Manual Definition of the Service Metamodel

In Asset -> Service we have seen how to define Service Properties for the Service using the graphical interface. Behind the scenes, the Data Analytics System converts what is entered interactively into a JSON metamodel. This JSON can also be manually defined and loaded from the UI.

To access the metamodel JSON editor, go to the Service registration panel and click on Service Definition

The editor for entering the metamodel will open:

Attribute	Description
name	Nunique name of the Service
version	Service version
description	Brief textual description of the Service
framework	Technology used to implement the Service
url	The URL of the Docker image of the Service, with prefix docker:// and version tag. Example: docker://dockerhub.alidalab.it/alida/restricted/services/great-expectations:1.1.0
accessLevel	Service visibility level
properties	property Array (the list of parameters and I/O ports of the Service)
area	deprecated
markdown	Textual documentation in markdown format associated with the Service and displayed in the UI

For the generic property, an element of the properties array, the attributes are:

Attribute	Description
key	Unique identifier of the property. Used as a reference in code and UI
description	Descriptive text shown in UI
type	Type of property (application, resource, static). Determines the behavior of the property
mandatory	Indicates if it is mandatoryv(true | false)
defaultValue	Default Value. Always string/number; JSON serialized as string
valueType	Expected value type (string | int | json)
invisible	Whether to show the property in the UI or not (true | false ). true for the ports input/output
model	Additional information if the property is key = input-model { "format": "sklearn" }

For model, the unique attribute is: | Attribute | Description | | --------- | ----------- | | format | sklearn | huggingface | etc ... | |

General JSON Syntax for the Metamodel