v.0.10

On 2022.08.17
This version has been deprecated. If used with the current API version it can produce unexpected behaviour or errors.
To get the new version  🎁
​🎁
​🎁
​
pip install --upgrade shimoku-api-python
pip install --upgrade shimoku-components-catalog
This version brings Machine Learning features through our API by the first time!!
Fixes
For readability, renamed s.file.* by s.io.*: 
Before: s.file.get_dataframe(...) 
Now: s.io.get_dataframe(...)
Fixed IO for large dataframes now you can POST and GET dataframes of large sizes without size limit.
Fix path and subpaths. Now Apps with multiple words can be sorted too: s.plt.set_apps_orders() and s.plt.set_sub_path_orders() 
Fixes in the default look&feel of different charts and with new features to see data, download image, zoom in and out and change chart type:
Heatmap s.plt.heatmap()
Barchart s.plt.bar() and Linechart s.plt.line() 
Improvements
Improvements in the Input Forms look&feel
Easier to initialize the Shimoku Client
import shimoku_api_python as Shimoku
​
# Still allowed
s = Shimoku.Client(
    config={'access_token': api_key},    
    universe_id=universe_id,    
    environment=environment
)
​
# Easier
s = Shimoku.Client(
    access_token=api_key,
    universe_id=universe_id,    
    environment=environment
)
Added IO methods for Machine Learning models:
from sklearn import svm
from sklearn import datasets
​
clf = svm.SVC()
X, y = datasets.load_iris(return_X_y=True)
clf.fit(X, y)
​
s.io.post_ai_model(
    business_id=business_id,
    app_name='test',
    model_name='model-object-test',
    model=clf,
)
​
model = s.io.get_ai_model(
    business_id=business_id,
    app_name='test',
    model_name='model-object-test',
)
New
Added method to clear a whole business (tabula rasa) shimoku.plt.clear_business()
Before
s.plt.clear_business()
Added method to retrieve input forms data
input_data: List[Dict] = s.plt.get_input_forms(menu_path='test/input-form')
​
​
Machine Learning!
AI for Machine Learning features. Now users can use Shimoku's Machine Learning from the SDK just passing a test dataframe and specifying the Model endpoint (ask to Shimoku about what endpoints are available). You can decide whether to add Explainability or not to the predictions
df_pred, df_error = s.ai.predict_categorical(
    df_test=df_test_fail,
    model_endpoint=model_endpoint,
    explain=False,
)
Also predictive tables have been added that create a table with the prediction
target_column: str = 'NRO_POL'
column_to_predict: str = 'churn_probability'
menu_path: str = 'AI/Churn'
order: int = 0
​
s.ai.predictive_table(
   df_test=df_test,
   model_endpoint=model_endpoint,    
   target_column=target_column,
   column_to_predict=column_to_predict,  
   prediction_type='categorical',
   explain=True,  
   menu_path=menu_path, order=order,    
   add_filter_by_column_to_predict=False,
   add_search_by_target_column=True,
   extra_filter_columns=None,
   extra_search_columns=None
)
v.0.11
v.0.10.4
Last modified 5mo ago