s

app.py

@@ -5,68 +5,59 @@ import numpy as np
 import lightgbm as lgb
 from autogluon.tabular import TabularPredictor
 
-loaded_model = pickle.load(open('model.pkl', 'rb'))
-
-def relapse(age,  early_stage = 0 , pathology = 'Mixed', B_symptoms = 0):
-    
-    # default values
-    pathology_nod = 0
-    pathology_others = 0
-    
-    if early_stage: 
-        early_stage = 1
-
-    if pathology == 'Nodular':
-        pathology_nod = 1
-        
-    if pathology == 'Others':
-        pathology_others = 1
-    
-    if B_symptoms: 
-        B_symptoms = 1
-    
-    # I will multiply age by std dev and then add the mean to reverse standard scalar 
-    
-    # I am using data to get the mean and std dev which is probably not right. 
-
-    age = (age - 14.892) / 1.635   
-
-    X_test_radio = pd.DataFrame({'Age' : age, 'early_advanced_early': early_stage,
-                                 'pathology_Nodular Sclerosis cHL' : pathology_nod,
-                                 'pathology_others': pathology_others , 'B_symptoms_Yes': B_symptoms,
-                                 'radio_Yes': 1, 'second_PET' : 1}, index = [0])
-
-    X_test_no_radio = pd.DataFrame({'Age' : age, 'early_advanced_early': early_stage,
-                                    'pathology_Nodular Sclerosis cHL' : pathology_nod,
-                                    'pathology_others': pathology_others, 'B_symptoms_Yes': B_symptoms,
-                                    'radio_Yes': 0, 'second_PET' : 1}, index = [0])
-    
-
-    pred_proba_radio = loaded_model.predict_proba(X_test_radio)[:, 1]
+loaded_model = pickle.load(open('Autogluon/models/XGBoost_BAG_L1/model.pkl', 'rb'))
+# pred_proba_radio = loaded_model.predict_proba(X_test_radio)
+# pred_proba_radio
+
+
+def relapse(age, pathology, B_symptoms):
+    
+    X_test_radio = pd.DataFrame.from_dict(
+        {
+            "Age": [age],
+            "pathology_Nodular Sclerosis cHL": [1 if pathology == 'Nodular' else 0],
+            "pathology_others": [1 if pathology == 'Others' else 0],
+            "B_symptoms_Yes" : [1 if B_symptoms else 0],
+            'radio_Yes': [1]
+        }
+    )
+    
+    X_test_no_radio = pd.DataFrame.from_dict(
+        {
+            "Age": [age],
+            "pathology_Nodular Sclerosis cHL": [1 if pathology == 'Nodular' else 0],
+            "pathology_others": [1 if pathology == 'Others' else 0],
+            "B_symptoms_Yes" : [1 if B_symptoms else 0],
+            'radio_Yes': [0]
+        }
+    )    
+
+    pred_proba_radio = loaded_model.predict_proba(X_test_radio)
     
     pred_proba_radio =  round(np.ndarray.item(pred_proba_radio),2)
 
     pred_radio = loaded_model.predict(X_test_radio)
     
-    pred_proba_no_radio = loaded_model.predict_proba(X_test_no_radio)[:, 1]   
+    pred_proba_no_radio = loaded_model.predict_proba(X_test_no_radio)  
     
     pred_proba_no_radio =  round(np.ndarray.item(pred_proba_no_radio),2)
 
     pred_no_radio = loaded_model.predict(X_test_no_radio)
     
-    return pred_proba_radio ,pred_proba_no_radio 
+    return {"Radio": pred_proba_radio, "No Radio": pred_proba_no_radio}
 
-iface = gr.Interface(title = 'Should we omit radiotherapy?', 
+iface = gr.Interface(
+    
+    title = 'Should we omit radiotherapy?', 
                      description = 'This model predicts relapse according to risk factors.', 
                     fn=relapse, 
                      inputs= [
                          gr.Number(label = 'Age', show_label = True),
-                         gr.Checkbox( label = 'Early stage', show_label = True),
-                         gr.Radio( choices = ['Mixed cellularity', 'Nodular', 'Others'], label = 'Pathology',
-                                      show_label = True),
+        gr.Radio( choices = ['Mixed cellularity', 'Nodular', 'Others'], label = 'Pathology', show_label = True),
                         gr.Checkbox(label = 'B_symptoms', show_label = True)],
-                         outputs= [gr.Label(label = 'Relapse probability with RTH'),
-                                  gr.Label(label = 'Relapse probability without RTH')],live = True)
-
+    outputs = gr.Label(label = 'Has a higher probability of relapse', show_label = True)
+    ,live = True,
+    interpretation="default",
+)
 
 iface.launch()