import category_encoders as ce
ce.TargetEncoder(smoothing=0).fit_transform(df.Color,df.Target)

enc=ce.OneHotEncoder().fit(df.Target.astype(str))
y_onehot=enc.transform(df.Target.astype(str))
y_onehot

class_names = y_onehot.columns
for class_ in class_names:
    enc = ce.TargetEncoder(smoothing = 0)
    print(enc.fit_transform(X,y_onehot[class_]))

def target_encode_multiclass(X,y): #X,y are pandas df and series
    y=y.astype(str)  #convert to string to onehot encode
    enc=ce.OneHotEncoder().fit(y)
    y_onehot=enc.transform(y)
    class_names=y_onehot.columns  #names of onehot encoded columns
    X_obj=X.select_dtypes('object') #separate categorical columns
    X=X.select_dtypes(exclude='object')
    for class_ in class_names:
        enc=ce.TargetEncoder()
        enc.fit(X_obj,y_onehot[class_]) #convert all categorical
        temp=enc.transform(X_obj)       #columns for class_
        temp.columns=[str(x)+'_'+str(class_) for x in temp.columns]
        X=pd.concat([X,temp],axis=1)    #add to original dataset
    return X