Update LJDC

This commit is contained in:
Gregory Soutade
2015-07-08 07:40:36 +02:00
parent b440ba1464
commit a20936c7f7
5 changed files with 149 additions and 47 deletions

View File

@@ -1,6 +1,7 @@
#!/usr/bin/perl
use LWP::Simple;
use HTML::Entities;
sub usage() {
print "usage: ./ljdc in_file [out.xml]\n"
@@ -66,6 +67,10 @@ while( defined( $address = <IN_FILE> ) ) {
$title = $1;
}
if (length($title) == 0 && $line =~ /<meta property="og:title" content="([^"]+)" \/>/) {
$title = $1;
}
# IMG
if ($line =~ /<p class="c1">.*<img.*src="([^"]+.gif)".*\/><\/p>/) {
$img = $1;
@@ -79,6 +84,42 @@ while( defined( $address = <IN_FILE> ) ) {
$img = $1;
last;
}
if ($line =~ /<div class="bodytype"> <p class="e"><img .*src="([^"]+.gif)".*>/) {
$img = $1;
last;
}
if ($line =~ /<div class="bodytype"> <p class="e"><img .*src="([^"]+.jpg)".*\/>/) {
$img = $1;
last;
}
if ($line =~ /<p class="e"><img.*src="([^"]+.jpg)">/) {
$img = $1;
last;
}
if ($line =~ /<p class="e"><img.*src="([^"]+.gif)">/) {
$img = $1;
last;
}
if ($line =~ /<\/source><img src="([^"]+\.gif)">/) {
$img = $1;
last;
}
if ($line =~ /<p class="e"><img src="([^"]+.gif)">/) {
$img = $1;
last;
}
if ($line =~ /<p class="c1">.*<img.*src="([^"]+.gif)".*><\/p>/) {
$img = $1;
last;
}
if ($line =~ /<p class="centredimg"><img.*src="([^"]+.gif)".*><\/p>/) {
$img = $1;
last;
}
if ($line =~ /^<p><img.*src="(http:\/\/ljdchost.com\/[^"]+.gif)".*><\/p>$/) {
$img = $1;
last;
}
}
if (length($title) == 0 || length($img) == 0)
@@ -86,9 +127,12 @@ while( defined( $address = <IN_FILE> ) ) {
print "Error with $address\n";
print "Title not defined\n" if (length($title) == 0);
print "IMG not defined\n" if (length($img) == 0);
print { STDERR } $contents;
exit 1;
}
else
{
$title = decode_entities($title);
print(OUT_FILE " <entry>\n");
print(OUT_FILE " <id>$id</id>\n");
print(OUT_FILE " <address>$address</address>\n");