Repository navigation
Conversation
…nformation detection.
There was a problem hiding this comment.
🟡 Changes recommended
Providers without explicit URLs are omitted, and distinct providers sharing a homepage are merged.
2 open issues
What changed in this PR
Adds provider metadata extraction so library consumers can classify organizations and follow links for additional metadata.
Changes:
- Extracts provider names, URLs, and deduplicated
see_alsolinks. - Adds optional query joins for related metadata.
- Tests provider extraction and link handling.
| File | Description |
|---|---|
| test/field_test.rb | Tests provider metadata and link variations. |
| test/extraction_test.rb | Verifies providers in extracted results. |
| lib/tess/rdf/extraction.rb | Adds provider extraction and optional joins. |
Review effort: Balanced
| [:provider, RDF::Vocab::SCHEMA.url, :url], | ||
| [:provider, RDF::Vocab::SCHEMA.sameAs, :same_as, { optional: true }], | ||
| optional: [ | ||
| [:provider, RDF::Vocab::SCHEMA.subjectOf, :subject_of], | ||
| [:subject_of, RDF::Vocab::SCHEMA.url, :subject_of_url, { optional: true }] | ||
| ]).each do |p| |
There was a problem hiding this comment.
I personally would leave it as it is, because to me having an explicit url would indicate a higher level of metadata quality. If needed I could imagine making the url optional, but I don't like the idea of setting the url based on the provider iri.
To be honest, my reasoning about this is highly influenced by the fact that host institutions are currently serialized without url, so the url requirement lets me ignore the problem of distinguishing providers and host institutions.
| [:provider, RDF::Vocab::SCHEMA.subjectOf, :subject_of], | ||
| [:subject_of, RDF::Vocab::SCHEMA.url, :subject_of_url, { optional: true }] | ||
| ]).each do |p| | ||
| provider_key = p[:url] |
There was a problem hiding this comment.
Establishing provider identity based on their URL is consistent with how this code is used for exchanging content providers later on. Two providers should not have the exact same url.
This is needed for ElixirTeSS/TeSS#1191. It is up to the user of this library to distinguish content providers, nodes and organizers. Additional information can be gathered about content providers by extracting the metadata from see_also links.