Server changes for deployment

Dockerfile

@@ -33,6 +33,6 @@ COPY . .
 ENV PYTHONPATH=/app
 ENV PYTHONUNBUFFERED=1
 
-EXPOSE 8000
+EXPOSE 7860
 
-CMD ["uvicorn", "src.app.server:app", "--host", "0.0.0.0", "--port", "8000"]
+CMD ["uvicorn", "server:app", "--host", "0.0.0.0", "--port", "7860"]

src/app/index.html → index.html

@@ -293,7 +293,7 @@
         let selectedFile = null;
 
         // API endpoint - update this to your server URL
-        const API_URL = 'http://192.168.1.12:8082';
+        const API_URL = 'https://mateo496-esc50-model.hf.space/';
 
         // Click to upload
         uploadArea.addEventListener('click', () => fileInput.click());

src/app/server.py → server.py

@@ -1,16 +1,18 @@
-import uvicorn
-import torch
+import os
 import tempfile
-import os 
 
-from pydub import AudioSegment
-from fastapi import FastAPI, File, UploadFile, HTTPException
+import torch
+import uvicorn
+from fastapi import FastAPI, File, HTTPException, UploadFile
 from fastapi.responses import FileResponse
 from fastapi.staticfiles import StaticFiles
+from fastapi.middleware.cors import CORSMiddleware
+from pydub import AudioSegment
 
-from src.models.predict import load_model, predict_file
-from src.config.config import esc50_labels
+from src.config.config import DatasetConfig
+from src.models.predict import AudioPredictor
 
+dataset_cfg = DatasetConfig()
 
 app = FastAPI(
     title="ESC50 Audio Classifier API",
@@ -18,17 +20,20 @@ app = FastAPI(
     version="1.0.0",
 )
 
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["GET", "POST"],
+    allow_headers=["*"],
+)
 
-
-model = None
-device = None
-model_path="models/cnn/saved/final_model.pt"
 device = "cuda" if torch.cuda.is_available() else "cpu"
-model = load_model(model_path, device)
+predictor = AudioPredictor("models/cnn/saved/final_model.pt", device=device)
+
 
 @app.get("/")
 async def root():
-    return FileResponse("src/app/index.html")
+    return FileResponse("index.html")
 
 @app.get("/api/status")
 async def status():
@@ -38,40 +43,40 @@ async def status():
 
 @app.post("/predict-top-k")
 async def predict_top_k(file: UploadFile = File(...), k: int = 5):
-    if model is None:
+    if predictor is None:
         raise HTTPException(status_code=503, detail="Model not loaded")
-    
+
+    suffix = os.path.splitext(file.filename)[1]
+
+    with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as tmp:
+        tmp.write(await file.read())
+        tmp_path = tmp.name
+
     try:
-        with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(file.filename)[1]) as tmp:
-            content = await file.read()
-            tmp.write(content)
-            tmp_path = tmp.name
+        wav_path = tempfile.mktemp(suffix=".wav")
 
-        tmp_wav_path = tempfile.mktemp(suffix=".wav")
-        audio = AudioSegment.from_file(tmp_path)
-        audio.export(tmp_wav_path, format="wav")
+        print("[1] Converting to wav...")
+        AudioSegment.from_file(tmp_path).export(wav_path, format="wav")
+        print("[2] Running inference...")
+        predicted_class, top_probs, top_indices = predictor.predict_file(wav_path, top_k=k)
+        print(f"[3] Done: {predicted_class} = {dataset_cfg.esc50_labels[predicted_class]}")
+        AudioSegment.from_file(tmp_path).export(wav_path, format="wav")
 
-        predicted_class, top_probs, top_indices = predict_file(
-            model, tmp_path, device=device, top_k=k
-        )        
+        predicted_class, top_probs, top_indices = predictor.predict_file(wav_path, top_k=k)
 
-        os.unlink(tmp_path)
-        
         return {
-            'top_predictions': [
-                {'class': esc50_labels[idx], 'confidence': float(prob)}
+            "predicted_class": dataset_cfg.esc50_labels[predicted_class],
+            "confidence": float(top_probs[0]),
+            "top_predictions": [
+                {"class": dataset_cfg.esc50_labels[idx], "confidence": float(prob)}
                 for prob, idx in zip(top_probs, top_indices)
             ],
-            'predicted_class': esc50_labels[predicted_class],
-            'confidence': float(top_probs[0])
         }
-        
-    except Exception as e:
-        raise HTTPException(status_code=400, detail=str(e))
-
-# uvicorn.run(
-#     app,
-#     host="0.0.0.0",
-#     port=8000,
-#     log_level="info"
-# )
+    finally:
+        os.unlink(tmp_path)
+        if os.path.exists(wav_path):
+            os.unlink(wav_path)
+
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=7860, log_level="info")