On 06/01/2025 10:59, zPlus wrote:
Apache Jena RIOT version 5.2.0
Debian GNU/Linux 12 (bookworm)
I'm trying to import some nodes with IDs like <file:/path/to/file> into a TDB2
database. However, the IDs are automatically changed to <file:///path/to/file>
once they're added to the database. As far as I understand, the "file:" URI
scheme allows for a missing "host" component, therefore I'm not sure if this is
a bug or not.
The "file:" URI schema has not had the attention it should have had, it
was old and unclear, until RFC 8089 cleaned this up but also has legacy
compatibility.
Both file:/ and and file:/// are URIs with a missing host name.
See
https://datatracker.ietf.org/doc/html/rfc8089.html#section-2
and examples in
https://datatracker.ietf.org/doc/html/rfc8089.html#appendix-B
But we need one form for URI matching otherwise "file:/path" does not
match "file:///path"
Jena prefers the "file:///" form.
(c.f. "http:/" which is resolved.)
Normally, the parser emits a warning
Bad IRI: <file:/pdf/....> Code: 57/REQUIRED_COMPONENT_MISSING in
AUTHORITY: A component that is required by the scheme is missing.
or soon to be:
Unwise IRI: <file:/....> file: URLs are of the form file:///path/...
but, for some reason, the JSON-LD path does not print warnings.
Recorded as:
https://github.com/apache/jena/issues/2922
Andy
{
"@context": {
"library": "dokk:vocab:library:",
"license": "dokk:vocab:license:",
"library:journal": { "@type": "@id" },
"license:licensed_under": { "@type": "@id" ,
"blob": "dokk:vocab:blob:"
},
"@type": [
"library:Item",
"library:JournalArticle"
],
"@id": "dokk:theoryofcomputing_v020a001",
"library:author": [
"Ivan Hu",
"Dieter van Melkebeek",
"Andrew Morgan"
],
"library:journal": "dokk:theoryofcomputing",
"license:licensed_under": "dokk:license:CC-BY-3.0",
"library:title": "Polynomial Identity Testing via Evaluation of Rational
Functions",
"blob:at": {
"@id": "file:/pdf/theoryofcomputing.org/v020a001.pdf",
"blob:primary_source":
"https://theoryofcomputing.org/articles/v020a001/v020a001.pdf";,
"blob:retrieval_date": "2025-01-01"
}
}
