GiellaLT provides an infrastructure for rule-based language technology aimed at minority and indigenous languages, and streamlines building anything from keyboards to speech technology. Read more about Why. See also How to get started, and our Privacy document.
This is a document that tries to first describe the present situation as of december 2010, and then lay out what we want to have in the not too distant future.
A number of default targets are defined for all languages. The list is language depended, ie defined for every language we presently support. The language is specified using the variable $TARGET.
The following describes how it is done using the Xerox tools (the HFST compilation is still too underdeveloped to really be comparable).
compile all LexC source files (in $SRCS) to a lexical transducer using lexc. The resulting transducer should have paths of the following form:
gieseme#ruevtie+CmpN/SgN+CmpN/SgG+CmpN/PlG+N+Der1+Der/Dim+N+Sg+Ill+PxSg2:gieseme#ruevt%^aLATUS%»tje%>assedh
(corresponding to the surface word form giesemeruavtatjassedh)
2.
A number of default targets are defined for all languages. The list is mainly language independed, but possibly with language extension. The variable used to specify the language should be renamed to $GTLANG.
The following describes how it is done using the Xerox tools (the HFST compilation is still too underdeveloped to really be comparable).