Piper TTS: runtime model download, no build-time network

Dockerfile

@@ -6,26 +6,12 @@ WORKDIR /app
 RUN apt-get update && apt-get install -y \
     libsndfile1 \
     espeak-ng \
-    wget \
     && rm -rf /var/lib/apt/lists/*
 
 # Install Python dependencies
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Pre-download voice models at build time (directly to /app)
-RUN cd /app && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx.json && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx.json && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ms/ms_MY/sirius/medium/ms_MY-sirius-medium.onnx && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ms/ms_MY/sirius/medium/ms_MY-sirius-medium.onnx.json && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ja/ja_JP/kenji/medium/ja_JP-kenji-medium.onnx && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ja/ja_JP/kenji/medium/ja_JP-kenji-medium.onnx.json && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ko/ko_KR/hui_hao/medium/ko_KR-hui-hao-medium.onnx && \
-    wget -q https://huggingface.co/rhasspy/piper-voices/resolve/main/ko/ko_KR/hui_hao/medium/ko_KR-hui-hao-medium.onnx.json
-
 # Copy application code
 COPY main.py .
 
@@ -33,7 +19,7 @@ COPY main.py .
 EXPOSE 7860
 
 # Health check
-HEALTHCHECK --interval=30s --timeout=3s --start-period=5s --retries=3 \
+HEALTHCHECK --interval=30s --timeout=3s --start-period=10s --retries=3 \
   CMD curl -f http://localhost:7860/health || exit 1
 
 # Run the application

main.py

@@ -6,30 +6,75 @@ import os
 import uuid
 from fastapi.responses import Response
 from io import BytesIO
+import requests
+from pathlib import Path
 
 app = FastAPI(title="Piper TTS Space")
 
-# Load voices on startup (download if needed)
+# Voice model info: (lang, model_name, url_onnx, url_json)
 VOICE_MODELS = {
-    "zh": "zh_CN-huayan-medium",      # Chinese
-    "en": "en_US-lessac-medium",      # English
-    "ms": "ms_MY-sirius-medium",      # Malay
-    "ja": "ja_JP-kenji-medium",       # Japanese
-    "ko": "ko_KR-hui-hao-medium",     # Korean
+    "zh": ("zh_CN-huayan-medium", 
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx.json"),
+    "en": ("en_US-lessac-medium",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx.json"),
+    "ms": ("ms_MY-sirius-medium",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ms/ms_MY/sirius/medium/ms_MY-sirius-medium.onnx",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ms/ms_MY/sirius/medium/ms_MY-sirius-medium.onnx.json"),
+    "ja": ("ja_JP-kenji-medium",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ja/ja_JP/kenji/medium/ja_JP-kenji-medium.onnx",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ja/ja_JP/kenji/medium/ja_JP-kenji-medium.onnx.json"),
+    "ko": ("ko_KR-hui-hao-medium",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ko/ko_KR/hui_hao/medium/ko_KR-hui-hao-medium.onnx",
+           "https://huggingface.co/rhasspy/piper-voices/resolve/main/ko/ko_KR/hui_hao/medium/ko_KR-hui-hao-medium.onnx.json"),
 }
 
+MODEL_DIR = Path("/tmp/piper_models")
+MODEL_DIR.mkdir(exist_ok=True)
 VOICES = {}
 
-print("Loading Piper voices...")
-for lang, model_name in VOICE_MODELS.items():
+def download_file(url: str, dest: Path):
+    """Download file with progress, follow redirects."""
+    with requests.get(url, stream=True, timeout=60, allow_redirects=True) as r:
+        r.raise_for_status()
+        with open(dest, 'wb') as f:
+            for chunk in r.iter_content(chunk_size=8192):
+                f.write(chunk)
+
+def get_voice(lang: str) -> PiperVoice:
+    """Load voice, downloading model if needed."""
+    if lang in VOICES:
+        return VOICES[lang]
+    
+    if lang not in VOICE_MODELS:
+        raise ValueError(f"Language '{lang}' not available. Available: {list(VOICE_MODELS.keys())}")
+    
+    model_name, url_onnx, url_json = VOICE_MODELS[lang]
+    onnx_path = MODEL_DIR / f"{model_name}.onnx"
+    json_path = MODEL_DIR / f"{model_name}.onnx.json"
+    
+    # Download if not cached
+    if not onnx_path.exists():
+        print(f"Downloading {model_name}...")
+        download_file(url_onnx, onnx_path)
+    if not json_path.exists():
+        print(f"Downloading {model_name}.json...")
+        download_file(url_json, json_path)
+    
+    voice = PiperVoice.load(str(onnx_path))
+    VOICES[lang] = voice
+    print(f"✓ Loaded {lang}: {model_name}")
+    return voice
+
+# Pre-load default voice (zh) on startup to warm up
+@app.on_event("startup")
+async def warmup():
     try:
-        voice = PiperVoice.load(f"{model_name}.onnx")
-        VOICES[lang] = voice
-        print(f"✓ Loaded {lang}: {model_name}")
+        get_voice("zh")
+        print("Warmup complete")
     except Exception as e:
-        print(f"✗ Failed to load {lang}: {e}")
-
-print(f"Loaded {len(VOICES)} voices: {list(VOICES.keys())}")
+        print(f"Warmup failed (will retry on first request): {e}")
 
 class TTSRequest(BaseModel):
     text: str
@@ -39,9 +84,10 @@ class TTSRequest(BaseModel):
 
 @app.post("/tts")
 async def tts(request: TTSRequest):
-    voice = VOICES.get(request.lang)
-    if not voice:
-        raise HTTPException(400, f"Language '{request.lang}' not available. Available: {list(VOICES.keys())}")
+    try:
+        voice = get_voice(request.lang)
+    except Exception as e:
+        raise HTTPException(400, f"Voice load failed: {str(e)}")
     
     try:
         audio = voice.synthesize(request.text, speaker_id=request.speaker_id)
@@ -58,13 +104,13 @@ async def tts(request: TTSRequest):
 
 @app.get("/health")
 async def health():
-    return {"status": "healthy", "voices": list(VOICES.keys())}
+    return {"status": "healthy", "loaded_voices": list(VOICES.keys()), "available_voices": list(VOICE_MODELS.keys())}
 
 @app.get("/")
 async def root():
     return {
         "message": "Piper TTS API - Fast multilingual TTS",
-        "voices": list(VOICES.keys()),
+        "available_voices": list(VOICE_MODELS.keys()),
         "endpoint": "/tts (POST)",
         "example": {"text": "你好，这是 Piper TTS 测试", "lang": "zh"}
     }

requirements.txt

@@ -3,3 +3,4 @@ uvicorn[standard]==0.30.1
 piper-tts==1.2.0
 soundfile==0.12.1
 pydantic==2.8.2
+requests==2.32.3