Metadata attributes

At Anjana Data, the metadata attributes are the basic units of information that describe the objects (entities and relationships) governed on the platform. These attributes allow the technical, functional and governance characteristics of each asset to be documented in a structured way, facilitating search, classification and quality control, among other things.

Attributes are configured in the Attribute Definitions table, and can be reused in different metadata templates, although they can never appear more than once in the same template.

Thanks to this modularity, the organization can create a consistent catalog of reusable attributes that support its governance model.

In Anjana Data, the different types of supported metadata attributes are as follows:

• Array of boolean: attribute to indicate one or more 'true' or 'false' values

• Array of date: attribute to indicate one or more dates

• Array of decimal: attribute to indicate one or more decimal numbers

• Array of entities: attribute to choose one or more entities approved in Anjana

• Array of file: attribute to attach one or more files

• Array of number: attribute to indicate one or more integer numbers

• Array of Organizational Unit: attribute to select one or more organizational units from the list of all of them

• Array of text: attribute to enter one or more short texts of up to 255 characters

• Array of URL: attribute to enter one or more links to browsable URLs

• Array of users: attribute to choose one or more users from the complete list of application users

• Boolean: 'true' or 'false'

• Date: date (year, month, day, hour and minutes)

• Decimal: number with decimals

• Enriched Text Area: attribute to enter a long text of up to 300 thousand characters enriched with formatting (bold, underline, italics, etc.)

• Entity Container: attribute to choose an entity approved in Anjana Data and generate native relationships between entities. This attribute can only be used in the DSA, instance and solution entities

• Entity Search: attribute to choose an entity approved in Anjana Data

• File: file stored internally in Anjana Data. The file can also be downloaded if the user has read permissions

• International Text: regular text box available for the different languages available in the application

• International Text Editor: attribute to enter a long text of up to 300 thousand characters enriched with formatting (bold, underline, italics, etc.) available for the different languages available in the application

• International Textarea: attribute to enter a long text of up to 300 thousand characters in the different languages available for the application

• MultiSelect: attribute to select one or more values from a list preconfigured in the Reference Metadata tab

• MultiSelect with icons: attribute to select one or more icons from a list preconfigured in the Reference Metadata tab

• MultiSelect with icons and text: attribute to select one or more values (icon+text) from a list preconfigured in the Reference Metadata tab

• Number: attribute to indicate an integer number

• Number range: selector for an integer number between a defined minimum and maximum

• Organizational Unit: attribute to select an organizational unit from the list of all of them

• Reference Metadata: list of possible values that must be defined for the attribute

• Selector with icon: list of possible icons that must be defined for the attribute

• Selector with icon and text: list of possible values (icon+text) that must be defined for the attribute

• Single taxonomy: taxonomy tree

• Multiple selection taxonomy: taxonomy tree where one or more values can be selected

• Text: regular text box

• Text Area: attribute to enter a long text of up to 300 thousand characters

• URL: text treated as a URL so the user can click on the attribute and a new tab opens with that URL

• User: complete list of Anjana Data users

Table Attribute Definitions of the Configuration Panel (Administrator view)

Sections are configured in the Sections table of the Configuration Panel. Defining sections is the second step, after creating the menus, to structure the metadata templates.

Structure of the Attribute Definitions

Each registered attribute is characterized by the following fields:

• Id: unique identifier of the attribute.

  • It is assigned automatically from the Configuration Panel.

• Name: internal name of the attribute.

• Type: type of attribute.

  • Determines the nature of the data that can be stored (text, number, date, user, organizational unit, file, taxonomy, etc.).

  • See the equivalence table below.

• Label: visible label of the attribute in the metadata forms.

  • It can be configured for internationalization using its value as config_key in the translations table, with one record per active language in the application.

• Place Holder: example text shown when the field is empty.

  • It can be configured for internationalization in translations.

• shortDescription: brief description of the purpose of the attribute.

  • Optional field, intended to help administrators and validators understand the attribute.

• Descritpion: more extensive description of the attribute.

  • It can be used to detail instructions or additional context and is visible within the template in the icon next to the attribute name.

  • As with Label and Place Holder, it can be configured as a translation key.

Special considerations by attribute type

• Attributes of type SELECT, SELECT_IMG, SELECT_IMG_TXT, TREE_SELECT, MULTI_SELECT, MULTI_SELECT_IMG, MULTI_SELECT_IMG_TXT and TREE_MULTISELECT:

  • They require defining the possible values in the Attribute Definition Value table.

• Attributes of type INPUT_RANGE

  • They must have minimum and maximum validations configured for the range selector to work.

