Upload app.py with huggingface_hub

app.py

@@ -1,9 +1,8 @@
 import gradio as gr
 import pickle
 import pandas as pd
-import xgboost as xgb
 
-# 1. Load Model Bundle XGBoost
+# 1. Load Model
 with open("best_model_xgb.pkl", "rb") as f:
     bundle = pickle.load(f)
 
@@ -11,30 +10,31 @@ model = bundle['model']
 impute_col = bundle['impute_col']
 feature_names = bundle['feature_names']
 
-# 2. Fungsi Prediksi untuk API
-def predict_api(*args):
-    # Mapping input ke dictionary berdasarkan nama fitur
-    input_data = dict(zip(feature_names, args))
-    df = pd.DataFrame([input_data])
+# 2. Fungsi Prediksi yang Menerima 1 JSON (Dictionary)
+def predict_api(input_dict):
+    # Langsung ubah dictionary yang diterima dari mobile ke DataFrame
+    df = pd.DataFrame([input_dict])
     
-    # Handle missing values berdasarkan impute_col
+    # Pastikan urutan kolom sesuai dengan yang diharapkan XGBoost
+    df = df.reindex(columns=feature_names)
+    
+    # Handle missing values
     for col in impute_col:
         if pd.isna(df[col].iloc[0]):
-            df[col] = 0  # Sesuaikan dengan nilai imputasi jika dibutuhkan
+            df[col] = 0  # Sesuaikan logika imputasi
             
-    # Prediksi dengan XGBoost
+    # Eksekusi
     prediction = model.predict(df)
     return float(prediction[0])
 
 # 3. Setup UI Gradio
-inputs = [gr.Number(label=col) for col in feature_names]
-
+# Kita ganti inputs menjadi gr.JSON agar API menerima dictionary utuh
 app = gr.Interface(
     fn=predict_api,
-    inputs=inputs,
+    inputs=gr.JSON(label="Input Data (JSON)"), 
     outputs=gr.Number(label="Hasil Prediksi"),
-    title="NutriSnapS XGBoost Prediction API",
-    api_name="predict_kidney"  # Endpoint API sudah disesuaikan
+    title="NutriSnapS XGBoost API",
+    api_name="predict_kidney"
 )
 
 if __name__ == "__main__":