# This is a regular expression made to remove the soft hyphen sign from the upper # level of the analysis. (It used to be called hard hyphens, but that is wrong - hard # hyphens are just regular hyphens, they are part of the word (as in "CD-spelar"), whereas # soft hyphens are potential hyphenation points, and what we are concerned with here. # Schematically, it looks like this: # 0 <- %^ ; Storuman+N+Gen nohardhyphen.fst # Stor^uman+N+Gen # .o. # lexc Stor^uman+N+Gen # Stor^umana # compose ==> sme.save # twol: %^:0 Stor^umana # Stor0umana # .o. .o. # caseconv.fst # .o. # downcase.fst # .o. # spellrelax.fst # Here comes the actual regex: 0 <- %^ ;