my_voice_assistant_v2/app/main.py
Dieter Schlüter 111ecd8edf perf(startup): lokale Modelle beim Start vorladen (Warm-up)
Lifespan-Hintergrundtask laedt profilabhaengig piper-Stimme + faster-whisper-Modell
vor (nur was die aktive Konfiguration nutzt; Cloud-Profile = No-op). Server ist
sofort verfuegbar; der erste Nutzer zahlt nicht mehr den Kaltstart.

Messung: erster Turn direkt nach Start first_audio 1,62 s (statt ~4,5 s mit Kaltstart).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-18 08:27:37 +02:00

76 lines
2.9 KiB
Python

import asyncio
import time
from contextlib import asynccontextmanager
from pathlib import Path
from fastapi import FastAPI, Request
from fastapi.staticfiles import StaticFiles
from app.core.warmup import warmup_local_models
from app.metrics import metrics
from app.api.health import router as health_router
from app.api.chat import router as chat_router
from app.api.transcribe import router as transcribe_router
from app.api.speak import router as speak_router
from app.api.devices import router as devices_router
from app.api.sessions import router as sessions_router
from app.api.config import router as config_router
from app.api.admin import router as admin_router
from app.api.me import router as me_router
from app.api.metrics import router as metrics_router
from app.api.ws import router as ws_router
@asynccontextmanager
async def _lifespan(app: FastAPI):
# Lokale Modelle im Hintergrund vorladen -> Server ist sofort verfuegbar,
# der erste Nutzer zahlt nicht den Kaltstart.
task = asyncio.create_task(warmup_local_models())
yield
if not task.done():
task.cancel()
app = FastAPI(title="Voice Assistant Gateway", lifespan=_lifespan)
@app.middleware("http")
async def record_metrics(request: Request, call_next):
start = time.perf_counter()
response = await call_next(request)
duration = time.perf_counter() - start
# Route-Template (z. B. /api/sessions/{session_id}/route) statt konkreter URL,
# um die Label-Kardinalitaet niedrig zu halten.
route = request.scope.get("route")
path = getattr(route, "path", request.url.path)
labels = {"method": request.method, "path": path}
metrics.inc("http_requests_total", {**labels, "status": response.status_code})
metrics.observe("http_request_duration_seconds", duration, labels)
return response
app.include_router(health_router)
app.include_router(chat_router, prefix="/api")
app.include_router(transcribe_router, prefix="/api")
app.include_router(speak_router, prefix="/api")
app.include_router(devices_router, prefix="/api")
app.include_router(sessions_router, prefix="/api")
app.include_router(config_router, prefix="/api")
app.include_router(admin_router, prefix="/api")
app.include_router(me_router, prefix="/api")
app.include_router(metrics_router, prefix="/api")
app.include_router(ws_router)
# Statische Web-UI (same-origin -> kein CORS). Muss NACH allen API-/WS-Routen
# gemountet werden, damit "/" nur die uebrigen Pfade abfaengt.
class _NoCacheStaticFiles(StaticFiles):
"""Liefert die UI ohne Caching aus -> Aenderungen sind sofort sichtbar (kein Safari-Cache)."""
def file_response(self, *args, **kwargs):
response = super().file_response(*args, **kwargs)
response.headers["Cache-Control"] = "no-cache, no-store, must-revalidate"
return response
_WEB_DIR = Path(__file__).resolve().parent / "web"
if _WEB_DIR.is_dir():
app.mount("/", _NoCacheStaticFiles(directory=str(_WEB_DIR), html=True), name="web")