Hi Bill, First of all, please tell me which KIM version do you use (I guess it is the official version form the site - 17.12.15 - then the online docs and yor local ones should be the same). We constantly update the documentation and the examples, and to the last official version there indeed are errors in "in.space.owl" example and some tricky steps are not mentioned.
> Then I was getting other errors on start which went > away when I added the http://in.space#; namespace to the DefaultNS > node in the sesame.inmem.conf file. That last bit wasn't mentioned in the > documentation, should I have done that? Well, actually it is mentioned, Extending the Ontology => Configuration => a link to Sematic Repository Configuration. Importing new modules and adding defaultNSs is described there, so no need to double this. To answer your question - yes, you did right. > Anyway, the newly defined device, Robot, dosen't show up on > in the web interface under Entity -> Object -> Product -> Device -> Robot. > I have attached the relevant files. Can you take a look at them and > let me know what I am doing wrong? I don't have those sttachments as I received this mail forwarded. But I gues the reason you don't see the newly defined entity is a matter of missed instructions that are present in the latest doc. Here they are, assume this a last paragraph in "extending-kim-platform.html": ========================================================== Clear cached entity and ontology files for KIM and Tomcat For the changes to the ontology and/or the trusted entities to be taken into account the next time KIM and Tomcat are started it is necessary to remove some cached items. This will ensure that they will be clearly reloaded, recognized and shown correctly. a.. delete the %KIM_CONTEXT%/populated/cache/kim.trusted.entities.cache file (see more about the context folder) b.. delete the cached file %TOMCAT_HOME%/webapps/KIM/screen/script.inc to see the updated ontology in KIM WebUI (for more details see the HowTo... section) ========================================================== I thought of attaching a ZIP with the latest docs but I gave it up as there are too many changes that will not match your version and may confuse you. > KIMService k = GetService.from(); > eAPI = k.getEntityAPI(); > eAPI.createEntity("string") Finally, about the EntityAPI - in the version you have it is incomplete and inmature. Better don't try using it as the following versions introduce major changes to it, examples are also provided. Here is a link you can have the latest version from, just bear in mind it will solve your current problems but might raise new ones as it is still not stable - thus it is not officially released yet. http://ontodemo.sirma.bg/KIM/download/2.1.6.26/ Please, post your further questions, problems, recommendations to the KIM-discussion list (in the CC), so that more people are able to view and answer them. HTH, Ilian ----- Original Message ----- From: "Atanas Kiryakov" <[EMAIL PROTECTED]> To: "William Blair" <[EMAIL PROTECTED]> Cc: "Ilian Kitchukov" <[EMAIL PROTECTED]> Sent: 08 July 2007, Sunday 12:16 Subject: Re: KIM Platform Registration Dear Bill first of all sorry for the huge delay in answering this mail. In principle, the best way to handle such mails is by means of subscribing for the KIM mailing lists and posting the requests there: http://www.ontotext.com/kim/mailing-lists-info.html I am CC-ing Ilian Kitchukov, who can best answer your questions Regards Naso ---------------------------------------------------------- Atanas Kiryakov Head of Ontotext Lab, http://www.ontotext.com Sirma Group Corp, http://www.sirma.bg Phone: (+359 2) 8091 555; Fax: 9768 311 ---------------------------------------------------------- We really would like to take advantage of the "analyzing, annotating, indexing and searching" capabilities of KIM in our application. To that end we have been trying to extend the ontology by following the documentation, specifically the page "extending-kim-platform.html". I created the files in.space.owl and in.space.nt and edited the sesame.inmem.conf to point to the new files. I was getting errors initially then I edited the in.space.owl doc and removed the extra quote after Person: <owl:ObjectProperty rdf:about="#createdBy" rdfs:label="createdBy"> <rdfs:domain rdf:resource="#Robot"/> <rdfs:range rdf:resource="&ptop;Person""/> </owl:ObjectProperty> Then I was getting other errors on start which went away when I added the http://in.space#; namespace to the DefaultNS node in the sesame.inmem.conf file. That last bit wasn't mentioned in the documentation, should I have done that? Anyway, the newly defined device, Robot, dosen't show up on in the web interface under Entity -> Object -> Product -> Device -> Robot. I have attached the relevant files. Can you take a