Open mcodescu opened 8 years ago
sounds good! A smooth extension of what is already there.
The leading "dg" was supposed to indicate that you would get a development graph, so maybe "theory" should be used in the first two cases (that may require node=iri
to be in the query part or would only output the theory of the first or a solitary node).
Agreed. A translation of the (full) theory of a node in a DG along a composition of comorphisms would be written GET /theory/coded-iri/command-list?options&node=iri&translation=iri1;...;irin
@clange, is the semicolon legal for separating IRIs?
According to RFC 3986 the semicolon is legal. However, for this same reason, it might occur in any of the iri1
to irin
. If it occurs there it will have to be escaped as %3B
, which shouldn't be a problem.
FYI, IIRC Florian Rabe makes a lot of use of ;
in his MMT API.
&
and ;
in a query string of an IRI have the same semantics/are interchangeable. We should not redefine the semantics of ;
in the Hets API. Maybe a better solution would be to use another HTTP-verb than GET
and supply the options in the request body.
... or simply to use a ,
like in the command line:
❯ hets --help
Usage: hets [OPTION ...] [FILE ...] [+RTS -?]
[...]
-t TRANS --translation=TRANS translation option
is a colon-separated list
of one or more from: SIMPLE-ID
Good, a comma is a good idea.
GET /dg/coded-iri/command-list?options&node=iri
GET /dg/coded-iri/command-list?options&node=iri&translation=iri
GET /dg/coded-iri/command-list?options&translation=iri