app.py

import gradio as gr
import torch
import soundfile as sf
import tempfile
import os
import numpy as np

os.environ["CUDA_VISIBLE_DEVICES"] = ""

print("Loading Qwen3-TTS Models...")

from qwen_tts import Qwen3TTSModel

print("Loading CustomVoice model...")
custom_model = Qwen3TTSModel.from_pretrained(
    "Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice",
    device_map="cpu",
    dtype=torch.float32,
)

print("Loading Base model...")
base_model = Qwen3TTSModel.from_pretrained(
    "Qwen/Qwen3-TTS-12Hz-0.6B-Base",
    device_map="cpu",
    dtype=torch.float32,
)

print("✅ Models loaded!")

SPEAKERS = {
    "Vivian": {"desc": "Bright young female voice", "lang": "Chinese", "gender": "Female"},
    "Serena": {"desc": "Warm gentle female voice", "lang": "Chinese", "gender": "Female"},
    "Uncle_Fu": {"desc": "Seasoned male, low mellow", "lang": "Chinese", "gender": "Male"},
    "Dylan": {"desc": "Youthful Beijing male", "lang": "Chinese", "gender": "Male"},
    "Eric": {"desc": "Lively Chengdu male", "lang": "Chinese", "gender": "Male"},
    "Ryan": {"desc": "Dynamic male, strong rhythm", "lang": "English", "gender": "Male"},
    "Aiden": {"desc": "Sunny American male", "lang": "English", "gender": "Male"},
    "Ono_Anna": {"desc": "Playful Japanese female", "lang": "Japanese", "gender": "Female"},
    "Sohee": {"desc": "Warm Korean female", "lang": "Korean", "gender": "Female"},
}

LANGUAGES = ["English", "Chinese", "Japanese", "Korean", "German", "French", "Russian", "Portuguese", "Spanish", "Italian"]

def generate_custom_voice(text, language, speaker, instruct, stream_enabled):
    if not text or not text.strip():
        return None, "❌ Enter text"
    
    try:
        kwargs = {"text": text, "language": language, "speaker": speaker}
        if instruct and instruct.strip():
            kwargs["instruct"] = instruct
        
        wavs, sr = custom_model.generate_custom_voice(**kwargs)
        
        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
            sf.write(f.name, wavs[0], sr)
            mode = "Streaming" if stream_enabled else "Full"
            return f.name, f"✅ {mode} generation complete"
    except Exception as e:
        return None, f"❌ {str(e)}"

def generate_voice_clone(text, language, ref_audio, ref_text, stream_enabled):
    if not text or not text.strip():
        return None, "❌ Enter text"
    if not ref_audio:
        return None, "❌ Upload reference audio"
    if not ref_text or not ref_text.strip():
        return None, "❌ Enter reference text"
    
    try:
        wavs, sr = base_model.generate_voice_clone(
            text=text, language=language,
            ref_audio=ref_audio, ref_text=ref_text,
        )
        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
            sf.write(f.name, wavs[0], sr)
            return f.name, "✅ Voice cloned"
    except Exception as e:
        return None, f"❌ {str(e)}"

# ============== CUSTOM HTML/CSS UI ==============

custom_css = """
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap');

* {
    margin: 0;
    padding: 0;
    box-sizing: border-box;
}

/* Dark Animated Background */
.gradio-container {
    background: #050508 !important;
    min-height: 100vh;
    font-family: 'Inter', sans-serif !important;
    position: relative;
    overflow: hidden;
}

.gradio-container::before {
    content: '';
    position: fixed;
    top: 0;
    left: 0;
    right: 0;
    bottom: 0;
    background: 
        radial-gradient(ellipse 80% 50% at 20% 40%, rgba(120, 0, 255, 0.15), transparent),
        radial-gradient(ellipse 60% 40% at 80% 60%, rgba(0, 200, 255, 0.1), transparent),
        radial-gradient(ellipse 50% 30% at 50% 80%, rgba(255, 0, 150, 0.08), transparent);
    pointer-events: none;
    z-index: 0;
}

/* Animated orbs */
.orb {
    position: fixed;
    border-radius: 50%;
    filter: blur(80px);
    opacity: 0.5;
    animation: float 20s ease-in-out infinite;
    pointer-events: none;
    z-index: 0;
}

