app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from PyPDF2 import PdfReader
from fpdf import FPDF
from deep_translator import GoogleTranslator
import tempfile

# Load the tokenizer and model for summarization
tokenizer = AutoTokenizer.from_pretrained("nsi319/legal-led-base-16384")
model = AutoModelForSeq2SeqLM.from_pretrained("nsi319/legal-led-base-16384")


# Function to extract text from PDF
def extract_text_from_pdf(pdf_file):
    reader = PdfReader(pdf_file)
    text = ''
    for page in reader.pages:
        text += page.extract_text()
    return text


# Function to summarize text
def summarize_text(text, max_input_length=16384, max_summary_length=512):
    text = ' '.join(text.split())
    inputs = tokenizer(text, return_tensors="pt", max_length=max_input_length, truncation=True)

    summary_ids = model.generate(
        inputs["input_ids"],
        max_length=max_summary_length,
        min_length=150,
        length_penalty=2.0,
        num_beams=5,
        early_stopping=True,
        no_repeat_ngram_size=3
    )

    summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
    return summary


# Function to save summary to PDF
def save_summary_to_pdf(summary_text, output_pdf_path):
    pdf = FPDF()
    pdf.set_auto_page_break(auto=True, margin=15)
    pdf.add_page()
    pdf.set_font("Arial", size=12)
    pdf.set_font("Arial", 'B', 16)
    pdf.cell(200, 10, txt="Legal Document Summary", ln=True, align="C")
    pdf.set_font("Arial", size=12)
    lines = summary_text.splitlines()

    for line in lines:
        pdf.multi_cell(0, 10, line)

    pdf.output(output_pdf_path)
    return output_pdf_path


# Function to process PDF for summarization
def process_pdf_summary(pdf_file):
    pdf_text = extract_text_from_pdf(pdf_file)
    summary = summarize_text(pdf_text)
    output_pdf_path = "legal_document_summary.pdf"
    return save_summary_to_pdf(summary, output_pdf_path)


# Function to translate text
def translate_text(text, language):
    translation = GoogleTranslator(source='auto', target=language).translate(text)
    return translation


# Function to create a PDF from translated text
def create_pdf(text):
    pdf_output = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
    pdf = FPDF()
    pdf.add_page()
    pdf.set_font("Arial", size=12)  # Use Arial instead of Latha

    for line in text.split('\n'):
        pdf.cell(200, 10, txt=line.encode('latin-1', 'replace').decode('latin-1'), ln=True)

    pdf.output(pdf_output)
    return pdf_output


# Function to process PDF for translation
def process_pdf_translation(pdf_file, language):
    extracted_text = extract_text_from_pdf(pdf_file)
    translated_text = translate_text(extracted_text, language)
    translated_pdf = create_pdf(translated_text)
    return translated_pdf, translated_text


# Create Gradio interface
with gr.Blocks() as app:
    gr.Markdown("# Legal Document Translator and Summarizer")
    gr.Markdown("Choose an operation to perform on your document:")

    with gr.Tab("Summarization"):
        gr.Markdown("### Upload PDF for Summarization")
        pdf_input_summary = gr.File(label="Upload PDF Document")
        summary_output = gr.File(label="Download Summary PDF")
        summarize_button = gr.Button("Summarize")

        summarize_button.click(process_pdf_summary, inputs=pdf_input_summary, outputs=summary_output)

    with gr.Tab("Translation"):
        gr.Markdown("### Upload PDF for Translation")
        pdf_input_translation = gr.File(label="Upload PDF Document")
        language_options = ["hi", "ta", "ml", "en"]  # Hindi, Tamil, Malayalam, English
        language_selector = gr.Dropdown(choices=language_options, label="Select Language", value="hi")
        translation_output = gr.File(label="Download Translated PDF")
        translated_text_output = gr.Textbox(label="Translated Text", lines=10)
        translate_button = gr.Button("Translate")

        translate_button.click(process_pdf_translation, inputs=[pdf_input_translation, language_selector],
                               outputs=[translation_output, translated_text_output])

# Launch the app
if __name__ == "__main__":
    app.launch()