app.py

import gradio as gr
from ai4bharat.transliteration import XlitEngine
import torch
import re
from transformers import MarianMTModel, MarianTokenizer

# Initialize Transliteration Engine
engine = XlitEngine("te")

# Load Model & Tokenizer
model_folder_path = "HackHedron/N_Te-N_En-Translator"  # Your model on Hugging Face
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
tokenizer = MarianTokenizer.from_pretrained(model_folder_path)
model = MarianMTModel.from_pretrained(model_folder_path).to(device)

# Text Cleaning Function
def clean_text(text):
    text = text.lower().strip()
    numbers = re.findall(r'\d+', text)
    text = re.sub(r'\d+', '<>', text)
    text = re.sub(r'[.!?]', '', text)
    text = re.sub(r'["\'“”‘’]', '', text)
    text = re.sub(r'\s+', ' ', text)
    return text, numbers

# Transliterate to Telugu Script
def transliterate_to_native_script(text):
    words = text.split()
    transliterated_words = [engine.translit_word(word, topk=1)["te"][0] for word in words]
    return " ".join(transliterated_words)

# Translate to English
def translate_input(user_input):
    cleaned, numbers = clean_text(user_input)
    telugu_script = transliterate_to_native_script(cleaned)
    inputs = tokenizer(telugu_script, return_tensors="pt", max_length=128, truncation=True).to(device)
    output_ids = model.generate(inputs["input_ids"], max_length=128, num_beams=5, early_stopping=True)
    translated = tokenizer.decode(output_ids[0], skip_special_tokens=True)

    def restore_numbers(text, numbers):
        placeholder_pattern = re.compile(r'(<<>>|<>|<<|>>|<|>)')
        for num in numbers:
            match = placeholder_pattern.search(text)
            if match:
                start, end = match.span()
                text = text[:start] + num + text[end:]
        return text

    final_output = restore_numbers(translated, numbers)
    return user_input, cleaned, telugu_script, final_output

# Gradio Interface
interface = gr.Interface(
    fn=translate_input,
    inputs=gr.Textbox(label="Enter Romanized Telugu Sentence"),
    outputs=[
        gr.Textbox(label="Original Input"),
        gr.Textbox(label="Cleaned Text"),
        gr.Textbox(label="Telugu Script"),
        gr.Textbox(label="Translated English"),
    ],
    title="Romanized Telugu to English Translator",
    description="Transliterates Romanized Telugu to Telugu script and translates it to English."
)

interface.launch(share=True)