omorfi icon indicating copy to clipboard operation
omorfi copied to clipboard

omorfi-hyphenate.sh references non-existent hfst

Open eschling opened this issue 10 years ago • 1 comments

omorfi-hyphenate.sh references $prefix/share/hfst/fi/hyphenate.dict.hfst, which doesn't exist. Should it be $prefix/share/hfst/fi_FI/hyphenation.rule.hfst?

eschling avatar Mar 24 '15 17:03 eschling

good point, hyphenation hasn't been maintained much and it seems to be in serious need of reworking; the current ruleset combined with dictionary will end up wasting hundreds of megs and isn't usable as such. The rules at fi_FI used separately like that would work, but I guess it's from an older installation? I've changed current version to rely on rules only but it really is something one could do with sed faster... or with tex or so, the tex patterns I feel are little less accurate on some cases but good enough. I'll leave the issue open though until the size issue for hyphenating dictionary is resolved. It can still be explicitly compiled by make generated/omorfi.hyphenate-dict.hfst and used with hfst-lookup.

flammie avatar Mar 25 '15 03:03 flammie