app.py

import os
import streamlit as st
from dotenv import load_dotenv
import base64
import requests
import time

# Load environment variables from the .env file
load_dotenv()

# Set your Hugging Face API token from the environment variable
HUGGINGFACE_API_TOKEN = os.getenv("HUGGINGFACE_API_TOKEN")

# Translation API URL (using LibreTranslate as an example)
TRANSLATION_API_URL = "https://libretranslate.com/translate"

# Function to get user input
def get_text():
    input_text = st.text_input("Enter text for speech generation:", key="input")
    return input_text

# Function to select the language model and target language for translation
def select_language():
    language_options = {
        "English": "en",
        "Spanish": "es",
        "French": "fr",
        "German": "de",
        "Italian": "it",
    }
    selected_language = st.selectbox("Select Language", list(language_options.keys()))
    return language_options[selected_language]

# Function to translate text to English (using LibreTranslate API as an example)
def translate_text(text, target_language):
    if target_language != "en":
        payload = {
            "q": text,
            "source": target_language,
            "target": "en",
        }
        response = requests.post(TRANSLATION_API_URL, json=payload)
        if response.status_code == 200:
            return response.json().get("translatedText", "")
        else:
            st.error(f"Error: {response.status_code} - Could not translate text.")
            return text
    return text  # If English, return the same text

# Function to send a request to the Hugging Face API with retry on 503 and error handling for 500
def generate_speech(model_id, text):
    url = f"https://api-inference.huggingface.co/models/{model_id}"
    headers = {"Authorization": f"Bearer {HUGGINGFACE_API_TOKEN}"}
    payload = {"inputs": text}

    while True:
        response = requests.post(url, headers=headers, json=payload, stream=True)

        if response.status_code == 200:
            return response.content  # Return the audio data if successful
        elif response.status_code == 503:
            # Parse the estimated wait time from the response and wait
            response_data = response.json()
            estimated_time = response_data.get("estimated_time", 10)
            st.info(f"Model is loading, please wait {estimated_time} seconds...")
            time.sleep(estimated_time)  # Wait before retrying
        elif response.status_code == 500:
            st.error("Internal Server Error: The model encountered an issue. Please try again later.")
            return None
        else:
            st.error(f"Error: {response.status_code} - {response.text}")
            return None

# Function to play audio in Streamlit
def play_audio(audio_bytes):
    b64 = base64.b64encode(audio_bytes).decode()
    audio_html = f"""
<audio controls autoplay>
<source src="data:audio/wav;base64,{b64}" type="audio/wav">
        Your browser does not support the audio element.
</audio>
    """
    st.markdown(audio_html, unsafe_allow_html=True)

# Initialize Streamlit UI
st.set_page_config(page_title="Multilingual Text-to-Speech", page_icon="🔊")
st.header("Multilingual Text-to-Speech Demo")

# Get user input and language selection
user_input = get_text()
selected_language = select_language()

# Create a button for generating speech
submit = st.button('Generate Speech')

# If the generate button is clicked and user input is not empty
if submit and user_input:
    with st.spinner("Translating and generating speech..."):
        # Translate the text to English if needed
        translated_text = translate_text(user_input, selected_language)
        
        # Display both original and translated text
        if selected_language != "en":
            st.write(f"Original Text: {user_input}")
            st.write(f"Translated Text: {translated_text}")

        # Generate speech with retry if the model is loading and handle 500 errors
        audio_data = generate_speech("myshell-ai/MeloTTS-English", translated_text)

        if audio_data:
            play_audio(audio_data)
elif submit:
    st.warning("Please enter some text.")  # Warning for empty input