# check version
from pycaret.utils import version
version()

from pycaret.datasets import get_data
index = get_data('index')

data = get_data('juice')

from pycaret.classification import *
clf1 = setup(data, target = 'Purchase', session_id=123, log_experiment=True, experiment_name='juice1')

best_model = compare_models()

lr = create_model('lr')

dt = create_model('dt')

rf = create_model('rf', fold = 5)

models()

models(type='ensemble').index.tolist()

ensembled_models = compare_models(whitelist = models(type='ensemble').index.tolist(), fold = 3)

tuned_lr = tune_model(lr)

tuned_rf = tune_model(rf)

bagged_dt = ensemble_model(dt)

boosted_dt = ensemble_model(dt, method = 'Boosting')

blender = blend_models(estimator_list = [boosted_dt, bagged_dt], method = 'soft')

stacker = stack_models(estimator_list = [boosted_dt,bagged_dt,tuned_rf], meta_model=rf)

plot_model(rf)

plot_model(rf, plot = 'confusion_matrix')

plot_model(rf, plot = 'boundary')

evaluate_model(rf)

catboost = create_model('catboost', cross_validation=False)

interpret_model(catboost)

interpret_model(catboost, plot = 'correlation')

interpret_model(catboost, plot = 'reason', observation = 12)

best = automl(optimize = 'Recall')
best

pred_holdouts = predict_model(lr)
pred_holdouts.head()

new_data = data.copy()
new_data.drop(['Purchase'], axis=1, inplace=True)
predict_new = predict_model(lr, data=new_data)
predict_new.head()

save_model(lr, model_name='best-model')

loaded_bestmodel = load_model('best-model')
print(loaded_bestmodel)

from sklearn import set_config
set_config(display='diagram')
loaded_bestmodel[0]

from sklearn import set_config
set_config(display='text')

deploy_model(lr, model_name = 'best-aws', authentication = {'bucket' : 'pycaret-test'})

X_train = get_config('X_train')
X_train.head()

get_config('seed')

from pycaret.classification import set_config
set_config('seed', 999)

get_config('seed')

get_system_logs()

!mlflow ui

# to generate csv file with experiment logs
get_logs()