novafarma/scripts/generate_voices_edge_tts.py

#!/usr/bin/env python3
"""
Edge TTS Voice Generator
Generate voice-over audio using Microsoft Edge TTS
"""

import asyncio
import edge_tts
from pathlib import Path

# Output directory
OUTPUT_DIR = Path("/Users/davidkotnik/repos/novafarma/assets/audio/voices")

# Voice configurations
VOICES = {
    "kai_en": "en-US-AvaNeural",  # English - female, young
    "kai_sl": "sl-SI-PetraNeural",  # Slovenian
    "ana_en": "en-US-JennyNeural",  # English - warm, friendly
    "ana_sl": "sl-SI-PetraNeural",
    "narrator_en": "en-US-GuyNeural",  # English - deep, storytelling
    "narrator_sl": "sl-SI-RokNeural",  # Slovenian male
}

async def generate_voice(text: str, voice: str, output_path: Path, rate: str = "+0%", pitch: str = "+0Hz"):
    """
    Generate voice audio from text
    
    Args:
        text: Text to convert to speech
        voice: Voice ID (e.g. "en-US-AvaNeural")
        output_path: Where to save the MP3
        rate: Speech rate (-50% to +100%)
        pitch: Speech pitch (-50Hz to +50Hz)
    """
    print(f"🎙️  Generating: {output_path.name}")
    print(f"   Voice: {voice}")
    print(f"   Text: {text[:50]}...")
    
    # Create output directory
    output_path.parent.mkdir(parents=True, exist_ok=True)
    
    # Generate speech
    communicate = edge_tts.Communicate(text, voice, rate=rate, pitch=pitch)
    await communicate.save(str(output_path))
    
    # Check file size
    size = output_path.stat().st_size
    print(f"   ✅ Saved: {size:,} bytes\n")

async def generate_test():
    """Generate test voice for Kai"""
    text = "My name is Kai, and I will find my sister."
    output = OUTPUT_DIR / "kai" / "kai_test_01.mp3"
    
    await generate_voice(
        text=text,
        voice=VOICES["kai_en"],
        output_path=output
    )

async def generate_kai_voices():
    """Generate Kai's voice lines"""
    lines = [
        "My name is Kai, and I will find my sister.",
        "Ana, where are you? I won't give up.",
        "This farm... it reminds me of home.",
        "I need to keep farming. For Ana.",
        "Another day, another harvest. But I won't forget."
    ]
    
    for i, text in enumerate(lines, 1):
        output = OUTPUT_DIR / "kai" / f"kai_{i:02d}.mp3"
        await generate_voice(text, VOICES["kai_en"], output)

async def generate_ana_voices():
    """Generate Ana's voice lines (memories)"""
    lines = [
        "Kai... can you hear me?",
        "Remember the farm... remember our home.",
        "I'm still here, Kai. Don't forget me.",
        "The valley holds secrets... find them."
    ]
    
    for i, text in enumerate(lines, 1):
        output = OUTPUT_DIR / "ana" / f"ana_{i:02d}.mp3"
        await generate_voice(text, VOICES["ana_en"], output, rate="-10%", pitch="-5Hz")

async def generate_narrator_voices():
    """Generate narrator voice lines"""
    lines = [
        "In the Valley of Death, a young farmer searches for answers.",
        "Long ago, this valley was green and full of life.",
        "But the dead walk now, and the living must survive."
    ]
    
    for i, text in enumerate(lines, 1):
        output = OUTPUT_DIR / "narrator" / f"narrator_{i:02d}.mp3"
        await generate_voice(text, VOICES["narrator_en"], output, rate="-5%")

async def list_available_voices():
    """List all available Edge TTS voices"""
    print("\n📋 Available Edge TTS Voices:\n")
    
    voices = await edge_tts.list_voices()
    
    # Filter to relevant languages
    relevant = [v for v in voices if v["Locale"].startswith(("en-", "sl-"))]
    
    for voice in relevant[:20]:  # Show first 20
        print(f"  {voice['ShortName']}")
        print(f"    Language: {voice['Locale']}")
        print(f"    Gender: {voice['Gender']}")
        print()

async def main():
    """Main execution"""
    print("="*60)
    print("🎙️  EDGE TTS VOICE GENERATOR")
    print("="*60)
    
    # Generate test first
    print("\n🧪 GENERATING TEST VOICE:\n")
    await generate_test()
    
    print("\n" + "="*60)
    print("✅ TEST COMPLETE! Check: assets/audio/voices/kai/kai_test_01.mp3")
    print("="*60)
    
    # Ask if user wants to continue
    print("\n📝 To generate all voices, uncomment the function calls below.")

if __name__ == "__main__":
    # Run async main
    asyncio.run(main())
    
    # UNCOMMENT BELOW TO GENERATE ALL VOICES:
    # asyncio.run(generate_kai_voices())
    # asyncio.run(generate_ana_voices())
    # asyncio.run(generate_narrator_voices())