import gradio as gr
from TTS.api import TTS
import os

# 1. Environment Variable set karna (License Agreement ke liye)
os.environ["COQUI_TOS_AGREED"] = "1"

# 2. Model Load karna (CPU Mode)
print("⏳ Model load ho raha hai, kripya wait karein...")
try:
    # Model download aur load
    tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=False)
    print("✅ Model load ho gaya!")
except Exception as e:
    print(f"❌ Model Error: {e}")

# 3. Voices folder check karna
def get_voices():
    voice_folder = "voices"
    # Agar folder nahi hai to bana do
    if not os.path.exists(voice_folder):
        os.makedirs(voice_folder)
        return []
    
    # Sirf .wav files dhoondo
    voices = [f for f in os.listdir(voice_folder) if f.endswith(".wav")]
    return voices

def generate_voice(text, voice_name, language):
    try:
        if not voice_name:
            return None, "⚠️ Kripya pehle koi Voice select karein!"

        speaker_audio_path = os.path.join("voices", voice_name)
        
        # Check agar file exist karti hai
        if not os.path.exists(speaker_audio_path):
            return None, f"⚠️ Error: '{voice_name}' file nahi mili!"

        output_path = "output.wav"
        
        # Audio generate karna
        tts.tts_to_file(
            text=text,
            file_path=output_path,
            speaker_wav=speaker_audio_path,
            language=language
        )
        return output_path, "✅ Audio Generated Successfully!"
    except Exception as e:
        return None, f"❌ Error: {str(e)}"

# 4. UI Design
with gr.Blocks(title="Deepak's AI Voice Tool") as demo:
    gr.Markdown("## 🎙️ AI Voice Cloning (11Labs Style)")
    
    with gr.Row():
        with gr.Column():
            text_input = gr.Textbox(
                label="Text likhein (Hindi/English)", 
                placeholder="Namaste, main Pappu bol raha hoon. Aaj hum AI seekhenge.",
                lines=3
            )
            
            # Character Selection Dropdown
            # Note: Agar list khali hai to error na de isliye check lagaya hai
            initial_voices = get_voices()
            voice_dropdown = gr.Dropdown(
                choices=initial_voices, 
                label="Character Select Karein (voices folder se)",
                value=initial_voices[0] if initial_voices else None,
                interactive=True
            )
            
            refresh_btn = gr.Button("🔄 Refresh Voice List")
            
            language_input = gr.Dropdown(
                choices=["en", "hi", "es", "fr", "de", "it", "ja"], 
                value="hi", 
                label="Language"
            )
            
            generate_btn = gr.Button("🔊 Generate Audio", variant="primary")
        
        with gr.Column():
            audio_output = gr.Audio(label="Output Audio")
            status_msg = gr.Textbox(label="Status", interactive=False)

    # Button Functions
    def refresh_dropdown():
        new_choices = get_voices()
        return gr.Dropdown(choices=new_choices, value=new_choices[0] if new_choices else None)

    refresh_btn.click(fn=refresh_dropdown, inputs=[], outputs=voice_dropdown)
    
    generate_btn.click(
        fn=generate_voice,
        inputs=[text_input, voice_dropdown, language_input],
        outputs=[audio_output, status_msg]
    )

# 5. App Launch (Docker Compatibility ke liye ye zaroori hai)
if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)