The IdentifierIssuer currently uses a counter to ensure that e.g. blank nodes are assigned a unique id when flattening a JSON-LD file in jsonld.js.
This does not work very well when using jsonld.js as backend for parsing JSON-LD into a graph representation (e.g in rdflib.js).
Every time a new file is loaded, the counter will start from zero, and blank node identifiers are created that already exist in the graph.
It would be great if the identifier issuer would instead of a counter use a unique string generator with a low likelihood of clashes (e.g. UUID or a shortened form thereof).
The
IdentifierIssuercurrently uses a counter to ensure that e.g. blank nodes are assigned a unique id when flattening a JSON-LD file injsonld.js.This does not work very well when using
jsonld.jsas backend for parsing JSON-LD into a graph representation (e.g inrdflib.js).Every time a new file is loaded, the counter will start from zero, and blank node identifiers are created that already exist in the graph.
It would be great if the identifier issuer would instead of a counter use a unique string generator with a low likelihood of clashes (e.g. UUID or a shortened form thereof).