Hi Reto, all If I understand correcly, this means that the resulting metadata obtained from the annotation service will not be stored in a triple store, but directly on the filesystem (in which format?). The "zz-way" to access these metadata is through this special TcProvider. Right? I think I can agree with this. Does anybody see any problem with this approach?
Hasan On Wed, Apr 6, 2011 at 10:45 AM, Reto Bachmann-Gmuer < [email protected]> wrote: > > > On Mon, Apr 4, 2011 at 11:49 AM, Hasan Hasan <[email protected]> wrote: > >> Hi Reto, >> >> On Mon, Apr 4, 2011 at 9:20 AM, Reto Bachmann-Gmuer < >> [email protected]> wrote: >> >>> Hi Hasan >>> >>> >>>> @Reto, do we want to have a dedicated TcProvider or rather define an >>>> ontology for >>>> the mapping of CAS into properties and classes ? >>>> I think we should discuss both alternatives or why to you prefer one >>>> over the other? >>>> >>> I don't see the conflict. How should A TcProvider map the entities >>> without an ontology? An what's the benefit of an ontology without a tool >>> that does the mapping? >>> >> >> Probably, I misunderstand you. I thought you would like to have a special >> TcProvider that does this mapping instead of using any TcProvider we have. >> > Yes, I'd rather have a special TcProvider gibing a life access to the CAS > data rather than copying this data to the MGraph provided by a storing > provider. > > Cheers, > Reto > >> >> cheers >> hasan >> >> >>> >>> Cheers, >>> reto >>> >>> >>> >>>> >>>> Kind regards >>>> Hasan >>>> >>>> >>>> On Sat, Apr 2, 2011 at 10:56 PM, Reto Bachmann-Gmuer < >>>> [email protected]> wrote: >>>> >>>>> Hi Tommaso >>>>> >>>>> Rdfizing the CAS and providing a TcProvider that provides a graph view >>>>> to >>>>> the UIMA CAS data sounds like a grat plan. I imagine there would be >>>>> interesting stuff that can be done implementing cas consumers on >>>>> clerezza. >>>>> Do you have ideas for usecases and demo apps? >>>>> >>>>> Cheers, >>>>> Reto >>>>> >>>>> On Fri, Apr 1, 2011 at 3:09 PM, Tommaso Teofili >>>>> <[email protected]>wrote: >>>>> >>>>> > Hi all, >>>>> > recently I've been discussing with Stanbol mates [1] about possible >>>>> usage >>>>> > of >>>>> > text mining algorithms in Stanbol on top of UIMA, so, since they're >>>>> already >>>>> > using Clerezza we agreed that it'd be good to implement a component >>>>> which >>>>> > converts objects from the UIMA CAS model to a Clerezza Graph. >>>>> > Since I am familiar with both I'd volunteer to implement such a thing >>>>> > inside >>>>> > the uima module. >>>>> > What do you think? >>>>> > Regards, >>>>> > Tommaso >>>>> > >>>>> > [1] : >>>>> > >>>>> > >>>>> http://markmail.org/thread/2sstx2girbiydath#query:+page:1+mid:wbky7miem5ggdsi2+state:results >>>>> > >>>>> >>>> >>>> >>> >> >
