Changeset 7935 for lang/perl/Acme-Shukugawa-Atom

Timestamp:

03/14/08 09:27:47 (9 months ago)

Author:

daisuke

Message:

lang/perl/Acme-Shukugawa-Atom; add tv stations, handle spaces

Location:

lang/perl/Acme-Shukugawa-Atom/trunk

Files:

• lib/Acme/Shukugawa/Atom.pm (modified) (3 diffs)
• t/01_basic.t (modified) (1 diff)

lang/perl/Acme-Shukugawa-Atom/trunk/lib/Acme/Shukugawa/Atom.pm

@@ -78 +78 @@
         '時間が' => '尺が',
         '用意する|準備する' => 'しこむ',
+        'フジテレビ' => 'CX',
+        'ニッポン放送' => 'LF',
+        'テレビ東京' => 'TX',
+        '文化放送' => 'QR',
+        '日本テレビ' => '日テレ',
     );
 }
@@ -131 +136 @@
     my $ret = '';
 
-    foreach my $text (split(/($RE_EXCEPTION)/, $$strref)) {
+    foreach my $text (split(/($RE_EXCEPTION|\s+)/, $$strref)) {
         if ($text =~ /$RE_EXCEPTION/) {
             $ret .= $text;
@@ -137 +142 @@
         }
 
+        if ($text !~ /\S/) {
+            $ret .= $text;
+            next;
+        }
+
         foreach (my $node = $mecab->parse($text); $node; $node = $node->next) {
             next unless $node->surface;
+
             my $surface = decode_utf8($node->surface);
             my $feature = decode_utf8($node->feature);

lang/perl/Acme-Shukugawa-Atom/trunk/t/01_basic.t

@@ -150 +150 @@
 ===
 --- input:    フジテレビ ニッポン放送 テレビ東京 文化放送 日本テレビ
---- expected: CX LF TX QR ニッテレ
+--- expected: CX LF TX QR 日テレ
 
 ===