app.py

import gradio as gr
from supertonic import TTS
import os
import numpy as np

try:
    tts = TTS(auto_download=True)
except Exception as e:
    print(f"Error initializing TTS: {e}")

VOICES = [
  "F6-zh", "F7-gb", "F8-gb", "mb4-id", # custom
  "M1", "M2", "M3", "M4", "M5",
  "F1", "F2", "F3", "F4", "F5",
]
CUSTOM_VOICES = VOICES[:4] # Changed to :4 to include "mb4-id"
CUSTOM_VOICES_DIR = "my_voice_styles"
# The 31 supported languages
LANGUAGES = {
    "English": "en", "Korean": "ko", "Japanese": "ja", "Arabic": "ar",
    "Bulgarian": "bg", "Czech": "cs", "Danish": "da", "German": "de",
    "Greek": "el", "Spanish": "es", "Estonian": "et", "Finnish": "fi",
    "French": "fr", "Hindi": "hi", "Croatian": "hr", "Hungarian": "hu",
    "Indonesian": "id", "Italian": "it", "Lithuanian": "lt", "Latvian": "lv",
    "Dutch": "nl", "Polish": "pl", "Portuguese": "pt", "Romanian": "ro",
    "Russian": "ru", "Slovak": "sk", "Slovenian": "sl", "Swedish": "sv",
    "Turkish": "tr", "Ukrainian": "uk", "Vietnamese": "vi"
}

def generate_speech(text, voice, language_name):
    if not text.strip():
        raise gr.Error("Please enter some text.")
        
    try:
        lang_code = LANGUAGES[language_name]
        
        # Get the voice style object
        if voice in CUSTOM_VOICES:
          style_path = os.path.join(CUSTOM_VOICES_DIR, f"{voice}.json")
          style = tts.get_voice_style_from_path(style_path)
        else:
          style = tts.get_voice_style(voice_name=voice)
        
        # Synthesize (Returns wav data and a numpy array for duration)
        wav, duration = tts.synthesize(text, voice_style=style, lang=lang_code, speed=0.95)
        
        # Save to a temporary path
        output_path = "output.wav"
        tts.save_audio(wav, output_path)
        
        # FIX: Convert numpy.ndarray duration to float for f-string compatibility
        readable_duration = float(duration[0])
        
        return output_path, f"Generation Successful! \nDuration: {readable_duration:.2f}s"
    
    except Exception as e:
        raise gr.Error(f"Generation failed: {str(e)}")

# Define the Gradio Interface
with gr.Blocks(theme='soft', title="Supertonic-3 Custom Voices") as demo:
    gr.Markdown("# 🔊 Supertonic-3 Custom Voices")
    gr.Markdown("""An on-device, lightweight Text-to-Speech system by **[Supertone](https://huggingface.co/Supertone/supertonic-3)**. Running on CPU via ONNX.
    You can create custom voice style from **[HERE](https://github.com/saurabhv749/supertonic3-voice-clone)**""")
    
    with gr.Row():
        with gr.Column(scale=1):
            input_text = gr.Textbox(
                label="Input Text", 
                placeholder="Type your message here...",
                value="In the same track, he reveals his father has been diagnosed with cancer.",
                lines=4
            )
            
            with gr.Row():
                voice_opt = gr.Dropdown(
                    choices=VOICES, 
                    value="F8-gb", 
                    label="Voice Style"
                )
                lang_opt = gr.Dropdown(
                    choices=sorted(list(LANGUAGES.keys())), 
                    value="English", 
                    label="Language"
                )
            
            btn = gr.Button("Synthesize Speech", variant="primary")
            
        with gr.Column(scale=1):
            audio_output = gr.Audio(label="Synthesized Audio", type="filepath")
            status_box = gr.Textbox(label="Status", interactive=False)

    # Setup Examples
    gr.Examples(
        examples=[
            ["Hello! This is a test of the Supertonic 3 system with custom voices, running locally.", "F7-gb", "English"],
            ["こんにちは、これは日本語の音声合成のテストです。", "F1", "Japanese"],
            ["पिछले कुछ सालों में फ़ीवर के इलाज के लिए कई नई, और पहले से ज़्यादा असरदार दवाएं उपलब्ध हुई हैं.", "F6-zh", "Hindi"],
            ["C'est un plaisir de vous rencontrer.", "M4", "French"],
            ["¡Hola! Esta es una prueba de voz en español.", "F8-gb", "Spanish"],
        ],
        inputs=[input_text, voice_opt, lang_opt]
    )

    btn.click(
        fn=generate_speech, 
        inputs=[input_text, voice_opt, lang_opt], 
        outputs=[audio_output, status_box]
    )

if __name__ == "__main__":
    demo.launch()