Emit all URLs as real URLs instead of strings

VladimirAlexiev commented 10 months ago

This pertains to eg:

  datacite:doi "https://doi.org/10.48550/arxiv.1611.03159";
  sp:crossrefType "https://api.crossref.org/types/posted-content"; 
  rdfs:seeAlso "https://en.wikipedia.org/wiki/Computer%20science"; 
  foaf:depiction "https://upload.wikimedia.org/wikipedia/commons/5/52/Conceptual_Diagram_-_Example.svg"; 
  sp:imageThumbnail "https://upload.wikimedia.org/wikipedia/commons/thumb/5/52/Conceptual_Diagram_-_Example.svg/100px-Conceptual_Diagram_-_Example.svg.png" . 
  fabio:hasURL "https://zenodo.org/record/7750707"; 
  sp:pdfUrl "https://zenodo.org/record/7750707/files/Towards%20InnoGraph%20A%20Knowledge%20Graph%20for%20AI%20Innovation.pdf"; 
  foaf:homepage "http://www.nyu.edu/";
  soa:ror "https://ror.org/0190ak572";

I think you know with certainty which OpenAlex fields are URLs?

The only downside is if that OpenAlex includes an INVALID URL, the repo will refuse to ingest this data (unless you turn off URL validation).
for fields where you have such suspicion, you can use a literal with appropriate datatype, eg "https://zenodo.org/record/7750707/files/Towards%20InnoGraph%20A%20Knowledge%20Graph%20for%20AI%20Innovation.pdf"^^xsd:anyURI

linnaung commented 9 months ago

We will implement suggested by you.

VladimirAlexiev commented 6 months ago

This is especially important for identifiers in other KGs, and you should use the exact semantic URLs of that KG. Eg:

<https://semopenalex.org/institution/I200650556> a soa:Institution;
  owl:sameAs makg:200650556, "https://www.wikidata.org/wiki/Q1470769";

The link should become owl:sameAs <https://www.wikidata.org/entity/Q1470769>

<https://semopenalex.org/geo/I200650556> a soa:Geo;
  <http://www.geonames.org/ontology#geonamesID> "3149318";
  <https://dbpedia.org/property/city> "Kristiansand";

The link should become owl:sameAs <https://sws.geonames.org/3149318/>

metaphacts / semopenalex

Emit all URLs as real URLs instead of strings #43