  • In the absence of validations, the default values apply:

    • MIN_RANGE and MAX_RANGE, defined in the appConfiguration table of the Configuration Panel.

• Attributes of type ENRICHED_TEXT_AREA and ENRICHED_TEXT_AREA_INTERNATIONAL

  • They support uploading files with extensions: .gif, .png, .jpg with size restrictions.

Primary key attributes

Some entities of the metamodel require a minimum set of attributes that act as a primary key.

• Dataset: name, infrastructure, path, technology, zone.

• Dataset field: name, infrastructure, path, technology, zone.

• DSA: name.

• Process: name, infrastructure, path, technology, zone.

• Process instance: name, processAri, solutionAri.

• Solution: name.

• Non-native entities defined by the client: name.

• Relationships: name, source, destination.

Mandatory attributes

Some attributes are mandatory for Anjana Data to work correctly, since the platform has internal logic associated with them.

• The name field must match exactly the expected names.

• The attribute_type field must have a specific value.

• Each mandatory attribute must only be applied to the entities or relationships detailed in the mandatory attributes table.

• Adding them to other objects can cause errors in the application.

Mandatory attributes for entities

The following table explains the purpose and type of the attributes indicated in the previous table:

Mandatory attributes for relationships

Mandatory attributes not referenced in templates

Some internal and reserved attributes must be included in Attribute Definition for the platform to work correctly.

Creating an attribute in the Attribute Definitions

Creating a new metadata attribute involves registering in the Attribute Definitions table a new field that can be reused in different entity or relationship templates.

To add a new attribute:

1. Click the New button in the top right corner. This will open a wizard (wizard) with the fields defined in the Structure of the Attribute Definitions table section.

2. Fill in the attribute fields according to the described structure:

   • Name: internal name of the attribute (e.g.: nivelRiesgo).

   • Type: type of attribute (e.g.: SELECT).

   • Label: label visible in the form. In case of internationalization, it must also be created in the translations table for each language.

   • Place Holder: example text when the field is empty (optional). In case of internationalization, it must also be created in the translations table for each language.

   • shortDescription: brief description of the attribute (optional).

   • Description: more extensive description or usage instructions. In case of internationalization, it must also be created in the translations table for each language. (e.g.: (PD) Indicates the level of risk posed by the model according to the taxonomy of the European regulation).

3. Click Save to save the attribute or Cancel to discard it.

Modifying an attribute in the Attribute Definitions

Modifying an attribute in the Attribute Definitions table must be done with special caution, since it can have different impacts on the templates where the attribute is being used and on objects already created in the Data Portal.

Fields that can be modified without high impact:

• Label, Place Holder, Description and shortDescriptioncan be adjusted, since they are only used as informational support in the forms:

  • If internationalization has not been configured, the texts can be modified directly.

  • If internationalization exists, the modification must be done in the translations table, ensuring that the config_key corresponding to it (defined from the Label/Place Holder originals) remain consistent.

Fields whose modification requires caution:

• Type: it is not recommended to modify the type of an attribute once it is in use, since:

  • It can invalidate validations configured in the templates.

  • It can generate inconsistencies in Solr indexing if the new type is not compatible with the previous one (example: changing a numeric attribute to alphanumeric).

  • If it is absolutely necessary to change it, follow the recommendations in the FAQ section.

• Name: must never be changed if the attribute is already in use in any template, since:

  • It can generate inconsistencies in Solr indexing

  • It can generate inconsistencies in the Filters configuration (Filter Confs table).

  • It can generate inconsistencies in the configuration of Depends On type validations (Template Attribute validations table).

  • Errors could occur in workflows.

  • If it were essential to change it, it would be necessary to execute a Reset data in the new Configuration Panel (Inicio > Operaciones de reseteo > Reestablecer datos) and reindex all the information.

Reset data

Configuring attributes via direct database access (Developer view)

The configuration of metadata attributes can also be done directly on the database, in the anjana.attribute_definition table. This table contains the definition of all the attributes reusable in the entity and relationship templates of the metamodel.

Below is an example script to configure a SELECT-type attribute to catalog the risk level of an AI System:

INSERT INTO anjana.attribute_definition
(id_attribute_definition, description, "label", last_modified_date, "name", place_holder, short_description, start_date, attribute_type)
VALUES(10321, '(PD) Indica el nivel de riesgo que supone el modelo de acuerdo con la taxonomía de la regulación europea', 'Nivel de riesgo', '2024-10-11 11:41:09.720', 'nivelRiesgo', NULL, NULL, '2024-10-11 11:41:09.720', 'SELECT');