.orb-1 {
    width: 400px;
    height: 400px;
    background: linear-gradient(135deg, #7c3aed, #3b82f6);
    top: -100px;
    left: -100px;
    animation-delay: 0s;
}

.orb-2 {
    width: 300px;
    height: 300px;
    background: linear-gradient(135deg, #06b6d4, #8b5cf6);
    bottom: -50px;
    right: -50px;
    animation-delay: -5s;
}

.orb-3 {
    width: 200px;
    height: 200px;
    background: linear-gradient(135deg, #ec4899, #8b5cf6);
    top: 50%;
    left: 50%;
    animation-delay: -10s;
}

@keyframes float {
    0%, 100% { transform: translate(0, 0) scale(1); }
    25% { transform: translate(50px, -30px) scale(1.1); }
    50% { transform: translate(-30px, 50px) scale(0.9); }
    75% { transform: translate(40px, 20px) scale(1.05); }
}

/* Main Container */
.main-container {
    position: relative;
    z-index: 1;
    max-width: 900px;
    margin: 0 auto;
    padding: 40px 20px;
}

/* Frosted Glass Card */
.glass-card {
    background: rgba(255, 255, 255, 0.03);
    backdrop-filter: blur(40px) saturate(150%);
    -webkit-backdrop-filter: blur(40px) saturate(150%);
    border-radius: 28px;
    border: 1px solid rgba(255, 255, 255, 0.08);
    box-shadow: 
        0 0 0 1px rgba(255, 255, 255, 0.05),
        0 20px 50px -10px rgba(0, 0, 0, 0.5),
        0 40px 80px -20px rgba(0, 0, 0, 0.3),
        inset 0 1px 0 rgba(255, 255, 255, 0.1);
    overflow: hidden;
    margin-bottom: 24px;
}

/* Header */
.app-header {
    text-align: center;
    padding: 40px 40px 30px;
    border-bottom: 1px solid rgba(255, 255, 255, 0.05);
}

.app-title {
    font-size: 52px;
    font-weight: 700;
    color: white;
    letter-spacing: -2px;
    margin-bottom: 8px;
    background: linear-gradient(135deg, #fff 0%, rgba(255,255,255,0.7) 100%);
    -webkit-background-clip: text;
    -webkit-text-fill-color: transparent;
    background-clip: text;
}

.app-subtitle {
    font-size: 16px;
    color: rgba(255, 255, 255, 0.5);
    font-weight: 400;
}

.app-badges {
    display: flex;
    gap: 12px;
    justify-content: center;
    margin-top: 20px;
    flex-wrap: wrap;
}

.badge {
    padding: 8px 16px;
    border-radius: 100px;
    font-size: 13px;
    font-weight: 500;
    border: 1px solid;
}

.badge-purple {
    background: rgba(139, 92, 246, 0.1);
    border-color: rgba(139, 92, 246, 0.3);
    color: #a78bfa;
}

.badge-cyan {
    background: rgba(34, 211, 238, 0.1);
    border-color: rgba(34, 211, 238, 0.3);
    color: #67e8f9;
}

.badge-pink {
    background: rgba(236, 72, 153, 0.1);
    border-color: rgba(236, 72, 153, 0.3);
    color: #f9a8d4;
}

/* Content Area */
.content-area {
    padding: 32px 40px;
}

/* Section Title */
.section-title {
    font-size: 13px;
    font-weight: 600;
    color: rgba(255, 255, 255, 0.4);
    text-transform: uppercase;
    letter-spacing: 1.5px;
    margin-bottom: 20px;
}

/* Tab Buttons */
.tab-container {
    display: flex;
    gap: 8px;
    margin-bottom: 32px;
    background: rgba(255, 255, 255, 0.02);
    padding: 6px;
    border-radius: 16px;
    border: 1px solid rgba(255, 255, 255, 0.05);
}

.tab-btn {
    flex: 1;
    padding: 14px 24px;
    border: none;
    background: transparent;
    color: rgba(255, 255, 255, 0.5);
    font-size: 15px;
    font-weight: 500;
    border-radius: 12px;
    cursor: pointer;
    transition: all 0.3s ease;
    font-family: inherit;
}

.tab-btn:hover {
    color: white;
    background: rgba(255, 255, 255, 0.05);
}

.tab-btn.active {
    background: rgba(255, 255, 255, 0.1);
    color: white;
    box-shadow: 0 0 0 1px rgba(255, 255, 255, 0.1);
}

