fix: Correctly handle cube metadata

[bprusinowski]

Closes #1138.
Closes #1131.

This PR:

• aligns the way of detecting literal dimensions by using shacl:nodeKind = shacl:Literal. This results in smaller queries when fetching observations and should thus improve overall performance of the application a bit,
• updates the rdf-to-visualize part of the documentation to be up-to-date when it comes to defining isLiteral and isNumerical dimension properties.

@Rdataflow I hope this should align the way we handle cubes with the cube spec – let me know if there is anything else you've identified (looks like the Data types section of the spec is not very extensive).

I didn't introduce any changes to the way we detect isNumerical property, as I believe it's done correctly (the Energy dimension mentioned in #1131 lacks the dataType property altogether, that's why it's marked as not numerical). I believe this should be fixed in the data.

How to test

1. Go to this dataset.
2. Click on Run SPARQL query button.
3. See that the query contains 16 dimensions.
4. Repeat the process on PROD environment.
5. See that this query has 21 dimensions (5 unnecessary, empty ones due to wrong Literal dimension type detection).

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
visualization-tool	✅ Ready (Inspect)	Visit Preview	Aug 24, 2023 0:01am