summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorChris Dyer <cdyer@Chriss-MacBook-Air.local>2013-04-19 17:06:35 -0400
committerChris Dyer <cdyer@Chriss-MacBook-Air.local>2013-04-19 17:06:35 -0400
commitf729d92ab891a714ced95bd0d4cd8e5d8470a52d (patch)
tree834deff8895c8ba8a4f17f4cee7f32aa206c42fc
parent5555d37ada60ea64ce7d8f25e40827a79ef003d8 (diff)
hindi
-rwxr-xr-xcorpus/support/tokenizer.pl8
1 files changed, 4 insertions, 4 deletions
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;