categorias={'workclass':{'gov':['State-gov','Local-gov','Federal-gov']},
            'education':{'pre-college':['5th-6th','11th','10th','1st-4th',
            			'12th','7th-8th','9th','Preschool'],
                         'assoc':['Assoc-acdm','Assoc-voc'],
                        'HS-grad':['Doctorate','HS-grad','Masters']},
            'marital.status':{'single':['Never-married','Divorced','Widowed',
            			'Married-spouse-absent', 'Separated']},
            'occupation':{'other':['Other-service','Priv-house-serv']},
            'native.country':{'countrys':nc}}

def preprocesamiento(df):
    df['capital.gain'] = df['capital.gain'].apply(lambda x: np.log(x + 1))
    df['capital.loss'] = df['capital.loss'].apply(lambda x: np.log(x + 1))
    colC = (df.describe(include=np.object).columns)
    columnsC = colC.tolist()
    for c in columnsC:
        df[c] = df[c].astype('category')
        df[c] = df[c].cat.codes
    if len(list(df.keys()))>12:
        income_raw = df['income']
        income = income_raw.apply(lambda x: 1 if x == 1 else 0)
        df['income'] = income
    return df

def treeModel(df):
    X= df.drop('income', axis=1)
    y = df['income']
    oversample = SMOTE()
    X, y = oversample.fit_resample(X, y)
    Xtrain, Xtest, ytrain, ytest =train_test_split(X,y,stratify=y)
    b_tree = DecisionTreeClassifier(criterion='entropy',max_depth=20,
                                    splitter='random',max_features=12,
                                    min_samples_split=6)
    b_tree.fit(x,y)
    y_pred = b_tree.predict(Xtrain)
    y_predT = b_tree.predict(Xtest)
    f1_t = f1_score(ytrain,y_pred,average='weighted')
    f1_e = f1_score(ytest,y_predT,average='weighted')
    at = accuracy_score(ytrain,y_pred)
    ae = accuracy_score(ytest,y_predT)
    return {'f1_t':f1_t,'f1_e':f1_e,'at':at,'ae':ae}

def adaboostTree(df):
    X = df.drop('income', axis=1)
    y = df['income']
    oversample = SMOTE()
    X, y = oversample.fit_resample(X, y)
    Xtrain, Xtest, ytrain, ytest =train_test_split(X,y,stratify=y)
    b_tree = DecisionTreeClassifier(criterion='entropy',max_depth=8,
                                    min_samples_leaf=5,min_samples_split=10)
    clf = AdaBoostClassifier(b_tree,learning_rate=0.09, 
                             n_estimators=1000, algorithm='SAMME')
    clf.fit(Xtrain,ytrain)
    y_pred = clf.predict(Xtrain)
    y_predT = clf.predict(Xtest)
    f1_t = f1_score(ytrain,y_pred,average='weighted')
    f1_e = f1_score(ytest,y_predT,average='weighted')
    at = accuracy_score(ytrain,y_pred)
    ae = accuracy_score(ytest,y_predT)
    return {'f1_t':f1_t,'f1_e':f1_e,'at':at,'ae':ae}