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

# In[ ]:


# !pip install pycaret


# In[ ]:


# !pip install pycaret[full] 


# In[1]:


from pycaret.utils import version
version()


# In[2]:


from pycaret.datasets import get_data
data = get_data('juice')


# In[3]:


data.dtypes


# In[4]:


data['Purchase'].value_counts()


# In[5]:


from pycaret.classification import *
s = setup(data, target = 'Purchase', log_experiment = True, experiment_name = 'dsc1',
          fold = 3)


# In[6]:


xgboost = create_model('xgboost')


# In[7]:


deploy_model(xgboost, 'dsc123', platform = 'aws', authentication = {'bucket' : 'pycaret-test'})


# In[ ]:





# In[ ]:





# In[ ]:





# In[ ]:


# add_metric()


# In[ ]:


get_ipython().run_cell_magic('time', '', 'best = compare_models()\n')


# In[ ]:


from sklearn.linear_model import LogisticRegression


# In[ ]:


asd = LogisticRegression()


# In[ ]:


create_model(asd)


# In[ ]:





# In[ ]:


lr = create_model('lr')


# In[ ]:


tuned_lr = tune_model(lr)


# In[ ]:


tuned_lr = tune_model(lr, search_library = 'optuna')


# In[ ]:


tuned_lr = tune_model(lr, search_library = 'scikit-optimize')


# In[ ]:


dt = create_model('dt')


# In[ ]:


dt


# In[ ]:


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


# In[ ]:


bagged_dt


# In[ ]:


plot_model(lr, plot = 'feature')


# In[ ]:


evaluate_model(lr)


# In[ ]:


dt = create_model('dt')
plot_model(dt, plot = 'boundary')


# In[ ]:


xgboost = create_model('xgboost')


# In[ ]:


interpret_model(xgboost)


# In[ ]:


save_model(xgboost, model_name = 'abc')


# In[ ]:


get_ipython().system('mlflow ui')


# In[ ]:





# In[ ]: