app.py

import streamlit as st
from PIL import Image
import os
from ocr_tamil.ocr import OCR
from deep_translator import GoogleTranslator

# Load OCR models
ocr_detect = OCR(detect=True, enable_cuda=False)
ocr_recognize = OCR(detect=False, enable_cuda=False)

# Language code mapping
lang_codes = {
    'English': 'en',
    'Kannada': 'kn',
    'Tamil': 'ta',
    'Telugu': 'te',
    'Hindi': 'hi',
    'Malayalam': 'ml',
    'Marathi': 'mr',
    'Bengali': 'bn',
    'Gujarati': 'gu',
    'Urdu': 'ur',
}

# Prediction function
def predict(image_path, target_language):
    try:
        texts = ocr_detect.predict(image_path)
        texts = [" ".join(texts[0])]
        detected_text = texts[0]

        translated_text = GoogleTranslator(source='auto', target=target_language).translate(detected_text)
        return detected_text, translated_text
    except Exception as e:
        return f"Error: {str(e)}", ""

# Streamlit UI
st.title("Tamil OCR + Translation")

uploaded_file = st.file_uploader("Upload a Tamil handwritten image", type=["jpg", "jpeg", "png"])

target_lang_name = st.selectbox("Target Language", list(lang_codes.keys()))

if uploaded_file:
    image = Image.open(uploaded_file)
    st.image(image, caption="Uploaded Image", use_column_width=True)
    image_path = "uploaded_image.jpg"
    image.save(image_path)

    if st.button("Submit"):
        detected, translated = predict(image_path, lang_codes[target_lang_name])
        st.success("Detected Text:")
        st.write(detected)
        st.success("Translated Text:")
        st.write(translated)