Azure Machine Learning supports web service parameters
torsdag 13. november 2014
Azure Machine Learning now allows input parameters to be defined when creating Azure ML Experiments for module properties, such as table names, account names, and paths to blob storage. Subsequently, the consumers of the web service can pass the values for these parameters with web service requests. The parameters can be defined as required or optional. (Optional parameters do not need to be passed in the web service call.) Code samples on the API Help page have been updated, and the samples and the test dialogue of the published web service now include the defined parameters. To set a web service parameter:
- Click the icon next to the property name, and then select Set as web service parameter.
- To provide a default value and make the parameter optional, click the icon next to the created parameter and select Provide a default.
- Certain properties cannot be set as parameters, including Column Selector options and any top-level property that would change the options for a group of properties, such as the data source property.
- The password is a special case and does not allow a default to be set. Therefore, if it set as a parameter, it is always required (for security reasons).