|
#!/usr/bin/perl
use strict;
use warnings;
use LWP::Simple;
use HTML::Tree;
use URI;
use Time::localtime;
#download html files
my $caturl = 'http://www.ahtv.cn/';
my $remotefile;
my $localfile;
my $remotedir ='';
my $root = 'www.ahtv.cn';
my $locdir ;
my $tm = localtime;
mkdir($root);
$locdir = sprintf('%s/%04d%02d%02d/',$root,$tm->year +1900,$tm->mon+1,$tm->mday);
mkdir($locdir);
my $html = get($caturl);
my $file ;
my $tree = HTML::TreeBuilder->new;
$tree->parse_content($html); # !
foreach my $iframe ( $tree->look_down('_tag', 'IFRAME',sub{$_[0]->attr('src') =~ /article\.shtml$/}))
{
if($iframe)
{
my $link = $iframe->attr('src'); # !
my $uri = URI->new($link);
my @path=$uri->path_segments;
# There will always be an empty first component.
shift(@path);
$localfile = pop(@path);
print $uri->as_string,"\n";
print $localfile,"\n";
getstore($uri->as_string, "$locdir$localfile");
}
}
$tree->delete; # clear memory!
|
