On 9/25/13 10:49 PM, Ray Fergerson wrote:
We are very happy to finally announce, after a year of planning and development work, the release of the reengineering of BioPortal. Because this is a huge change we care labeling this release BioPortal 4.0.The primary distinguishing feature of this release is the replacement of several backend storage and caching mechanisms with a single triple-store based backend built with on the open source 4store platform. In addition we are simultaneously releasing a new version of the BioPortal REST API. This API is also built on top of the triple store. It provides a much more uniform and REST-full interface to the underlying BioPortal data than our previous REST API.The old API will continue to function through the end of 2013. This API will only be able to access "old content" (that is, ontologies submitted or updated before today). We encourage API users to transition as quickly as possible to the new API. The new API is the only way to access both new content (submitted after today) and old content. The reports from beta users of the new API have been uniformly positive. Draft documentation for the new API is available athttp://data.bioontology.org/documentationBioPortal users will note a few changes from the previous version. Some of these changes are temporary (we didn't quite get everything done...) while others are permanent. There are also a number of minor and cosmetic changes. The list of the more important changes is available below.With a release of this size, it is inevitable that some bugs have crept in. Please report any that you run across to the supp...@bioontology.org <mailto:supp...@bioontology.org> mailing list. Have confidence that we will be working diligently to prioritize and fix the problems that arise.The NCBO Team -------------------------- Temporary Changes-All handling of obsolete classes in UI and API is not supported yet. Obsolete classes are currently indistinguishable from normal classes.-"Slices" and ontology sets (user defined lists of ontologies of interest) are not supported yet-The tree browse UI does not yet distinguish between is-a and part-of relations for OBO ontologies. Previously (and probably in the future) icons next to the class name indicated the relationship type.-We do not yet have a VM version for this release. We expect to produce a new VM version by the end of the year. This VM will include the triple-store based storage mechanism. In addition VM users will be able to query the VM triple-store directly with SPARQL.-Parsing of new ontologies and new versions remains disabled for the next few days. We will re-enable parsing by early next week.Permanent Additions-New ontologies and new versions of ontologies will be typically available for text annotation the next day after submission. Previously the process took several weeks to several months.-We now have the latest versions UMLS ontologies. Previously we updated UMLS ontologies on request only, which left us with some very old versions. Going forward we will try to update our set of UMLS ontologies within a month of the UMLS release.Permanent Subtractions-Programmatic and browse access to classes in non-current versions of ontologies is removed. We still retain metadata about older versions of ontologies and we still allow you to download the source files for old versions.-We no longer support Protégé Frames format ontologies. Owners of these ontologies will need to convert them to OWL or OBO to load them.-We now use the owlAPI (http://owlapi.sourceforge.net/) to load OWL and OBO ontologies. This API supports OWL version 2.0. For almost all features, OWL 2.0 is a superset of OWL 1.0. There are some edge cases though where ontologies that marginally parsed with an OWL 1.0 parser will not parse at all with the OWL 2.0 parser. We leave it to OWL ontology owners to provide valid OWL 2.0 ontologies.
Am I correct in assuming that this still isn't a Linked Data offering? I say that because of what I get back from:
curl -iLH "Accept: text/turtle" http://purl.bioontology.org/ontology/SNOMEDCT/154501005
That returns HTML. Our Vapour variant for Linked Data verification reports:http://linkeddata.uriburner.com:8000/vapour?uri=http%3A%2F%2Fpurl.bioontology.org%2Fontology%2FSNOMEDCT%2F154501005&validateRDF=1&defaultResponse=dontmind&userAgent=vapour.sourceforge.net .
Re. SPARQL access, what's the SPARQL endpoint URL? I couldn't find that in your documentation (of course, I might have just missed it).
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca handle: @kidehen Google+ Profile: https://plus.google.com/112399767740508618350/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
smime.p7s
Description: S/MIME Cryptographic Signature