/* Input Groups */
.input-group {
    margin-bottom: 24px;
}

.input-label {
    display: block;
    font-size: 14px;
    font-weight: 500;
    color: rgba(255, 255, 255, 0.8);
    margin-bottom: 10px;
}

.glass-input {
    width: 100%;
    padding: 16px 20px;
    background: rgba(255, 255, 255, 0.03);
    border: 1px solid rgba(255, 255, 255, 0.08);
    border-radius: 14px;
    color: white;
    font-size: 15px;
    font-family: inherit;
    transition: all 0.3s ease;
    resize: none;
}

.glass-input:focus {
    outline: none;
    border-color: rgba(139, 92, 246, 0.5);
    background: rgba(255, 255, 255, 0.05);
    box-shadow: 0 0 0 4px rgba(139, 92, 246, 0.1);
}

.glass-input::placeholder {
    color: rgba(255, 255, 255, 0.3);
}

/* Dropdown */
.glass-dropdown {
    width: 100%;
    padding: 16px 20px;
    background: rgba(10, 10, 15, 0.8);
    border: 1px solid rgba(255, 255, 255, 0.08);
    border-radius: 14px;
    color: white;
    font-size: 15px;
    font-family: inherit;
    cursor: pointer;
    appearance: none;
    background-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='24' height='24' viewBox='0 0 24 24' fill='none' stroke='rgba(255,255,255,0.5)' stroke-width='2'%3E%3Cpath d='M6 9l6 6 6-6'/%3E%3C/svg%3E");
    background-repeat: no-repeat;
    background-position: right 16px center;
}

/* Speaker Grid */
.speaker-grid {
    display: grid;
    grid-template-columns: repeat(3, 1fr);
    gap: 12px;
    margin-bottom: 24px;
}

.speaker-card {
    padding: 16px;
    background: rgba(255, 255, 255, 0.02);
    border: 1px solid rgba(255, 255, 255, 0.05);
    border-radius: 14px;
    cursor: pointer;
    transition: all 0.3s ease;
    text-align: center;
}

.speaker-card:hover {
    background: rgba(255, 255, 255, 0.05);
    border-color: rgba(139, 92, 246, 0.3);
    transform: translateY(-2px);
}

.speaker-card.selected {
    background: rgba(139, 92, 246, 0.15);
    border-color: rgba(139, 92, 246, 0.5);
}

.speaker-name {
    font-size: 15px;
    font-weight: 600;
    color: white;
    margin-bottom: 4px;
}

.speaker-meta {
    font-size: 12px;
    color: rgba(255, 255, 255, 0.4);
}

.speaker-gender {
    font-size: 11px;
    color: rgba(255, 255, 255, 0.3);
    margin-top: 4px;
}

/* Generate Button */
.generate-btn {
    width: 100%;
    padding: 18px 32px;
    background: linear-gradient(135deg, #8b5cf6 0%, #06b6d4 100%);
    border: none;
    border-radius: 14px;
    color: white;
    font-size: 16px;
    font-weight: 600;
    font-family: inherit;
    cursor: pointer;
    transition: all 0.3s ease;
    position: relative;
    overflow: hidden;
}

.generate-btn::before {
    content: '';
    position: absolute;
    top: 0;
    left: 0;
    right: 0;
    bottom: 0;
    background: linear-gradient(135deg, #a78bfa 0%, #22d3ee 100%);
    opacity: 0;
    transition: opacity 0.3s ease;
}

.generate-btn:hover::before {
    opacity: 1;
}

.generate-btn span {
    position: relative;
    z-index: 1;
}

.generate-btn:hover {
    transform: translateY(-2px);
    box-shadow: 0 10px 30px -10px rgba(139, 92, 246, 0.5);
}

/* Audio Output */
.audio-output {
    margin-top: 24px;
    padding: 20px;
    background: rgba(255, 255, 255, 0.02);
    border: 1px solid rgba(255, 255, 255, 0.05);
    border-radius: 16px;
}

.audio-output audio {
    width: 100%;
    border-radius: 12px;
}

/* Status */
.status-text {
    padding: 12px 16px;
    background: rgba(255, 255, 255, 0.02);
    border-radius: 10px;
    font-size: 14px;
    color: rgba(255, 255, 255, 0.7);
    margin-top: 16px;
    text-align: center;
}

