Hi Thilo, I and some friend are try to develop an extension of UIMA that allow users to obtain automatically (in a way similar to JCas Gen) an analysis engine based on regular expression over feature structure (UIMA type).
I try to explain with an example : In my type system descriptor I define three type of annotation : 1) Car with features color:string and model:string 2) Bus with features color:string and seat_number:integer 3) Traffic with some other feature I would obtain an analysis engine that match with the left hand side regexp add a new annotation as described in the right hand side (<Car color=blue> | <Bus seat_num=30> )* --> Traffic I hope this example explain better what I'm trying to implement(something similar to JAPE for GATE ). Thanks 2011/2/18 Thilo Götz <[email protected]> > On 2/18/2011 16:53, Alessandro Rossi wrote: > > Hi to all, > > > > I would like to have some information > > > > 1) There exists any methods to define regex on UIMA type (not for simple > > text) ? > > Please be more explicit, not clear what you mean. > > > > > 2) There exists any way to do unification and subsumption on UIMA type? > > Short answer: no. Long answer: subsumption is easy, you can implement > that in a couple of hours. Unification is more difficult. If you > want to do it yourself, we can discuss where the major issues are. > > --Thilo > > > > > Thanks. > > > > Bye > > >
