summaryrefslogtreecommitdiff
path: root/corpus
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2013-05-02 09:09:59 +0200
committerPatrick Simianer <p@simianer.de>2013-05-02 09:09:59 +0200
commit0ce66778da6079506896739e9d97dc7dff83cd72 (patch)
treef435457bb23dab0c566c9896f9d38cece9d15885 /corpus
parentb6754386f1109b960b05cdf2eabbc97bdd38e8df (diff)
parentb7ea2615bc9bb69031ff714ddce1539c9f1bda2d (diff)
Merge remote-tracking branch 'upstream/master'
Diffstat (limited to 'corpus')
-rwxr-xr-xcorpus/lowercase.pl2
-rw-r--r--corpus/support/token_list152
-rwxr-xr-xcorpus/support/tokenizer.pl8
3 files changed, 157 insertions, 5 deletions
diff --git a/corpus/lowercase.pl b/corpus/lowercase.pl
index 688e493b..9fd91dac 100755
--- a/corpus/lowercase.pl
+++ b/corpus/lowercase.pl
@@ -2,7 +2,7 @@
use strict;
binmode(STDIN,":utf8");
binmode(STDOUT,":utf8");
-while(<>) {
+while(<STDIN>) {
$_ = lc $_;
print;
}
diff --git a/corpus/support/token_list b/corpus/support/token_list
index 366cd7ff..43dd80d9 100644
--- a/corpus/support/token_list
+++ b/corpus/support/token_list
@@ -294,3 +294,155 @@ Z.
т.н.
т.ч.
н.э.
+# Swahili
+A.D.
+Afr.
+A.G.
+agh.
+A.H.
+A.M.
+a.s.
+B.A.
+B.C.
+Bi.
+B.J.
+B.K.
+B.O.M.
+Brig.
+Bro.
+bt.
+bw.
+Bw.
+Cap.
+C.C.
+cCM.
+C.I.A.
+cit.
+C.M.S.
+Co.
+Corp.
+C.S.Sp.
+C.W.
+D.C.
+Dk.
+Dkt.
+Dk.B.
+Dr.
+E.C.
+e.g.
+E.M.
+E.n.
+etc.
+Feb.
+F.F.U.
+F.M.
+Fr.
+F.W.
+I.C.O.
+i.e.
+I.L.C.
+Inc.
+Jan.
+J.F.
+Jr.
+J.S.
+J.V.W.A.
+K.A.R.
+K.A.U.
+K.C.M.C.
+K.k.
+K.K.
+k.m.
+km.
+K.m.
+K.N.C.U.
+K.O.
+K.S.
+Ksh.
+kt.
+kumb.
+k.v.
+kv.
+L.G.
+ltd.
+Ltd.
+M.A.
+M.D.
+mf.
+Mh.
+Mhe.
+mil.
+m.m.
+M.m.
+Mm.
+M.M.
+Mr.
+Mrs.
+M.S.
+Mt.
+Mw.
+M.W.
+Mwl.
+na.
+Na.
+N.F.
+N.J.
+n.k.
+nk.
+n.k.w.
+N.N.
+Nov.
+O.C.D.
+op.
+P.C.
+Phd.
+Ph.D.
+P.J.
+P.o.
+P.O.
+P.O.P.
+P.P.F.
+Prof.
+P.s.
+P.S.
+Q.C.
+Rd.
+s.a.w.
+S.A.W.
+S.D.
+Sept.
+sh.
+Sh.
+SH.
+shs.
+Shs.
+S.J.
+S.L.
+S.L.P.
+S.s.
+S.S.
+St.
+s.w.
+s.w.T.
+taz.
+Taz.
+T.C.
+T.E.C.
+T.L.P.
+T.O.H.S.
+Tsh.
+T.V.
+tz.
+uk.
+Uk.
+U.M.C.A.
+U.N.
+U.S.
+Ush.
+U.W.T.
+Viii.
+Vol.
+V.T.C.
+W.H.
+yamb.
+Y.M.C.A.
diff --git a/corpus/support/tokenizer.pl b/corpus/support/tokenizer.pl
index 0350a894..acc537fb 100755
--- a/corpus/support/tokenizer.pl
+++ b/corpus/support/tokenizer.pl
@@ -226,7 +226,7 @@ sub proc_token {
}
## step 1: check the most common case
- if($token =~ /^[a-z0-9\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}]+$/i){
+ if($token =~ /^[a-z0-9\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}\p{Devanagari}]+$/i){
### most common cases
return $token;
}
@@ -246,7 +246,7 @@ sub proc_token {
## number
return $token;
}
- if($token =~ /^(@|#)[A-Za-z0-9_\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}]+.*$/){
+ if($token =~ /^(@|#)[A-Za-z0-9_\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}\p{Devanagari}]+.*$/){
## twitter hashtag or address
return proc_rightpunc($token);
}
@@ -277,7 +277,7 @@ sub proc_token {
}
#my $t1 = '[\x{0600}-\x{06ff}a-z\d\_\.\-]';
- my $t1 = '[a-z\d\_\-\.\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}]';
+ my $t1 = '[a-z\d\_\-\.\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}\p{Devanagari}]';
if($token =~ /^\/(($t1)+\/)+($t1)+\/?$/i){
### /nls/p/....
return $token;
@@ -361,7 +361,7 @@ sub deep_proc_token {
}
##### step 0: if it mades up of all puncts, remove one punct at a time.
- if($line !~ /[\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}a-zA-Z\d]/){
+ if($line !~ /[\p{Cyrillic}\p{Greek}\p{Hebrew}\p{Han}\p{Arabic}\p{Devanagari}a-zA-Z\d]/){
if($line =~ /^(\!+|\@+|\++|\=+|\*+|\<+|\>+|\|+|\?+|\.+|\-+|\_+|\&+)$/){
## ++ @@@@ !!! ....
return $line;