X-Git-Url: http://dolda2000.com/gitweb/?a=blobdiff_plain;f=ANN.pm;h=aefd04efca61573395bb688e492644f6f00f531e;hb=edeeab7e12944f6a0935f4c27108032f0b8ec0e0;hp=fbf08e55a88d0d943c556006168fd93890ff8a2e;hpb=23d4abb2e29634e8d2e11879be91cec9cfd6ad13;p=utils.git

diff --git a/ANN.pm b/ANN.pm
index fbf08e5..aefd04e 100644
--- a/ANN.pm
+++ b/ANN.pm
@@ -1,21 +1,45 @@
 package Anime::ANN;
 
+use POSIX;
+use Digest::MD5;
 use LWP::UserAgent;
 use HTML::Entities;
 
 $ver = "0.1";
+$ua = LWP::UserAgent->new;
+$ua->agent("ANNData/$ver ");
+$usecache = 1;
 
 sub _get
 {
-    my($uri, $ua, $res);
+    my($uri, $cname, $res);
     ($uri) = @_;
 
-    $ua = LWP::UserAgent->new;
-    $ua->agent("ANNData/$ver ");
-
+    $cname = $ENV{"HOME"} . "/.ann/";
+    mkdir $cname unless -e $cname;
+    $cname .= "cache/";
+    mkdir $cname unless -e $cname;
+    $cname .= Digest::MD5::md5_hex $uri;
+    if($usecache && -e $cname) {
+	my(@s);
+	@s = stat $cname;
+	if((time - $s[9]) < 86400) {
+	    my($buf);
+	    open CACHE, "<:utf8", $cname;
+	    $res .= $buf while read CACHE, $buf, 1024;
+	    close CACHE;
+	    return $res;
+	}
+    }
+    
     $res = $ua->request(HTTP::Request->new("GET", "$uri"));
-
-    die "could not fetch $uri\n" unless $res->is_success;
+    
+    if(open CACHE, ">:utf8", $cname) {
+	print CACHE $res->content;
+	close CACHE;
+    }
+    
+    return undef unless $res->is_success;
     return $res->content;
 }
 
@@ -26,14 +50,16 @@ sub getlist
     
     $il = uc(($name =~ /^(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
-    $html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il";
+    if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
+	return undef;
+    }
     
     # 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;
     
-    return(@ret);
+    return @ret;
 }
 
 sub getid
@@ -43,7 +69,9 @@ sub getid
     
     $il = uc(($name =~ /^(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
-    $html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il";
+    if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
+	return undef;
+    }
     
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
@@ -53,6 +81,14 @@ sub getid
     return ($url =~ /\?id=(\d+)$/)[0];
 }
 
+sub geturl
+{
+    my($id);
+    ($id) = @_;
+    
+    return "http://www.animenewsnetwork.com/encyclopedia/anime.php?id=$id";
+}
+
 sub getthemes
 {
     my($html, $kind, @ret);
@@ -67,13 +103,13 @@ sub getthemes
 		$ct = {};
 		$ct->{"num"} = $parts[1] if defined $parts[1];
 		if(defined $parts[5]) {
-		    $ct->{"tit"} = $parts[5];
+		    $ct->{"tit"} = decode_entities($parts[5]);
 		    $ct->{"jat"} = decode_entities($parts[2]) if defined $parts[2];
 		} else {
-		    $ct->{"tit"} = $parts[2] if defined $parts[2];
+		    $ct->{"tit"} = decode_entities($parts[2]) if defined $parts[2];
 		}
-		$ct->{"ent"} = $parts[7] if defined $parts[7];
-		$ct->{"prf"} = $parts[8] if defined $parts[8];
+		$ct->{"ent"} = decode_entities($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];
 		push @ret, $ct;
@@ -86,19 +122,30 @@ sub getthemes
 
 sub getseries
 {
-    my($id, $buf, $html, %ret);
+    my($id, $buf, $html, %ret, @cp);
     ($id) = @_;
     
-    $html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?id=$id";
+    if(!($html = _get geturl $id)) {
+	return undef;
+    }
     
-    $ret{"url"} = "http://www.animenewsnetwork.com/encyclopedia/anime.php?id=$id";
-    ($ret{"name"}) = ($html =~ /\<TITLE\>Anime News Network - ([^<]*)<\/TITLE>/);
+    $ret{"url"} = geturl $id;
+    ($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)) {
 	$ret{"vtg"} = $buf;
     }
     if(($buf) = ($html =~ /number of episodes:<\/b>\n([^<]+)</is)) {
 	$ret{"eps"} = $buf;
     }
+    if(($buf) = ($html =~ /genres:<\/b>\n([^<]+)</is)) {
+	@cp = split /,\s*/, $buf;
+	$ret{"gnr"} = \@cp;
+    }
     $buf = getthemes $html, "opening";
     $ret{"op"} = $buf if(@{$buf});
     $buf = getthemes $html, "ending";