Update app.py

app.py

@@ -1,22 +1,67 @@
-from fastapi import FastAPI
-from pydantic import BaseModel
-import uvicorn
+from fastapi import FastAPI, UploadFile, File
+from fastapi.responses import JSONResponse
+import pandas as pd
+from sklearn.linear_model import LogisticRegression
+from sklearn.preprocessing import StandardScaler
+import io
+import joblib
 
-# تعريف الـ app
-app = FastAPI(title="My AI API", description="Permanent FastAPI endpoint", version="1.0")
+# ===== FastAPI app =====
+app = FastAPI(title="Employee of the Month API", version="1.0")
 
-# نموذج البيانات اللي هتجيلك من المستخدم
-class InputData(BaseModel):
-    text: str
+# ===== تدريب الموديل مرة واحدة عند تشغيل السيرفر =====
+df_train = pd.DataFrame({
+    'EmployeeID': [101, 102, 103, 104, 105, 106],
+    'PerformanceScore': [90, 85, 95, 80, 88, 92],
+    'ProjectsCompleted': [5, 6, 7, 4, 6, 5],
+    'Attendance': [98, 92, 95, 90, 97, 96],
+    'EmployeeOfTheMonth': [0, 0, 1, 0, 0, 0]
+})
 
-# الدالة اللي بتتعامل مع البيانات
+X_train = df_train[['PerformanceScore', 'ProjectsCompleted', 'Attendance']]
+y_train = df_train['EmployeeOfTheMonth']
+
+scaler = StandardScaler()
+X_train_scaled = scaler.fit_transform(X_train)
+
+model = LogisticRegression(class_weight='balanced', random_state=42)
+model.fit(X_train_scaled, y_train)
+
+# Optional: حفظ الموديل والمقياس
+joblib.dump(model, 'model.pkl')
+joblib.dump(scaler, 'scaler.pkl')
+
+# ===== Endpoint =====
 @app.post("/predict")
-def predict(data: InputData):
-    text = data.text
-    # هنا بتحط كود الموديل بتاعك الحقيقي 👇
-    result = f"Prediction for: {text}"
-    return {"result": result}
+async def predict_employee(file: UploadFile = File(...)):
+    try:
+        # قراءة Excel
+        contents = await file.read()
+        df_new = pd.read_excel(io.BytesIO(contents))
+        
+        # Scale البيانات الجديدة
+        X_new_scaled = scaler.transform(df_new[['PerformanceScore', 'ProjectsCompleted', 'Attendance']])
+        
+        # توقع الاحتمالات
+        probs = model.predict_proba(X_new_scaled)[:,1]
+        df_new['ProbabilityOfBeingBest'] = probs
+        
+        # اختيار أفضل موظف
+        best_employee = df_new.loc[df_new['ProbabilityOfBeingBest'].idxmax()]
+        
+        # معاملات الموديل
+        coef_dict = dict(zip(['PerformanceScore','ProjectsCompleted','Attendance'], model.coef_[0].round(3)))
+        
+        return JSONResponse({
+            "predictions_all": df_new.to_dict(orient="records"),
+            "best_employee": best_employee.to_dict(),
+            "feature_coefficients": coef_dict
+        })
+    
+    except Exception as e:
+        return JSONResponse({"error": str(e)})
 
-# لتشغيل السيرفر محليًا
+# ===== تشغيل السيرفر محليًا =====
 if __name__ == "__main__":
+    import uvicorn
     uvicorn.run(app, host="0.0.0.0", port=7860)