#!/usr/bin/env python
# coding: utf-8

# # PyCaret 2 Classification Example
# This notebook is created using PyCaret 2.0. Last updated : 31-07-2020

# In[1]:


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


# # 1. Data Repository

# In[2]:


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


# In[4]:


data = get_data('juice')


# # 2. Initialize Setup

# In[5]:


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


# # 3. Compare Baseline

# In[6]:


best_model = compare_models()


# # 4. Create Model

# In[7]:


lr = create_model('lr')


# In[8]:


dt = create_model('dt')


# In[9]:


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


# In[10]:


models()


# In[11]:


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


# In[12]:


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


# # 5. Tune Hyperparameters

# In[13]:


tuned_lr = tune_model(lr)


# In[14]:


tuned_rf = tune_model(rf)


# # 6. Ensemble Model

# In[15]:


bagged_dt = ensemble_model(dt)


# In[16]:


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


# # 7. Blend Models

# In[17]:


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


# # 8. Stack Models

# In[18]:


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


# # 9. Analyze Model

# In[19]:


plot_model(rf)


# In[20]:


plot_model(rf, plot = 'confusion_matrix')


# In[21]:


plot_model(rf, plot = 'boundary')


# In[42]:


plot_model(rf, plot = 'feature')


# In[43]:


plot_model(rf, plot = 'pr')


# In[44]:


plot_model(rf, plot = 'class_report')


# In[22]:


evaluate_model(rf)


# # 10. Interpret Model

# In[23]:


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


# In[24]:


interpret_model(catboost)


# In[25]:


interpret_model(catboost, plot = 'correlation')


# In[26]:


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


# # 11. AutoML()

# In[27]:


best = automl(optimize = 'Recall')
best


# # 12. Predict Model

# In[28]:


pred_holdouts = predict_model(lr)
pred_holdouts.head()


# In[30]:


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


# # 13. Save / Load Model

# In[31]:


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


# In[32]:


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


# In[33]:


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


# In[34]:


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


# # 14. Deploy Model

# In[35]:


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


# # 15. Get Config / Set Config

# In[36]:


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


# In[37]:


get_config('seed')


# In[38]:


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


# In[39]:


get_config('seed')


# # 16. MLFlow UI

# In[ ]:


# !mlflow ui


# # End
# Thank you. For more information / tutorials on PyCaret, please visit https://www.pycaret.org