import gradio as gr
import torch
from bark import SAMPLE_RATE, generate_audio
from scipy.io.wavfile import write as write_wav
import tempfile


def generate_voice(text_prompt, language):
    if not text_prompt.strip() or text_prompt.isnumeric():
        return None, "Please enter valid text."
    
    try:
        audio_array = generate_audio(text_prompt)
        audio_array = audio_array.cpu().numpy()
        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as fp:
            write_wav(fp.name, SAMPLE_RATE, audio_array)
            return fp.name, "Audio generated successfully!"
    except Exception as e:
        return None, f"Error: {str(e)}"


with gr.Blocks() as demo:
    gr.Markdown("# 🌐 Text to Speech Generator (Multi-Language Bark TTS GPU) 🚀")
    txt = gr.Textbox(label="Enter Text")
    lang = gr.Dropdown(
        choices=["English", "Tamil", "Hindi", "French"],
        label="Select Language"
    )
    btn = gr.Button("Generate Audio")
    audio = gr.Audio(label="Generated Audio", type="filepath")
    output_text = gr.Textbox(label="Status Message")

    btn.click(fn=generate_voice, inputs=[txt, lang], outputs=[audio, output_text])

demo.launch()