X-Git-Url: http://dolda2000.com/gitweb/?p=utils.git;a=blobdiff_plain;f=ANN.pm;h=6dd02e94fc951f0b1773cf7f5131f6ecf3359386;hp=f41147a8c988549abc09ea34ee759b9d1f757726;hb=8b09dc89e6081cfc76fee513be8dac9e39b6305f;hpb=edcecb65d18376f5401aefbcbc2c563cf10e74f0

diff --git a/ANN.pm b/ANN.pm
index f41147a..6dd02e9 100644
--- a/ANN.pm
+++ b/ANN.pm
@@ -35,12 +35,12 @@ sub _get
     $res = $ua->request(HTTP::Request->new("GET", "$uri"));
     
     if(open CACHE, ">:utf8", $cname) {
-	print CACHE $res->content;
+	print CACHE $res->decoded_content;
 	close CACHE;
     }
     
     return undef unless $res->is_success;
-    return $res->content;
+    return $res->decoded_content;
 }
 
 sub getlist
@@ -49,7 +49,7 @@ sub getlist
     ($name) = @_;
     
     $name = ($name =~ /^(the\s+)?(.*)$/i)[1];
-    $il = uc(($name =~ /^(.)/)[0]);
+    $il = uc(($name =~ /^\W*(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
     if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
 	return undef;
@@ -58,9 +58,9 @@ sub getlist
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    while($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*>(\<small\>.*\<\/small\>)?([^<]+)<\//ig) {
-	if((substr "" . lc $3 , 0, length $name) eq lc $name) {
-	    push @ret, $3;
+    while($html =~ /<A\s+CLASS="HOVERLINE"\s+HREF="([^"]+)"[^>]*>(<FONT[^>]*>)?(<small>[^<]*<\/small>)?\s*([^<]+)<\//ig) {
+	if((substr "" . lc $4 , 0, length $name) eq lc $name) {
+	    push @ret, $4;
 	}
     }
     # push @ret, $1 while $html =~ /<A\s.*CLASS=HOVERLINE\s.*>.*<FONT.*>([^<>]*$name[^<>]*)<\/FONT/ig;
@@ -74,7 +74,7 @@ sub getid
     ($name) = @_;
     
     $name = ($name =~ /^(the\s+)?(.*)$/i)[1];
-    $il = uc(($name =~ /^(.)/)[0]);
+    $il = uc(($name =~ /^\W*(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
     if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
 	return undef;
@@ -83,8 +83,8 @@ sub getid
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    while($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*>(\<small\>.*\<\/small\>)?([^<]+)<\//ig) {
-	if((substr "" . lc $3 , 0, length $name) eq lc $name) {
+    while($html =~ /<A\s+CLASS="HOVERLINE"\s+HREF="([^"]+)"[^>]*>(<FONT[^>]*>)?(<small>[^<]*<\/small>)?\s*([^<]+)<\//ig) {
+	if((substr "" . lc $4 , 0, length $name) eq lc $name) {
 	    return ($1 =~ /id=(\d+)$/)[0];
 	}
     }
@@ -107,10 +107,10 @@ sub getthemes
     
     if($html =~ /$kind theme:<\/strong>\s*\n/igc) {
 	my(@parts, $ct, $buf);
-	while($html =~ /\G\s*\<div\>(([^<>]|\<i\>|<\/i>)+)<\/div>/igc) {
+	while($html =~ /\G\s*\<div class=\"tab\"\>(([^<>]|\<i\>|<\/i>)+)(<span[^<>]*>[^<>]*<img[^<>]*>[^<>]*<\/span>)?<\/div>/igc) {
 	    $buf = $1;
-	    #                     0  1            2           3     4     5        6          7                      8              9         10    11
-	    if(@parts = ($buf =~ /(\#(\d+):)?\s*\"([^\"\(]+\S)(\s*\((\<i\>(.*)<\/i>( - \s*)?)?([^<>]+)?\))?\"\s+by\s+([^\(]*[^\(\s])(\s*\(eps (\d+)-(\d+)?\))?/i)) {
+	    #                     0  1            2           3     4     5        6          7                      8              9          10   1112
+	    if(@parts = ($buf =~ /(\#(\d+):)?\s*\"([^\"\(]+\S)(\s*\((\<i\>(.*)<\/i>( - \s*)?)?([^<>]+)?\))?\"\s+by\s+([^\(]*[^\(\s])(\s*\(eps? (\d+)(-(\d+))?\))?/i)) {
 		$ct = {};
 		$ct->{"num"} = $parts[1] if defined $parts[1];
 		if(defined $parts[5]) {
@@ -122,7 +122,7 @@ sub getthemes
 		$ct->{"ent"} = decode_entities($parts[7]) if defined $parts[7];
 		$ct->{"prf"} = decode_entities($parts[8]) if defined $parts[8];
 		$ct->{"fep"} = $parts[10] if defined $parts[10];
-		$ct->{"lep"} = $parts[11] if defined $parts[11];
+		$ct->{"lep"} = $parts[12] if defined $parts[12];
 		push @ret, $ct;
 	    }
 	}