REST api for curriculum data, uses the GraphQL server in https://github.com/slonl/curriculum-graphql-server. It is written in javascript, using nodejs, with the express framework.
The REST api uses the GraphQL server as a backend, the reverse of the traditional role. The advantage is that the code base is very small and simple. You can easily extend / alter the REST api. Each route is generally a thin wrapper around a specific GraphQL query.
The exception is the legacy/ route. The source data was less strict in how the UUID's were used. This meant that you needed all the parent ID's to uniquely identify the data linked to a specific UUID. This dataset was split so that each UUID is now uniquely tied to a single set of data. When querying such replaced entities by their original UUID, you get a replacedBy array that lists all new ID's linked to the original entity. To find the correct ID that matches the data in the legacy dataset, you must specify its original parent ID's as well. e.g.:
/legacy/vak/bk:2b363227-8633-4652-ad57-c61f1efc02c8/vakkern/bk:25661339-1281-434c-be08-b7f5ac8cd432/vaksubkern/bk:f719f7e0-ba85-493d-8d78-b0910a38c7d8
The legacy route will find the new entities for each of the parent ID's and filter the result entities to only those that have the same (updated) parent relations.
You can read more about the data architecture and the replacedBy/replaces mechanic in the dataset repositories: