register
Description
This command lets you regsiter a Python function into your data warehouse as a User Defined Function (UDF) or a Stored Procedure (SProc).
Modelstar provides a bunch of builtin ML functions that can be deployed and used in your warehouse. For a list of the functions, refer to: ML-SQL-Functions
Command
For registering a Modelstar builtin function:
modelstar register <modelstar_function_pointer>
Synopsis
<modelstar_function_pointer>
This refers to a colon separated function name reference that you would like to register.
Available builtin ML functions are:
forecast:univariate_time_series_forecast: univariate_time_series_forecast
Example
Usage of the command
$ modelstar register forecast:univariate_time_series_forecast
Output of the command
Registering `univariate_time_series_forecast` into MODELSTAR_TEST.PUBLIC of your data warehouse.
+----------------------------------------------------------------+
| status |
+----------------------------------------------------------------+
| Function UNIVARIATE_TIME_SERIES_FORECAST successfully created. |
+----------------------------------------------------------------+
Stored Procedure available at: `MODELSTAR_TEST.PUBLIC`
Notes
The function is regsitered to the database, schema and the stage that is set in the configuration file modelstar.config.yaml. Before you run this, you need to first set the session using the modelstar use <session-config-name> described here.