On 13.09.12 12:08, Alessandro Adamou wrote: Hi Alessandro,
> To begin with, http://purl.org/rss/1.0/ is not returning RDF for me. > Even if I specify it in the Accept header it still returns some HTML > with the embedded RDF schema. Try http://purl.org/rss/1.0/schema.rdf ah damn thanks for the hint, didn't even try that. Assumed that it would :-) > Then, the parameter for the URL is not "location" but "url". ok then I would say this is wrong in the documentation, you talk about location there (in REST API page on Ontonet): Parameters location: the physical URL of the ontology to be loaded. > Finally, you should use -F instead of -d, as the service expects > multipart form data. If the ontology is not very large you can use -d > (which passes it as form URLencoded), but then the payload must be the > actual ontology content, and you must specify the MIME type. ah ok my bad, didn't understand the difference between that yet. Thanks for the hint > So, either: > > curl -X POST -F "url=http://purl.org/rss/1.0/schema.rdf"; > http://data.gemeindezukunft.ch:9090/ontonet/ > > or you download the RDF file and then > > curl -H "Content-type: application/rdf+xml" -d @schema.rdf > http://data.gemeindezukunft.ch:9090/ontonet/ great that works, thanks! > I thought I had updated the REST live documentation to state this, > perhaps I forgot some spots though? at least in my latest trunk version there is absolutely no example on how to pass the arguments so the CURL examples are not complete IMHO. Would be great if you could update that so copy/past the example would result in a working POST. > Yes, you can request an ontology either standalone or managed (i.e. > within an ontology collector -- scope/space/session -- that is > referencing it). They will mostly be the same, however the managed > version can have some additional axioms and imports that could change > the interpretation of the ontology if you give it to, e.g. Protégé or a > reasoner. ah ok, I see the idea of grouping but didn't get yet why I would want to use that. > Even better, a good reasoner will infer that :Adrian and :Alessandro are > foaf:Person s ok great, I start to see the idea and will definitely play more with that. But I'm not really far in reasoning yet, didn't do much more than some simple rules and SPARQL CONSTRUCTS so far. > Of course you could import FOAF in your ontology as well, but the > problem is that not many people think about doing that, especially in > Linked Data. you mean make something a subproperty of a FOAF property for example if I get you correctly? That is indeed rarely done currently. > Also, having a managed ontology can help for maintaining multiple > versions simultaneously. ah yes that makes a lot of sense > There's more to that, but I hope this helps for a start. It does, thanks a lot. Also the same question to you: If I figure out stuff I prefer to document it somewhere so others don't have to do the same learnings again. Is there a place where I could try doing that @ stanbol? Otherwise I will simply start doing it for myself. > Right. You can POST either in /ontonet or in /ontonet/ontology/{ScopeID} > or in /ontonet/session/{SessionID} ok I think I get that. I posted to /ontonet/ontolgy before and it created an ID which I now see in /ontonet as well directly. > Note that some endpoints will change their URLs, most notably "ontonet" > will be replaced with something else for legal reasons. We haven't > decided what with, yet. right I remember this discussion on the list. > Oh and before you ask: ontology deletion is being implemented because > there are some dependency issues between ontologies to be handled. > Originally the feature was in the ontologymanater/store, but since it is > not part of canonical Stanbol, it looks like it will have to be > implemented in ontonet. ok is there a way to drop them right now manually? For testing this is a bit annoying :-D cu Adrian
