feature/model update v3

[fdiehr]

PR Context

We need to update our metadata model due to new insights deriving from a) consulting more standards and schemas specific to the modelling of health data, like upcoming profile "HealthDCAT-AP" in relation to the development of the EHDS (European Health Data Space).
Further, we introduce a new entity called "bibliographic reference". This new entity allows us to describe bibliographic references, e.g. publications (books, book sections, journal articles). With that, we added a new core entity to our metadata catalogue, which provides valuable information about RKI's research.

Added

• Entity bibliographic reference
• Vocabulary bibliographic-reference-type: a list of different publication types.
• Vocabulary resource-creation-method: we want to have a better distinction between the usage of resourceTypeGeneral as a generic type for resources and the context/method of how the resources were created.
• Vocabulary personal-data: we want to be able to say, if a resource contains personal data or not.
• Added JSON Schema property description to all entities and properties, since we delete the specification of the model (docs/index.md)

resource

• Property conformsTo: Providing information about standards used with the resource.
• Property doi: For published resources.
• Property hasLegalBasis: Information about the legal basis on which the data was processed; inspired by HealthDCAT-AP draft and Data Privacy Vocabulary
• Property hasPersonalData: Information if the resource contains personal data or not ;inspired by HealthDCAT-AP draft and Data Privacy Vocabulary
• Property maxTypicalAge and minTypicalAge: Information about the age of the researched population; inspired by HealthDCAT-AP draft
• Property populationCoverage: Information about the researched population; inspired by HealthDCAT-AP draft
• Property resourceCreationMethod: Information about the context or method used in order to create the resource

Changes

all entities

• Property identifierInPrimarySource has now a "minLength" and "maxLength" and has a pattern to restrict newlines

variable

• Property dataType does not expect a vocabulary item anymore, but a string

activity

• Property publication is now expecting the identifier of bibliographic-reference item

resource

• Property publication is now expecting the identifier of a bibliographic reference item
• Property loincId: new pattern

distribution

• Properties accessURL and downloadURL now have the type "array", because there can be multiple repositories for accessing the distribution (e.g. GitHub, Zenodo).
• Deleted some properties: We want only to have properties that are relevant for the distribution's access. Hence, all properties concerning actors involved in the creation of the resource were deleted.

vocabularies

• theme: It had too many concepts and was not suitable for our purposes. We decided to delete most of the concepts and keep only the main topics of RKI.
• activity: Deleted concepts due to discussions: We want to focus on the distinction between third-party funded projects and internal projects.
• resource-type-general: Deleted and added concepts due to discussions: We have consulted other standards and schemas (e.g. NFDI4Health, DataCite) to find more suitable resource types; although we didn't find a perfect fit for the vocabulary, we found a few useful concepts and added some of them.

date properties

• Defined new patterns for dates: YearMonthDay, YearMonth, Year. Delete "format": "date" and "format": "datetime" to ensure validation.

index.md
Deleted all information that were relevant to the "specification" of the model. We decided, that the maintenance of the specification is not worth its benefit (mainly interoperability with standards). In the future, it is more feasible and fitting to our use case to maintain mappings to the portals we will contribute to.
But we want to keep the index file, because we may add some more description on the schema on a more general level.

README.md

• Transferred contact and publisher information from index.md to README

Removed

• Vocabulary data-type.json because we decided against having a controlled vocabulary for data types.

[LeipoldL]

changes approved