forked from GitHub/gf-rgl
8fb8ddd808
- NP: added field isLight in order to push negation behind light nps;
this had been done in gf-3.9 using field isPron, but isPron is now
used to put accusative pronoun before dative pronoun. Removed field
adv: adverbial extensions cannot be extracted (todo: also for CN).
Reduced isLight*isPron to w:Weight with 3 values: WPron, WLight, WHeavy.
- added param Control and field ctrl:Control to classify V2V-verbs into
subject- and object-contol verbs, use ctrl to make reflexives agree
with subject resp. object in VPSlash, and refine ComplSlash.
- Verb: new versions of ComplVV, SlashV2V and SlashVV to give better
(nested) infinitives (extracting infzu and correcting object order).
a) nested SlashVV doesn't work properly;
b) SlashV2VNP may have to be commented out to prevent a stack overflow
when compiling.
Intended change of SlashV2VNP in tests/german/TestLangGer could not
be tested due to size problems with the compiler.
- VP: changed field a1 : Polarity => Str to a1:Str to collect the adverbs
coming before negation, using (negation : Polarity => Str) in mkClause.
Use objCtrl:Bool instead of missingAdv to let reflexives agree with object.
- ResGer: insertObjNP reorganized, infzuVP added
- DictVerbsGer: some corrections (helft -> hilft, *sprecht -> *spricht)
- Some potential passive rules in tests/german/TestLangGer|Eng
- ExtraGer needs to be cleaned up with repect to the modified mkClause.
24 lines
1.4 KiB
Plaintext
24 lines
1.4 KiB
Plaintext
-- Use gf --run < passive.gfs or gf> eh passive.gfs
|
|
--? echo "loading TestLangGer.gf and TestLangEng.gf ..."
|
|
--i TestLangGer.gf TestLangEng.gf
|
|
|
|
? echo "extracting positive, negative and dubious examples from passive.txt ..."
|
|
? grep accept passive.txt | sed s/\ --\ [\*a-zA-Z\(\)\ \.\:\,\;\<\>\\_0-4\\+\\?\\-]*// | cpif passive.pos.txt
|
|
? wc -l passive.pos.txt
|
|
? grep reject passive.txt | sed s/\ --\ [\*a-zA-Z\(\)\ \.\:\,\;\<\>\\_0-4\\+\\?\\-]*// | cpif passive.neg.txt
|
|
? wc -l passive.neg.txt
|
|
? grep dubious passive.txt | sed s/\ --\ [\*a-zA-Z\(\)\ \.\:\,\;\<\>\\_0-4\\+\\?\\-]*// | cpif passive.dub.txt
|
|
? wc -l passive.dub.txt
|
|
|
|
? echo "parsing positive examples ...; storing trees in passive.pos.new ..."
|
|
rf -lines -file="passive.pos.txt" | p -lang=Ger | l -treebank -lang="Ger,Eng" | wf -file="passive.pos.new"
|
|
? echo "diff passive.pos.out passive.pos.new:" ; diff passive.pos.out passive.pos.new
|
|
|
|
? echo "parsing negative examples ...; storing trees in passive.neg.new ..."
|
|
rf -lines -file="passive.neg.txt" | p -lang=Ger | l -treebank -lang="Ger,Eng" | wf -file="passive.neg.new"
|
|
? echo "diff passive.neg.out passive.neg.new:" ; diff passive.neg.out passive.neg.new
|
|
|
|
? echo "parsing dubious examples ...; storing trees in passive.dub.new ..."
|
|
rf -lines -file="passive.dub.txt" | p -lang=Ger | l -treebank -lang="Ger,Eng" | wf -file="passive.dub.new"
|
|
? echo "diff passive.dub.out passive.dub.new:" ; diff passive.dub.out passive.dub.new
|