Update app.py

app.py

@@ -3,29 +3,70 @@ import pandas as pd
 import numpy as np
 import joblib
 import onnxruntime as ort
+import os
+import logging
 
-# Load the ONNX model and scaler outside the function for efficiency
+# Configure logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+
+# Set feature names. CRUCIAL - must match your training data
+feature_names = ['Age', 'Sex', 'CD4+ T-cell count', 'Viral load', 'WBC count', 'Hemoglobin', 'Platelet count']
+
+# Initialize model and scaler (set to None initially)
+ort_session = None
+scaler = None
+model_loaded = False
+scaler_loaded = False
+
+# --- Attempt to Load Model and Scaler ---
 try:
-    ort_session = ort.InferenceSession("hiv_model.onnx")
-    scaler = joblib.load("hiv_scaler.pkl")
-    feature_names = ['Age', 'Sex', 'CD4+ T-cell count', 'Viral load', 'WBC count', 'Hemoglobin', 'Platelet count']  # Match your training data
+    # 1. Set the current working directory (as a precaution)
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+    os.chdir(script_dir)
+    logging.info(f"Current working directory set to: {os.getcwd()}")
+
+    # 2. Check if files exist
+    model_path = "hiv_model.onnx"
+    scaler_path = "hiv_scaler.pkl"
+
+    if not os.path.exists(model_path):
+        logging.error(f"Model file not found: {model_path}")
+        raise FileNotFoundError(f"Model file not found: {model_path}")
+
+    if not os.path.exists(scaler_path):
+        logging.error(f"Scaler file not found: {scaler_path}")
+        raise FileNotFoundError(f"Scaler file not found: {scaler_path}")
+
+    # 3. Load the model and scaler
+    ort_session = ort.InferenceSession(model_path)
+    scaler = joblib.load(scaler_path)
 
     model_loaded = True
     scaler_loaded = True
-except Exception as e:
-    print(f"Error loading model or scaler: {e}")
+
+    logging.info("Model and scaler loaded successfully.")
+
+except FileNotFoundError as e:
+    logging.error(f"File not found: {e}")
+    ort_session = None
+    scaler = None
     model_loaded = False
-    scaler_loaded = False
+    scaler_loaded = False # Make sure these are false if loading fails!
+
+except Exception as e:
+    logging.exception(f"An error occurred while loading the model or scaler: {e}")
     ort_session = None
     scaler = None
-    feature_names = []  # Set to empty to avoid errors later
+    model_loaded = False
+    scaler_loaded = False
+    # Log the full exception traceback for debugging
+# --- End Model Loading ---
 
 def predict_risk(age, sex, cd4_count, viral_load, wbc_count, hemoglobin, platelet_count):
-    """
-    Predicts HIV risk probability based on input features.
-    """
+    """Predicts HIV risk probability based on input features."""
+
     if not model_loaded or not scaler_loaded:
-        return "Model or scaler not loaded. Please ensure 'hiv_model.onnx' and 'hiv_scaler.pkl' are in the same directory."
+         return "Model or scaler not loaded.  Check the logs for errors. Ensure 'hiv_model.onnx' and 'hiv_scaler.pkl' are in the same directory."
 
     try:
         # 1. Create a DataFrame
@@ -41,8 +82,8 @@ def predict_risk(age, sex, cd4_count, viral_load, wbc_count, hemoglobin, platele
         input_df = pd.DataFrame(input_data)
 
         # 2. Standardize the data
-        scaled_values = scaler.transform(input_df[feature_names])
-        scaled_df = pd.DataFrame(scaled_values, columns=feature_names)
+        scaled_values = scaler.transform(input_df[feature_names]) #Use ALL features now.
+        scaled_df = pd.DataFrame(scaled_values, columns=feature_names)  #Use ALL feature names now.
 
         # 3. ONNX Prediction
         input_array = scaled_df[feature_names].values.astype(np.float32)  # Enforce float32
@@ -56,8 +97,8 @@ def predict_risk(age, sex, cd4_count, viral_load, wbc_count, hemoglobin, platele
         return f"High Risk Probability: {risk_probability:.4f}"
 
     except Exception as e:
-        return f"An error occurred during prediction: {e}"
-
+        logging.exception(f"An error occurred during prediction: {e}")
+        return f"An error occurred during prediction: {e}.  Check the logs for details."
 
 # Define Gradio inputs
 age_input = gr.Number(label="Age", value=30)