/* Settings Panel */
.settings-row {
    display: flex;
    align-items: center;
    justify-content: space-between;
    padding: 16px 20px;
    background: rgba(255, 255, 255, 0.02);
    border: 1px solid rgba(255, 255, 255, 0.05);
    border-radius: 14px;
    margin-bottom: 24px;
}

.settings-label {
    font-size: 14px;
    color: rgba(255, 255, 255, 0.8);
}

.settings-toggle {
    position: relative;
    width: 52px;
    height: 28px;
    background: rgba(255, 255, 255, 0.1);
    border-radius: 14px;
    cursor: pointer;
    transition: all 0.3s ease;
    border: 1px solid rgba(255, 255, 255, 0.1);
}

.settings-toggle.active {
    background: linear-gradient(135deg, #8b5cf6, #06b6d4);
    border-color: transparent;
}

.settings-toggle::after {
    content: '';
    position: absolute;
    top: 3px;
    left: 3px;
    width: 20px;
    height: 20px;
    background: white;
    border-radius: 50%;
    transition: all 0.3s ease;
    box-shadow: 0 2px 4px rgba(0, 0, 0, 0.2);
}

.settings-toggle.active::after {
    left: 27px;
}

/* Footer */
.app-footer {
    text-align: center;
    padding: 24px;
    color: rgba(255, 255, 255, 0.3);
    font-size: 13px;
}

/* Hide Gradio default styles */
.gradio-container .contain {
    background: transparent !important;
    border: none !important;
    box-shadow: none !important;
}

.gradio-container .form {
    background: transparent !important;
    border: none !important;
}

.gradio-container input, .gradio-container textarea, .gradio-container select {
    background: rgba(255, 255, 255, 0.03) !important;
    border: 1px solid rgba(255, 255, 255, 0.08) !important;
    border-radius: 14px !important;
    color: white !important;
}

.gradio-container button.primary {
    background: linear-gradient(135deg, #8b5cf6 0%, #06b6d4 100%) !important;
    border: none !important;
    border-radius: 14px !important;
}

.gradio-container audio {
    border-radius: 12px !important;
    background: rgba(255, 255, 255, 0.02) !important;
}

.gradio-container .tabs {
    background: transparent !important;
    border: none !important;
}

.gradio-container .tabitem {
    background: transparent !important;
    border: none !important;
}

.gradio-container label {
    color: rgba(255, 255, 255, 0.8) !important;
}

/* Responsive */
@media (max-width: 768px) {
    .speaker-grid {
        grid-template-columns: repeat(2, 1fr);
    }
    
    .app-title {
        font-size: 36px;
    }
    
    .content-area {
        padding: 24px 20px;
    }
}
"""

# ============== GRADIO APP ==============

with gr.Blocks(css=custom_css, title="Qwen3-TTS", theme=gr.themes.Base()) as demo:
    
    stream_state = gr.State(False)
    selected_speaker = gr.State("Ryan")
    
    # Main Container
    with gr.Column(elem_classes="main-container"):
        
        # Animated orbs (HTML)
        gr.HTML("""
        <div class="orb orb-1"></div>
        <div class="orb orb-2"></div>
        <div class="orb orb-3"></div>
        """)
        
        # Main Glass Card
        with gr.Column(elem_classes="glass-card"):
            
            # Header
            gr.HTML("""
            <div class="app-header">
                <h1 class="app-title">Qwen3-TTS</h1>
                <p class="app-subtitle">Multilingual Text-to-Speech with Voice Cloning</p>
                <div class="app-badges">
                    <span class="badge badge-purple">🎙️ 10 Languages</span>
                    <span class="badge badge-cyan">🎨 Voice Clone</span>
                    <span class="badge badge-pink">⚡ 9 Speakers</span>
                </div>
            </div>
            """)
            
            # Content
            with gr.Column(elem_classes="content-area"):
                
                # Streaming Toggle
                with gr.Row(elem_classes="settings-row"):
                    gr.HTML('<span class="settings-label">🎧 Enable Streaming</span>')
                    stream_toggle = gr.HTML('<div class="settings-toggle" id="streamToggle" onclick="this.classList.toggle(\'active\')"></div>')
                
