On this page
Special tags and flags for FST
This is a list of flags that are common for all languages in giellalt and
have special rules in automake scripts in giella-core. All tags and flags are
also documented in language specific documentation in
https://giellalt.github.io. Most of the handling described here is using
automata in fst-filters/ directory and called from am-shared/.
Use tags
+Use/NG- no MT / TTS / oahpa- filtered out from some MT automata. some TTS automata and oahpa automata but included in all generators and analysers otherwise?
+Use/NA- no analysis- filtered out from all analysis automata,
src/fst/analyser-*.hfst??
- filtered out from all analysis automata,
+Use/-Spell- no spell checking- filtered out from all spell-checker automata,
tools/spellers/*.zhfst
- filtered out from all spell-checker automata,
+Use/MT- only MT- filtered out of other than machine translation automata, only in
tools/mt
- filtered out of other than machine translation automata, only in
+Use/TTS- only TTS- filtered out of non-TTS automata only in
src/fst/*-tts-*.hfst??andtools/tts/*
- filtered out of non-TTS automata only in
+Use/SpellNoSugg- no speller suggestions- filtered out from spell-checker suggestions (programmatically), but silently accepted if found
+Use/-GC- no grammar checker,- filtered out of grammar checker automata, e.g. in
src/fst/*-gramcheck-*.hfst??andtools/grammarchecker/*.pmhfst
- filtered out of grammar checker automata, e.g. in
+Use/PMatch- only tokeniser- used only by tokenisers in
tools/tokenisers/*(and other tools that use tokenisers)
- used only by tokenisers in
Err tags
+Err/*- error forms- filtered from normative automata, e.g.
src/fst/*-gt-norm*.hfst??andtools/spellcheckers/*.zhfst - Err tags are optional in descriptive generator
- Err tags may be used by grammatical error correction logics
- filtered from normative automata, e.g.
Dialect tags
Dialects need to be set up in configure.ac
Dial/*- dialect tags- may be filtered into dialect specific automata, e.g.
src/fst/*-dial_XXX.hfst??where XXX is an ad hoc dialect code, alsotools/spellers/*-x-XXX.zhfst, etc.
- may be filtered into dialect specific automata, e.g.
Area tags
Areas need to be set up in configure.ac
Area/*- area tags- may be filtered into area-specific automata, e.g.
src/fst/*_XX.hfst??where XX is the standard country code, alsotools/spellers/*-XX.zhfstetc.
- may be filtered into area-specific automata, e.g.
Orth tags
Orthographies need to be setup in configure.ac
Orth/*- orthography tagsAltOrth/*- alternative orthography tags