Add app.py

app.py

@@ -0,0 +1,191 @@
+#!/usr/bin/env python3
+"""
+HOLLY TTS API - Maya1 FastAPI Service
+Production-ready TTS microservice for HOLLY AI
+"""
+
+from fastapi import FastAPI, HTTPException
+from fastapi.responses import Response
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel, Field
+from typing import Optional
+import os
+import io
+import soundfile as sf
+
+from holly_voice_generator import HollyVoiceGenerator, HOLLY_VOICE_DESCRIPTION
+
+# Initialize FastAPI
+app = FastAPI(
+    title="HOLLY TTS API",
+    description="Self-hosted Maya1 TTS microservice for HOLLY AI",
+    version="1.0.0"
+)
+
+# CORS middleware
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],  # In production, restrict to your domains
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Global voice generator (lazy load)
+voice_generator: Optional[HollyVoiceGenerator] = None
+
+
+def get_generator() -> HollyVoiceGenerator:
+    """Get or initialize the voice generator"""
+    global voice_generator
+    if voice_generator is None:
+        voice_generator = HollyVoiceGenerator()
+    return voice_generator
+
+
+class TTSRequest(BaseModel):
+    """TTS generation request"""
+    text: str = Field(..., description="Text to synthesize", min_length=1, max_length=5000)
+    description: Optional[str] = Field(
+        None,
+        description="Voice description (defaults to HOLLY's signature voice)"
+    )
+    temperature: float = Field(0.4, ge=0.1, le=1.0, description="Sampling temperature")
+    top_p: float = Field(0.9, ge=0.1, le=1.0, description="Nucleus sampling threshold")
+
+
+class TTSResponse(BaseModel):
+    """TTS generation response metadata"""
+    success: bool
+    duration_seconds: float
+    sample_rate: int = 24000
+    message: str
+
+
+@app.on_event("startup")
+async def startup_event():
+    """Preload model on startup"""
+    print("🚀 HOLLY TTS API starting up...")
+    # Optionally preload model here
+    # get_generator()
+    print("✅ HOLLY TTS API ready!")
+
+
+@app.get("/")
+async def root():
+    """Health check endpoint"""
+    return {
+        "service": "HOLLY TTS API",
+        "status": "online",
+        "model": "maya-research/maya1",
+        "version": "1.0.0",
+        "voice": "HOLLY (Female, 30s, American, confident, intelligent, warm)"
+    }
+
+
+@app.get("/health")
+async def health():
+    """Health check for monitoring"""
+    return {"status": "healthy", "model_loaded": voice_generator is not None}
+
+
+@app.post("/generate", response_class=Response)
+async def generate_speech(request: TTSRequest):
+    """
+    Generate speech from text using HOLLY's voice
+    
+    Returns WAV audio (24kHz, mono)
+    """
+    try:
+        # Get generator
+        generator = get_generator()
+        
+        # Generate audio
+        audio = generator.generate(
+            text=request.text,
+            description=request.description,
+            temperature=request.temperature,
+            top_p=request.top_p
+        )
+        
+        # Convert to WAV bytes
+        wav_buffer = io.BytesIO()
+        sf.write(wav_buffer, audio, 24000, format='WAV')
+        wav_bytes = wav_buffer.getvalue()
+        
+        # Return audio
+        return Response(
+            content=wav_bytes,
+            media_type="audio/wav",
+            headers={
+                "Content-Disposition": "inline; filename=holly_speech.wav",
+                "X-Duration-Seconds": str(len(audio) / 24000),
+                "X-Sample-Rate": "24000"
+            }
+        )
+    
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"TTS generation failed: {str(e)}")
+
+
+@app.post("/generate/info")
+async def generate_speech_info(request: TTSRequest) -> TTSResponse:
+    """
+    Generate speech and return metadata (without audio bytes)
+    Useful for testing and monitoring
+    """
+    try:
+        generator = get_generator()
+        
+        audio = generator.generate(
+            text=request.text,
+            description=request.description,
+            temperature=request.temperature,
+            top_p=request.top_p
+        )
+        
+        duration = len(audio) / 24000
+        
+        return TTSResponse(
+            success=True,
+            duration_seconds=duration,
+            message=f"Generated {len(audio)} samples ({duration:.2f}s)"
+        )
+    
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"TTS generation failed: {str(e)}")
+
+
+@app.get("/voice/info")
+async def voice_info():
+    """Get HOLLY's voice profile information"""
+    return {
+        "voice_name": "HOLLY",
+        "description": HOLLY_VOICE_DESCRIPTION,
+        "model": "maya-research/maya1",
+        "sample_rate": 24000,
+        "supported_emotions": [
+            "laugh", "laugh_harder", "chuckle", "giggle",
+            "whisper", "sigh", "gasp",
+            "angry", "cry",
+            "confident", "warm", "intelligent"
+        ],
+        "usage_example": {
+            "text": "Hello Hollywood! <chuckle> Let's build something amazing.",
+            "description": HOLLY_VOICE_DESCRIPTION
+        }
+    }
+
+
+if __name__ == "__main__":
+    import uvicorn
+    
+    port = int(os.environ.get("PORT", 8000))
+    
+    uvicorn.run(
+        "app:app",
+        host="0.0.0.0",
+        port=port,
+        workers=1,  # Maya1 is memory-intensive, use 1 worker
+        log_level="info"
+    )