look at them and let me know what I am doing wrong? Ultimately we want to add the wordnet and DMOZ to the WKB in order to give us additional knowledge to start out with. Once we do this do you think we will run into performance issues in using the in-memory sail and might want to change it to a database based sail? And finally, I was looking at the api that is provided and I see that there is a method for adding entities: KIMService k = GetService.from(); eAPI = k.getEntityAPI(); eAPI.createEntity("string") But I have no idea what the format of this string should be. Is it an n-tripple or an rdfs doc? If you have some sample code that would be much appreciated. Thanks, Bill Blair i33 ------------------------------------------------------------------------------ From: Atanas Kiryakov [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 16, 2007 1:30 PM To: William Blair Subject: Re: KIM Platform Registration Dear Bill, I will address your comments inline below One of the unique uses we were looking for was to take advantage of the "Massive world Knowledge" part of the KIM platform. I think that this is part of the Proton module, is that right? PROTON is an upper-level ontology, which defines classes as Person and Organization The major volume of common world knowledge is encoded in the so called KIM World Knowledge Base (WKB). It countains definitions of thousands of concrete entities (or instances, if you prefer) - e.g. Asia, Bill Clinton, USD (the currency), March (the calendar month), etc. What I am looking to do is to have a user type in a word, like "birds" and it would suggest similar, related words like "animals" or "robins". Is this something that it could do? It looks like something that Sesame could do but, of course, you have to build up the right repository with all of the right data. KIM is about analysing, annotating, indexing, and searching text. Ituses ontologies, and more generally structured knowledge, to annotate the text. It also extracts new facts from the text and extends the ontology. Still, when it comes to playing with specific ontologies and knowledge bases, our most relevant product is OWLIM, http://www.ontotext.com/owlim - a semantic repository handling structured data and answering queries about it. It is available as storage and inference layer of Sesame, e.g. it can be used though its interfaces. To this point KIM can contribute only: - the WKB - its web UI Linking "birds" to "animals" requires a dictionary-type lexical knowledge. I.e. knowledge about general terms/words, e.g. "apple" and "bird". On the contrary, WKB models specific objects, e.g. conrete people and organisations. The most popular "lexical ontology" is Wordnet. http://wordnet.princeton.edu/. It is a really good, large, well-structured dictionary. You can play with the dictionary online athttp://wordnet.princeton.edu/perl/webwn. There is a RDF/OWL encoding of this dictionary at http://www.w3.org/TR/wordnet-rdf/. We have recently tested SwiftOWLIM to load this representation of Wordnet - it works perfectly. One thing you should consider is that there are many different type of relations or associations between the different words. E.g. hypernymy (appler->fruit), meronimy (body->head; sand->beatch), antonymy (good->bad). In Wordnet there are about 20-30 of those. You can use each one of them as an evidence of similarity, but those which best match your application need to be identified taking into consideration the domain, the context, the background of the users, etc. We are also trying to load up Sesame with the RDF dump from the DMOZ to accomplish the same goal but it might be a waste of time if KIM can do that as well because KIM has a lot of other features that we will want to take advantage of. Loading DMOZ in Sesame is not a bad thing to do - it is just like extending the WKB of KIM with additional knowledge. KIM is designed to be able to cope with extensions and customization of its pre-define ontologies and KBs. I would recommend you to use SwiftOWLIM, instead of just Sesame, as it is much more efficient and scalable and allows more inference. Should you have further questions, please, do not hesitate to contact me. Regards Naso ---------------------------------------------------------- Atanas Kiryakov Head of Ontotext Lab, http://www.ontotext.com Sirma Group Corp, http://www.sirma.bg Phone: (+359 2) 8091 555; Fax: 9768 311 ---------------------------------------------------------- __________ NOD32 2384 (20070708) Information __________ This message was checked by NOD32 antivirus system. http://www.eset.com _______________________________________________ NOTE: Please REPLY TO ALL to ensure that your reply reaches all members of this mailing list. KIM-discussion mailing list [email protected] http://ontotext.com/mailman/listinfo/kim-discussion_ontotext.com
