Update app.py

app.py

@@ -1,4 +1,3 @@
-
 from flask import Flask, request, jsonify
 import joblib
 import pandas as pd
@@ -6,27 +5,63 @@ import os
 
 app = Flask(__name__)
 
+# Default model path (can be changed via Space environment variable)
 MODEL_PATH = os.getenv("MODEL_PATH", "best_model_random_forest.joblib")
 
+# Load model if it exists
 if not os.path.exists(MODEL_PATH):
-    raise FileNotFoundError(f"Model file not found: {MODEL_PATH}. Please ensure it's uploaded to the Space.")
+    raise FileNotFoundError(
+        f"❌ Model file not found: {MODEL_PATH}. Please ensure it's uploaded to the Space root directory."
+    )
 
 model_bundle = joblib.load(MODEL_PATH)
-if isinstance(model_bundle, dict) and "pipeline" in model_bundle:
-    pipe = model_bundle["pipeline"]
-else:
-    pipe = model_bundle
+
+# Support both raw model or model dict with a pipeline
+pipe = model_bundle["pipeline"] if isinstance(model_bundle, dict) and "pipeline" in model_bundle else model_bundle
+
+@app.route("/", methods=["GET"])
+def home():
+    return jsonify({
+        "status": "ok",
+        "message": "Backend Flask API is running successfully 🎉",
+        "model_loaded": True,
+        "model_path": MODEL_PATH
+    })
 
 @app.route("/health", methods=["GET"])
 def health():
-    return jsonify({"status": "ok", "model": MODEL_PATH})
+    return jsonify({"status": "ok", "model_path": MODEL_PATH})
 
 @app.route("/predict", methods=["POST"])
 def predict():
+    """
+    Example input:
+    {
+        "feature1": value1,
+        "feature2": value2,
+        ...
+    }
+    or
+    [
+        {"feature1": value1, "feature2": value2},
+        {"feature1": value3, "feature2": value4}
+    ]
+    """
     data = request.get_json(force=True)
+
+    # Normalize input to a DataFrame
     df = pd.DataFrame([data]) if isinstance(data, dict) else pd.DataFrame(data)
-    preds = pipe.predict(df)
-    return jsonify({"predictions": [float(p) for p in preds]})
 
+    try:
+        preds = pipe.predict(df)
+        # Convert numpy types to native Python floats
+        predictions = [float(p) for p in preds]
+        return jsonify({"predictions": predictions})
+    except Exception as e:
+        return jsonify({"error": str(e)}), 400
+
+
+# Entry point for local development; Spaces uses gunicorn automatically
 if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=5000)
+    port = int(os.getenv("PORT", 7860))  # HF Spaces use port 7860 by default
+    app.run(host="0.0.0.0", port=port, debug=False)