feat: version update, rate limit, etc

.gitignore

@@ -0,0 +1 @@
+*.pyc

Dockerfile

@@ -1,19 +1,14 @@
-FROM ghcr.io/remsky/kokoro-fastapi-cpu:latest
+FROM ghcr.io/remsky/kokoro-fastapi-cpu:v0.7.2
 
-# Environment setup
-ENV HOST=0.0.0.0
-ENV PORT=7860
+# 2 vCPU Space: keep torch from oversubscribing threads
+ENV OMP_NUM_THREADS=2
+ENV MKL_NUM_THREADS=2
 
-# ONNX optimization for 2 CPU environment
-ENV ONNX_MAX_INSTANCES=2
-ENV ONNX_NUM_THREADS=2
-ENV ONNX_INTER_OP_THREADS=1
-ENV VOICE_CACHE_SIZE=1
+# Default is DEBUG, which logs every caller's submitted text
+ENV API_LOG_LEVEL=INFO
 
-# Copy startup script
-COPY hf_start.py /app/hf_start.py
+# Root redirect to the bundled web player
+COPY hf_app.py /app/hf_app.py
 
-EXPOSE 7860
-
-# Run FastAPI with root UI
-CMD ["python", "/app/hf_start.py"]
+EXPOSE 8880
+CMD ["python", "-m", "uvicorn", "hf_app:app", "--host", "0.0.0.0", "--port", "8880"]

README.md

@@ -4,6 +4,7 @@ emoji: 📼
 colorFrom: gray
 colorTo: pink
 sdk: docker
+app_port: 8880
 pinned: true
 short_description: Web Utility that's living on Kokoro-FastAPI
 models:

hf_app.py

@@ -0,0 +1,47 @@
+# HF Space shim: serve the web player at / instead of a bare 404.
+import os
+import time
+from collections import deque
+
+from fastapi import Request
+from fastapi.responses import JSONResponse, RedirectResponse
+
+from api.src.main import app
+
+MAX_BODY_BYTES = int(os.getenv("MAX_BODY_BYTES", 8192))
+RATE_LIMIT_PER_MIN = int(os.getenv("RATE_LIMIT_PER_MIN", 20))
+
+_hits: dict[str, deque] = {}
+
+
+def _client_ip(request: Request) -> str:
+    forwarded = request.headers.get("x-forwarded-for", "").split(",")[-1].strip()
+    return forwarded or (request.client.host if request.client else "unknown")
+
+
+@app.middleware("http")
+async def throttle(request: Request, call_next):
+    if request.method != "POST":
+        return await call_next(request)
+
+    length = request.headers.get("content-length")
+    if length and int(length) > MAX_BODY_BYTES:
+        return JSONResponse({"detail": "Request body too large"}, status_code=413)
+
+    now = time.monotonic()
+    for stale in [k for k, v in _hits.items() if not v or v[-1] < now - 60]:
+        del _hits[stale]
+    hits = _hits.setdefault(_client_ip(request), deque())
+    while hits and hits[0] < now - 60:
+        hits.popleft()
+    if len(hits) >= RATE_LIMIT_PER_MIN:
+        return JSONResponse(
+            {"detail": "Rate limit exceeded"}, status_code=429, headers={"Retry-After": "60"}
+        )
+    hits.append(now)
+    return await call_next(request)
+
+
+@app.get("/", include_in_schema=False)
+async def root():
+    return RedirectResponse(url="/web/")

hf_start.py

@@ -1,12 +0,0 @@
-from api.src.main import app
-from fastapi.staticfiles import StaticFiles
-from fastapi.responses import RedirectResponse
-
-# Add root redirect to web UI
-@app.get("/")
-async def root():
-    return RedirectResponse(url="/web/")
-
-if __name__ == "__main__":
-    import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=7860)