Update app.py

app.py

@@ -20,21 +20,24 @@ from sklearn.tree import DecisionTreeClassifier
 from sklearn.ensemble import RandomForestClassifier
 
 def read(file,dep,ord):
-    df = pd.read_csv(file.name)
+    df = file
     cat = list()
     dep_type = str(df.dtypes[dep])
     for col in df.columns.values:
         if str(df.dtypes[col]) == 'bool' or str(df.dtypes[col]) == 'object':
             cat.append(col)
-    if ord == "":
+    new_df = df.dropna(axis=0)
+    if ord == "" and (dep_type == 'bool' or dep_type == 'object'):
+        ord = list()
+        ord.append(dep)
+    elif ord == "":
         ord = list()
     else:
-        ord = ord.split(',')
-    ord.append(dep)
+        pass
+    if len(ord)!=0:
+        le = LabelEncoder()
+        new_df[ord] = new_df[ord].apply(lambda col: le.fit_transform(col))
     nom = list(set(cat).difference(set(ord)))
-    le = LabelEncoder()
-    new_df = df.dropna(axis=0)
-    new_df[ord] = new_df[ord].apply(lambda col: le.fit_transform(col))
     if len(nom) == 0:
         pass
     else:
@@ -48,7 +51,7 @@ def read(file,dep,ord):
         text = "regression"
         result = regression(new_df,dep)
     return df.sample(5),new_df.sample(5),result, text, cat, ord, nom
-
+    
 def classification(df,dep):
     X = df.drop(dep,axis=1)
     y = df[dep]
@@ -113,12 +116,12 @@ def regression(df,dep):
         },
     ]
 
-    search = GridSearchCV(pipe, param_grid=parameters, cv=5, n_jobs=-1, scoring='r2')
+    search = GridSearchCV(pipe, param_grid=parameters, cv=5, n_jobs=-1, scoring='neg_mean_absolute_percentage_error')
     search.fit(X_train,y_train)
 
     result = pd.DataFrame(search.cv_results_)[['params','rank_test_score','mean_test_score']]
     
-    result['mean_test_score']= (result['mean_test_score'])*100
+    result['mean_test_score']= (result['mean_test_score']+1)*100
     result = result.astype({'params': str})
     
     result.sort_values('rank_test_score',inplace=True)