skodapetr/data-schema

Data Schema

Data schema is designed to be a re-usable data description focused model.

data-schema's Issues

Introduce a serialization container

Serialization container represents a serialization into a given format.

We consider container to be format specific.
As a result we distinguish container from an Entity.
We can use container to represent a data serialization for a particular format.
This would allow for format specific options (named graph, root node, array, ...).

Codelists, patterns and value limitations

A user should be able to limit values of certain type to a predefined list (codelist) or pattern.

We should work with types not properties, since we see values from a certain codelist as a type not a feature of a property.

Other application is to limit range of a number.

Allow for domain specific names to technical types

A user can specify RDF types for an entity or URL for a predicate.

We must not use IRIs of RDF classes directly as it would render it impossible to use it in multiple schemas.
For example SHACL use custom resources and only link the the class.

The class similar to the technical name.
Yet unlike technical name it make sense to have multiple types for an entity.

Prisma schema

https://www.prisma.io/docs/orm/prisma-schema

Partial entity definition

A single object type can be defined at multiple places.
We need to be able to aggregate those definitions while preserving provenance information.

Consider support for code generation.

We should be able to generate code based on the data model.

LinkMK

https://linkml.io/linkml/index.html