X-Git-Url: http://dolda2000.com/gitweb/?p=utils.git;a=blobdiff_plain;f=ANN.pm;h=6dd02e94fc951f0b1773cf7f5131f6ecf3359386;hp=ce85378553e89a9361808570c757c5345ec91ca6;hb=HEAD;hpb=b41d8fa9e6e09d448819e495c22adf23c4b02a10

diff --git a/ANN.pm b/ANN.pm
index ce85378..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
@@ -48,7 +48,8 @@ sub getlist
     my($name, $il, $html, @ret);
     ($name) = @_;
     
-    $il = uc(($name =~ /^(.)/)[0]);
+    $name = ($name =~ /^(the\s+)?(.*)$/i)[1];
+    $il = uc(($name =~ /^\W*(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
     if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
 	return undef;
@@ -57,7 +58,12 @@ sub getlist
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    push @ret, $1 while $html =~ /<A\s.*CLASS=HOVERLINE\s.*>.*<FONT.*>([^<>]*$name[^<>]*)<\/FONT/ig;
+    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;
     
     return @ret;
 }
@@ -67,7 +73,8 @@ sub getid
     my($name, $il, $html, $url);
     ($name) = @_;
     
-    $il = uc(($name =~ /^(.)/)[0]);
+    $name = ($name =~ /^(the\s+)?(.*)$/i)[1];
+    $il = uc(($name =~ /^\W*(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
     if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
 	return undef;
@@ -76,9 +83,13 @@ sub getid
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    (($url) = ($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*$name/i)) || return;
+    while($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];
+	}
+    }
     
-    return ($url =~ /\?id=(\d+)$/)[0];
+    return undef;
 }
 
 sub geturl
@@ -94,12 +105,12 @@ sub getthemes
     my($html, $kind, @ret);
     ($html, $kind) = @_;
     
-    if($html =~ /$kind theme:<\/b>\n/igc) {
+    if($html =~ /$kind theme:<\/strong>\s*\n/igc) {
 	my(@parts, $ct, $buf);
-	while($html =~ /\G\<br\>&nbsp; &nbsp; (([^<>]|\<i\>|<\/i>)+)/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+\((\<i\>(.*)<\/i>(;\s*)?)?([^<>]+)?\))?\"\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]) {
@@ -111,7 +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;
 	    }
 	}
@@ -130,19 +141,19 @@ sub getseries
     }
     
     $ret{"url"} = geturl $id;
-    ($buf) = ($html =~ /\<TITLE\>Anime News Network - ([^<]*)<\/TITLE>/);
+    ($buf) = ($html =~ /\<title\>([^<]*) - Anime News Network<\/title>/);
     if($buf =~ /\([^\)]+\)$/) {
 	($ret{"name"}, $ret{"type"}) = ($buf =~ /^(.*[^\s])\s*\(([^\)]+)\)$/);
     } else {
 	$ret{"name"} = $buf;
     }
-    if(($buf) = ($html =~ /vintage:<\/b>\n([^<]+)</is)) {
+    if(($buf) = ($html =~ /vintage:<\/strong>\s*\n\s*<span>([^<]+)</is)) {
 	$ret{"vtg"} = $buf;
     }
-    if(($buf) = ($html =~ /number of episodes:<\/b>\n([^<]+)</is)) {
+    if(($buf) = ($html =~ /number of episodes:<\/strong>\s*\n\s*<span>([^<]+)</is)) {
 	$ret{"eps"} = $buf;
     }
-    if(($buf) = ($html =~ /genres:<\/b>\n([^<]+)</is)) {
+    if(($buf) = ($html =~ /genres:<\/strong>\s*\n\s*([^<]+)</is)) {
 	$ret{"gnr"} = [split /, /, $buf];
     }
     $buf = getthemes $html, "opening";