                # Tab Selection
                with gr.Tabs() as tabs:
                    
                    # Custom Voice Tab
                    with gr.TabItem("🗣️ Custom Voice"):
                        with gr.Row():
                            with gr.Column(scale=2):
                                text_input = gr.Textbox(
                                    label="Text to Speak",
                                    lines=4,
                                    placeholder="Enter your text here...",
                                    elem_classes="glass-input"
                                )
                                lang_dropdown = gr.Dropdown(
                                    LANGUAGES,
                                    value="English",
                                    label="Language",
                                    elem_classes="glass-dropdown"
                                )
                                instruct_input = gr.Textbox(
                                    label="Style Instruction (optional)",
                                    lines=2,
                                    placeholder='e.g., "Speak slowly with emotion"',
                                    elem_classes="glass-input"
                                )
                            
                            with gr.Column(scale=1):
                                speaker_dropdown = gr.Dropdown(
                                    list(SPEAKERS.keys()),
                                    value="Ryan",
                                    label="Speaker",
                                    elem_classes="glass-dropdown"
                                )
                                speaker_info_display = gr.HTML("""
                                <div style="padding: 16px; background: rgba(255,255,255,0.02); border-radius: 14px; border: 1px solid rgba(255,255,255,0.05);">
                                    <div style="font-size: 14px; color: white; font-weight: 600; margin-bottom: 8px;">Ryan</div>
                                    <div style="font-size: 12px; color: rgba(255,255,255,0.5);">Male • English</div>
                                    <div style="font-size: 12px; color: rgba(255,255,255,0.4); margin-top: 8px;">Dynamic male voice with strong rhythm</div>
                                </div>
                                """)
                        
                        generate_btn = gr.Button("🎤 Generate Speech", variant="primary", elem_classes="generate-btn")
                        
                        with gr.Row():
                            audio_output = gr.Audio(label="Output", elem_classes="audio-output")
                        
                        status_output = gr.Textbox(visible=False)
                    
                    # Voice Clone Tab  
                    with gr.TabItem("🎨 Voice Clone"):
                        with gr.Row():
                            with gr.Column():
                                clone_text = gr.Textbox(
                                    label="Text to Speak",
                                    lines=4,
                                    placeholder="Enter text...",
                                    elem_classes="glass-input"
                                )
                                clone_lang = gr.Dropdown(
                                    LANGUAGES,
                                    value="English",
                                    label="Language",
                                    elem_classes="glass-dropdown"
                                )
                                ref_audio = gr.Audio(
                                    label="Reference Audio (3-10 sec)",
                                    type="filepath",
                                    elem_classes="audio-output"
                                )
                                ref_text = gr.Textbox(
                                    label="Reference Text",
                                    lines=2,
                                    placeholder="What's spoken in the audio...",
                                    elem_classes="glass-input"
                                )
                                clone_btn = gr.Button("🎨 Clone & Generate", variant="primary", elem_classes="generate-btn")
                            
                            with gr.Column():
                                clone_output = gr.Audio(label="Output", elem_classes="audio-output")
                                clone_status = gr.Textbox(visible=False)
                
                # Footer
                gr.HTML("""
                <div class="app-footer">
                    Running on CPU • Generation takes ~15-30 seconds
                </div>
                """)
    
    # Event handlers
    def update_speaker_info(speaker):
        info = SPEAKERS[speaker]
        return f"""
        <div style="padding: 16px; background: rgba(255,255,255,0.02); border-radius: 14px; border: 1px solid rgba(255,255,255,0.05);">
            <div style="font-size: 14px; color: white; font-weight: 600; margin-bottom: 8px;">{speaker}</div>
            <div style="font-size: 12px; color: rgba(255,255,255,0.5);">{info['gender']} • {info['lang']}</div>
            <div style="font-size: 12px; color: rgba(255,255,255,0.4); margin-top: 8px;">{info['desc']}</div>
        </div>
        """
    
    speaker_dropdown.change(update_speaker_info, [speaker_dropdown], [speaker_info_display])
    
    generate_btn.click(
        generate_custom_voice,
        [text_input, lang_dropdown, speaker_dropdown, instruct_input, stream_state],
        [audio_output, status_output]
    )
    
    clone_btn.click(
        generate_voice_clone,
        [clone_text, clone_lang, ref_audio, ref_text, stream_state],
        [clone_output, clone_status]
    )

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)