french tokenizer small change

TALP-UPC · Mar 31, 2016 · f2ba4d0 · f2ba4d0
1 parent 79581b7
commit f2ba4d0
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/data/fr/tokenizer.dat b/data/fr/tokenizer.dat
@@ -32,7 +32,7 @@ MAILS 	        0  {ALPHANUM}+([\._]{ALPHANUM}+)*@{ALPHANUM}+([\._]{ALPHANUM}+)*
 LON_FR		1  (l'on){NOALPHANUM} CI
 APOSTR_FR       1  ((qu|[cdlmtsjn])')({ALPHA}) CI
 URLS1 	        0  ((mailto:|(news|http|https|ftp|ftps)://)[\w\.\-/]+|^(www(\.[\w\-/]+)+))
-URLS2           1  ([\w\.\-/]+\.(com|org|net))[\s]
+URLS2           1  ([\w\.\-/]+\.(com|org|net|fr))[\s]
 KEEP_COMPOUNDS  0  {ALPHA}+(['_\-\+]{ALPHA}+)+
 *ABREVIATIONS1  0  (({ALPHA}+\.)+)(?!\.\.)
 WORD            0  {ALPHANUM}+[\+]*