path)
else:
return "", f"Unknown mode: {mode}"
except Exception as e:
# Surface errors to the UI
return "", f"Error generating PDF: {e}"
# Extract text to show how the copied/extracted text looks
extracted = _extract_text_from_pdf(output_path)
return output_path, extracted
def build_demo():
"""Construct and return the Gradio Blocks demo"""
with gr.Blocks() as demo:
gr.Markdown("# PDF Humanizer: Attack demo\nGenerate PDFs that look normal but extract differently when copied")
with gr.Row():
txt = gr.Textbox(lines=8, label="Input text", value="Enter or paste text here...")
with gr.Column():
mode = gr.Radio(choices=['normal', 'attacked', 'targeted'], value='attacked', label='Mode')
attack_factor = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.7, label='Attack factor (attacked mode)')
target_text = gr.Textbox(lines=2, label='Target text (targeted mode)')
generate = gr.Button('Generate PDF')
download_file = gr.File(label='Download generated PDF')
extracted_preview = gr.Textbox(lines=8, label='Extracted text preview')
def _on_generate(text, mode, attack_factor, target_text):
path, extracted = generate_pdf(text=text, mode=mode, attack_factor=attack_factor, target_text=target_text)
if not path:
# Return empty file and error message in preview
return None, extracted
return path, extracted
generate.click(fn=_on_generate, inputs=[txt, mode, attack_factor, target_text], outputs=[download_file, extracted_preview])
return demo
if __name__ == '__main__':
app = build_demo()
app.launch(server_name='0.0.0.0', server_port=7860)