src/streamlit_app.py

import streamlit as st
import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from langdetect import detect
# ==============================
# 🔐 HIDE STREAMLIT MENU (SOLUTION 2)
# ==============================
st.markdown("""
<style>
#MainMenu {visibility: hidden;}
header {visibility: hidden;}
footer {visibility: hidden;}
.stDeployButton {display:none;}
</style>
""", unsafe_allow_html=True)

# ==============================
# PAGE CONFIG
# ==============================
st.set_page_config(page_title="🌍 Multilingual Translator", layout="wide")

# ==============================
# LOAD MODEL (CACHED)
# ==============================
@st.cache_resource
def load_model():
    tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
    model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
    return tokenizer, model

tokenizer, model = load_model()

# ==============================
# LANGUAGE MAP (50+ LANGUAGES)
# ==============================
lang_map = {
    # Indian Languages
    "English": "eng_Latn",
    "Tamil": "tam_Taml",
    "Hindi": "hin_Deva",
    "Telugu": "tel_Telu",
    "Kannada": "kan_Knda",
    "Malayalam": "mal_Mlym",
    "Marathi": "mar_Deva",
    "Bengali": "ben_Beng",
    "Gujarati": "guj_Gujr",
    "Punjabi": "pan_Guru",
    "Urdu": "urd_Arab",

    # European
    "French": "fra_Latn",
    "Spanish": "spa_Latn",
    "German": "deu_Latn",
    "Italian": "ita_Latn",
    "Portuguese": "por_Latn",
    "Dutch": "nld_Latn",
    "Romanian": "ron_Latn",
    "Czech": "ces_Latn",
    "Polish": "pol_Latn",
    "Hungarian": "hun_Latn",
    "Finnish": "fin_Latn",
    "Swedish": "swe_Latn",
    "Norwegian": "nob_Latn",
    "Danish": "dan_Latn",

    # Middle East
    "Arabic": "arb_Arab",
    "Persian": "pes_Arab",
    "Hebrew": "heb_Hebr",
    "Turkish": "tur_Latn",

    # Asian
    "Chinese (Simplified)": "zho_Hans",
    "Chinese (Traditional)": "zho_Hant",
    "Japanese": "jpn_Jpan",
    "Korean": "kor_Hang",
    "Thai": "tha_Thai",
    "Vietnamese": "vie_Latn",
    "Indonesian": "ind_Latn",

    # Others
    "Russian": "rus_Cyrl",
    "Ukrainian": "ukr_Cyrl",
    "Bulgarian": "bul_Cyrl",
    "Greek": "ell_Grek",
    "Swahili": "swh_Latn",
    "Amharic": "amh_Ethi"
}

# ==============================
# DETECT LANGUAGE
# ==============================
detect_map = {
    "en": "eng_Latn",
    "ta": "tam_Taml",
    "hi": "hin_Deva",
    "te": "tel_Telu",
    "kn": "kan_Knda",
    "ml": "mal_Mlym",
    "mr": "mar_Deva",
    "bn": "ben_Beng",
    "gu": "guj_Gujr",
    "pa": "pan_Guru",
    "ur": "urd_Arab",
    "fr": "fra_Latn",
    "es": "spa_Latn",
    "de": "deu_Latn",
    "it": "ita_Latn",
    "pt": "por_Latn",
    "nl": "nld_Latn",
    "ro": "ron_Latn",
    "ru": "rus_Cyrl",
    "zh-cn": "zho_Hans",
    "zh-tw": "zho_Hant",
    "ja": "jpn_Jpan",
    "ko": "kor_Hang",
    "ar": "arb_Arab",
    "tr": "tur_Latn",
    "vi": "vie_Latn"
}

def detect_language(text):
    if len(text.strip()) < 3:
        return "eng_Latn"
    try:
        lang = detect(text)
    except:
        lang = "en"
    return detect_map.get(lang, "eng_Latn")

# ==============================
# TRANSLATION FUNCTION
# ==============================
def translate(text, target_lang):

    if len(text.strip()) < 2:
        return "⚠️ Please enter valid text.", "N/A"

    src_lang = detect_language(text)
    tgt_lang = lang_map[target_lang]

    tokenizer.src_lang = src_lang

    encoded = tokenizer(text, return_tensors="pt")

    generated_tokens = model.generate(
        **encoded,
        forced_bos_token_id=tokenizer.convert_tokens_to_ids(tgt_lang),
        max_length=200,
        num_beams=4,   # improves accuracy
        early_stopping=True
    )

    translated = tokenizer.batch_decode(
        generated_tokens,
        skip_special_tokens=True
    )[0]

    return translated, src_lang

# ==============================
# UI
# ==============================
st.title("🌍 40+ AI Powered-Multilingual Translators App")

col1, col2 = st.columns(2)

with col1:
    input_text = st.text_area("Enter Text", height=200)

with col2:
    target_lang = st.selectbox("Select Target Language", list(lang_map.keys()))

if st.button("Translate"):
    if not input_text.strip():
        st.warning("Please enter text")
    else:
        with st.spinner("Translating..."):
            output, src_lang = translate(input_text, target_lang)

            st.success("✅ Translation")
            st.write(output)

            st.info(f"Detected Language Code: {src_lang}")