Commit Graph

3779 Commits

Author SHA1 Message Date
bjorn
832f25fc2a Don't use string sharing in LexGF.
Profiling showed that when loading a large .gfo file, shareString was responsible for
15-18% of the CPU time, and a lot of the allocation. Since we already use ByteStrings for 
reading the source files, shareString mostly has the effect of creating lots 
of small ByteStrings instead of one large one. Since the plain size of the .gfo is seldom
a problem (unlike when it was read as a String), it is ok to keep the whole file 
as one ByteString in RAM, and have all tokens point into that.
Profiling after the change showed 15-20% reduction in CPU time and in total allocation.
2008-11-26 16:19:54 +00:00
bjorn
499d6ade49 My profiling showed that the BinTree operations were responsible for about 60% of the CPU time when reading a large .gfo file. Replacing BinTree by Data.Map reduced this to about 6%, which meant about 50% reduction in total CPU time. 2008-11-26 15:44:22 +00:00
bjorn
22269d277e Get 'make prof' working with GHC 6.8 2008-11-26 14:56:46 +00:00
bjorn
c63c3d9f58 Use OALD in ParseEng. 2008-11-26 14:56:34 +00:00
bjorn
54e994ef15 OALD: split lexicon into two: one for open classes and one for closed classes. 2008-11-26 14:35:35 +00:00
krasimir
4616041bdd add Russian in next-lib 2008-11-26 14:43:05 +00:00
bjorn
74c3956144 Added generated OALD files (Oald.gf and OaldEng.gf) to repo. 2008-11-26 14:17:55 +00:00
bjorn
6d13809091 Added original OALD files to repo. 2008-11-26 14:17:35 +00:00
bjorn
859cd04ae7 OALD: More info in generated header. 2008-11-26 14:17:16 +00:00
bjorn
dbbf7b5368 OALD: clean up asc2gf script. 2008-11-26 14:00:10 +00:00
bjorn
52536e8776 OALD: remove * and add comment for guessed irregular noun plural forms. 2008-11-26 13:38:34 +00:00
bjorn
4ab34d11bf OALD: Use verbs from IrregEng when possible. 2008-11-26 13:33:53 +00:00
bjorn
97a5fbca79 Added ParadigmsEng.mkV : Str -> V -> V, which adds a prefix to an existing verb. 2008-11-26 13:33:22 +00:00
bjorn
93222bb0ed First working version of OALD dictionary conversion. Still missing: irregular verbs, irregular nouns, better handling of singular- and plural-only nouns. 2008-11-25 20:09:50 +00:00
bjorn
734ac4cfc2 Added ParadigmsEng.mkConj. 2008-11-25 19:53:22 +00:00
bjorn
e80abfffbd First version of OALD alsmost working. 2008-11-25 14:50:16 +00:00
krasimir
c4f5c03700 add preposition "na" when the case is dative (QuestionBul) 2008-11-24 13:27:33 +00:00
krasimir
177e8573ec replace spec with qform 2008-11-24 12:55:17 +00:00
krasimir
5d8868e38c replace latin 'a' with cyrillic 'a' 2008-11-24 12:53:57 +00:00
aarne
a993ee2151 links in demo.txt 2008-11-21 17:12:34 +00:00
aarne
94320c3889 links to demos 2008-11-21 17:10:02 +00:00
bjorn
c03f2768af Started working on oald2gf script. 2008-11-21 16:50:07 +00:00
bjorn
0a73c96ee3 Added CompCN to ParseEng. Perhaps this should move to NounEng. 2008-11-21 12:33:56 +00:00
bringert
fee1483bd8 lighttpd conf tweak. 2008-11-20 16:49:12 +00:00
bjorn
de09e64892 pgf.fcgi: don't install any signal handlers. Since we didn't look at the variable they set, we just created zombies. 2008-11-20 15:00:53 +00:00
bjorn
1203d31d77 gwt: rearragne html files and paths to get nicer URLs. 2008-11-20 14:06:06 +00:00
bjorn
8f1d5c79ef Added CAdvNP and CAdvSSlash to ParseEng. Maybe they should go in Noun.gf instead? 2008-11-20 10:32:15 +00:00
bjorn
dd58077add Fix permissions of pgf server error log. 2008-11-20 09:52:18 +00:00
bjorn
7181e8540a gf-server lighttpd.conf: put fcgi socket in basedir, with PID in name. This helps avoid clashed when running muyltiple gf servers on the same machine. 2008-11-20 08:56:17 +00:00
bringert
d6acf9ac89 Include MachDeps.h in Data/Binary.hs, since it uses WORD_SIZE_IN_BITS. Without this, the Binary Word instance would be wrong on 64-bit machines. 2008-11-19 20:30:57 +00:00
bjorn
b5a6e15757 Updated CompatibilityEng to compile with case-inflected Num and Ord. 2008-11-19 14:28:50 +00:00
bjorn
842d7e434b Fix good_A in BigLexEng. 2008-11-19 13:57:42 +00:00
bringert
de52b219f3 Merge SyntaxEng.mkOrd and ParadigmsEng.mkOrd in TryEng. 2008-11-19 14:16:28 +00:00
bjorn
972d300ba0 Updated ParseEng for case-inflected Num. 2008-11-19 13:15:36 +00:00
bjorn
eaed41caaa Added case inflection to English numerals and adjectives, in order to get genitive forms of stand-alone determiners that use Num or Ord right. 2008-11-19 13:11:45 +00:00
aarne
cad49a225d restored "?" as system pipe command 2008-11-18 15:31:30 +00:00
bjorn
081b5707c3 Added overloaded mkQuant to ParadigmsEng. 2008-11-18 15:30:53 +00:00
bjorn
eba86d2f24 The stand-alone form of no_Quant should be "none", not "no". 2008-11-18 15:20:40 +00:00
bjorn
da4b858316 Remove "at least" and "at most" from parsing grammar, since they are now in structural. 2008-11-18 15:20:14 +00:00
bjorn
6cd27ad4f3 Remove no_Det, nobody_NP and nothing_NP from the parsing grammar, sicne they are now in Structural. 2008-11-18 12:38:03 +00:00
aarne
65aefea2d6 fixed a bug with type checking function types 2008-11-18 10:58:50 +00:00
peb
fae330f733 Fix in Prolog output of the atom "'" 2008-11-17 13:33:53 +00:00
aarne
ce50264e91 updated demo page and summer school 2008-11-18 10:00:43 +00:00
aarne
31c010d61e new version of summerschool announcement 2008-11-17 17:58:11 +00:00
aarne
d67c87b7d8 Attempto complete up to spec 2008-11-15 20:08:09 +00:00
aarne
b13c62f71b PP lists in Attempto via iteration 2008-11-15 17:20:38 +00:00
aarne
9cf2fd879a new CAdv type in Dan and Nor 2008-11-15 11:24:22 +00:00
aarne
fc60a33c2a comparative adverbs generalized to cover more of Attempto 2008-11-15 11:19:10 +00:00
aarne
6b21e170e8 attempto almost complete to spec now 2008-11-14 17:39:05 +00:00
aarne
65c6f8548a rules for constructing structural words (separate from Paradigms - is this desirable?) 2008-11-14 17:32:34 +00:00