app.py

import gradio as gr
import os
import tempfile
from kokoro_engine import KokoroEngine
from processor import ScriptProcessor
import numpy as np

# Initialize components
engine = KokoroEngine()
processor = ScriptProcessor(engine)

def tts_process(text, voice, speed, lang, long_script_mode):
    try:
        if long_script_mode:
            audio, sr = processor.process_long_script(text, voice, speed, lang)
        else:
            audio, sr = engine.generate(text, voice, speed, lang)
        
        # Save to temp file
        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
            processor.save_audio(audio, sr, tmp.name)
            return tmp.name
    except Exception as e:
        return f"Error: {str(e)}"

def clone_process(audio_ref):
    if audio_ref is None:
        return "Please upload an audio file for cloning."
    # Placeholder for actual cloning logic
    voice_id = engine.clone_voice_placeholder(audio_ref)
    return f"Voice cloned successfully! Reference ID: {voice_id}. You can now use this voice (currently defaults to {voice_id})."

# Flatten voice list for dropdown
all_voices = []
for category, voices in engine.voices.items():
    for v in voices:
        all_voices.append(v)

# Premium CSS for high-end look
custom_css = """
.container { 
    max-width: 900px !important; 
    margin: auto !important; 
    padding-top: 2rem !important; 
}
.header { 
    text-align: center; 
    margin-bottom: 2rem; 
}
.header h1 { 
    font-size: 3rem !important; 
    font-weight: 800 !important; 
    background: linear-gradient(90deg, #ff00cc, #3333ff); 
    -webkit-background-clip: text; 
    -webkit-text-fill-color: transparent; 
    margin-bottom: 0.5rem !important; 
}
.header p { 
    font-size: 1.1rem !important; 
    color: #888; 
}
.input-group { 
    border-radius: 12px !important; 
    border: 1px solid #333 !important; 
    background: #111 !important; 
    padding: 1rem !important; 
    margin-bottom: 1.5rem !important; 
}
.footer { 
    visibility: hidden; 
}
button.primary { 
    background: linear-gradient(90deg, #ff00cc, #3333ff) !important; 
    border: none !important; 
    font-weight: bold !important; 
    border-radius: 8px !important; 
}
button.primary:hover { 
    transform: translateY(-2px); 
    box-shadow: 0 4px 15px rgba(255, 0, 204, 0.4); 
}
"""

with gr.Blocks(title="Kokoro TTS Premium") as demo:
    with gr.Column(elem_classes="container"):
        with gr.Column(elem_classes="header"):
            gr.Markdown("# 🌸 Kokoro TTS")
            gr.Markdown("High-fidelity neural speech synthesis powered by Kokoro-82M")

        with gr.Column(elem_classes="input-group"):
            text_input = gr.Textbox(
                label="Script Content", 
                placeholder="Paste your story, script, or text here...", 
                lines=10,
                elem_id="text-input"
            )
            
            with gr.Row():
                voice_select = gr.Dropdown(
                    choices=all_voices, 
                    value="af_heart", 
                    label="Voice Archetype", 
                    scale=2
                )
                speed_slider = gr.Slider(
                    minimum=0.5, 
                    maximum=2.0, 
                    value=1.0, 
                    step=0.1, 
                    label="Pacing (Speed)", 
                    scale=1
                )
            
            with gr.Accordion("⚙️ Engine Configurations", open=False):
                with gr.Row():
                    lang_select = gr.Dropdown(
                        choices=[
                            ("🇺🇸 English (US)", "en-us"), 
                            ("🇬🇧 English (UK)", "en-gb"), 
                            ("🇨🇳 Chinese", "zh"), 
                            ("🇮🇳 Hindi", "hi"), 
                            ("🇯🇵 Japanese", "ja"), 
                            ("🇪🇸 Spanish", "es"), 
                            ("🇫🇷 French", "fr"), 
                            ("🇮🇹 Italian", "it"), 
                            ("🇵🇹 Portuguese", "pt")
                        ], 
                        value="en-us", 
                        label="Linguistic Context"
                    )
                    long_script_toggle = gr.Checkbox(
                        label="Optimize for Long Duration (Safe Chunking)", 
                        value=False
                    )
        
        generate_btn = gr.Button("⚡ Generate Neural Audio", variant="primary", size="lg")
        
        with gr.Column(variant="compact"):
            audio_output = gr.Audio(
                label="Master Audio Output", 
                type="filepath"
            )

    generate_btn.click(
        tts_process, 
        inputs=[text_input, voice_select, speed_slider, lang_select, long_script_toggle], 
        outputs=audio_output
    )

if __name__ == "__main__":
    demo.launch(ssr_mode=False, css=custom_css)