import gradio print("Gradio version:", gradio.__version__) import re import pytesseract from pdfminer.high_level import extract_text from PIL import Image from transformers import pipeline import gradio as gr # Load Summarization Model summarizer = pipeline("summarization", model="facebook/bart-large-cnn") def clean_text(text): text = re.sub(r'http[s]?://\S+', '', text) text = re.sub(r'www\.\S+', '', text) text = re.sub(r'\n+', ' ', text).strip() return text def extract_text_from_pdf(pdf_file): text = extract_text(pdf_file.name) return clean_text(text) def extract_text_from_image(image_file): image = Image.open(image_file) text = pytesseract.image_to_string(image) return clean_text(text) def chunk_text(text, chunk_size=1000): return [text[i:i+chunk_size] for i in range(0, len(text), chunk_size)] def summarize_text(text, max_length=200, min_length=50): chunks = chunk_text(text) summaries = [] for chunk in chunks: try: summary = summarizer(chunk, max_length=max_length, min_length=min_length, do_sample=False) summaries.append(summary[0]['summary_text']) except Exception as e: print(f"Error summarizing chunk: {e}") return " ".join(summaries) def gradio_handle_input(input_type, text_input, image_input, pdf_input): if input_type == "Text": text = text_input elif input_type == "Image": if image_input is None: return "", "Please upload an image." text = extract_text_from_image(image_input) elif input_type == "Pdf": if pdf_input is None: return "", "Please upload a PDF file." text = extract_text_from_pdf(pdf_input) else: return "", "Invalid input type." text = clean_text(text) if not text.strip(): return "", "Error: extracted text is empty." summary = summarize_text(text) return text, summary def toggle_inputs(choice): return ( gr.update(visible=choice == "Text", value=""), gr.update(visible=choice == "Image"), gr.update(visible=choice == "Pdf"), gr.update(value=""), # Clear preview if needed "", "" ) def download_summary(summary): if not summary.strip(): return gr.update(visible=False), None file_path = "summary.txt" with open(file_path, "w", encoding="utf-8") as f: f.write(summary) return gr.update(visible=True, value=file_path) with gr.Blocks(css="footer {visibility: hidden;}", title="Legal Document Summarization") as demo: gr.Markdown("## 📄 Legal Document Summarization") gr.Markdown(""" ### 📘 Steps to Use: 1. Choose input type (✍️ Text / 📄 PDF / 🖼️ Image) 2. Enter your text or upload a file 3. Click **📝 Summarize** 4. View original input and ✂️ summary below """) with gr.Row(): with gr.Column(scale=1): input_type = gr.Radio(choices=["Text", "Pdf", "Image"], label="Select Input Type", value="Text") text_input = gr.Textbox(label="✍️ Enter Text", lines=15) image_input = gr.Image(label="📷 Upload Image", visible=False, type="filepath") pdf_input = gr.File(label="📄 Upload PDF", file_types=[".pdf"], visible=False) submit_btn = gr.Button("📝 Summarize") with gr.Column(scale=1): gr.Markdown("⏳ *Note: Summarization may take some time depending on file size and system performance.*") gr.Markdown("🔍 **Summary Output**") output = gr.Textbox(label="", lines=15, interactive=False) download_btn = gr.Button("⬇️ Download Summary") file_output = gr.File(label="Download Summary", visible=False) gr.Markdown(""" > ⚠️ **Disclaimer:** This tool provides automatic summarizations of legal documents **for educational purposes only.** > It does **not** constitute legal advice or professional consultation. > Please consult a qualified legal professional for any legal matters. """) input_type.change( fn=toggle_inputs, inputs=input_type, outputs=[text_input, image_input, pdf_input, output, text_input, output] ) submit_btn.click( fn=gradio_handle_input, inputs=[input_type, text_input, image_input, pdf_input], outputs=[text_input, output] ) download_btn.click( fn=download_summary, inputs=output, outputs=file_output ) if __name__ == "__main__": demo.launch()