Create app.py

app.py

@@ -0,0 +1,80 @@
+# app.py
+from fastapi import FastAPI, UploadFile, File
+import numpy as np
+from keras.models import load_model
+from graph import zeropad, zeropad_output_shape
+from config import get_config
+
+app = FastAPI(
+    title="ECG Classification Backend",
+    description="REST API for ECG heartbeat classification",
+    version="1.1.0"
+)
+
+# ======= Load the model at startup =======
+config = get_config()
+MODEL_PATH = "MLII-latest.keras"
+
+print("🔹 Loading model:", MODEL_PATH)
+model = load_model(
+    MODEL_PATH,
+    custom_objects={
+        "zeropad": zeropad,
+        "zeropad_output_shape": zeropad_output_shape
+    },
+    compile=False
+)
+
+# ECG class mappings
+CLASSES = ["N", "V", "/", "A", "F", "~"]
+CLASS_NAMES = {
+    "N": "Normal sinus beat",
+    "V": "Premature Ventricular Contraction (PVC)",
+    "/": "Paced beat (Pacemaker)",
+    "A": "Atrial Premature Beat",
+    "F": "Fusion of Ventricular & Normal Beat",
+    "~": "Unclassifiable / Noise"
+}
+
+@app.get("/")
+async def root():
+    return {"message": "✅ ECG Inference API is running successfully!"}
+
+
+@app.post("/predict-ecg/")
+async def predict_ecg(file: UploadFile = File(...)):
+    """
+    Accepts a CSV or TXT file containing ECG signal samples.
+    Each value should be a single float per line.
+    """
+    content = await file.read()
+    text = content.decode("utf-8").strip().splitlines()
+
+    # Parse numeric values
+    try:
+        data = np.array([float(x.strip()) for x in text if x.strip() != ""])
+    except Exception:
+        return {"error": "Invalid file format. Please upload numeric ECG values only."}
+
+    # Normalize signal length (model expects 256 samples)
+    max_len = 256
+    if len(data) > max_len:
+        data = data[:max_len]
+    elif len(data) < max_len:
+        data = np.pad(data, (0, max_len - len(data)))
+
+    data = data.reshape(1, max_len, 1)
+
+    # Run model inference
+    preds = model.predict(data, verbose=0)
+    label_idx = int(np.argmax(preds))
+    confidence = float(np.max(preds))
+    label = CLASSES[label_idx]
+    description = CLASS_NAMES[label]
+
+    return {
+        "label": label,
+        "description": description,
+        "confidence": round(confidence, 4),
+        "samples_used": len(data[0])
+    }