addbookmarktitle

gistfile1.pl

#!/usr/bin/perl

use warnings;
use strict;
use utf8;
use LWP::UserAgent;
use Mozilla::CA;
use Encode;
use Encode::Guess;

my $InputFile = $ARGV[0];
my $OutputFile = $ARGV[1];

open(my $InputHandle, "<$InputFile") or die $!;
open(my $OutputHandle, ">$OutputFile") or die $!;

my $n = 1;
my $ua = LWP::UserAgent->new;
$ua->agent("Mozilla/5.0");
$ua->ssl_opts(SSL_ca_file => Mozilla::CA::SSL_ca_file());
$ua->timeout(10);

while(my $line = <$InputHandle>){
	chomp($line);
#	last if ($n > 4);
	if ($line =~ /^(.*?)<A HREF=\"(.*?)\"(.*?)>.*?<\/A>$/) {
		my ($mae, $url, $ushiro) = ($1, $2, $3);
		my $req = HTTP::Request->new(GET => $url);
		print "$n. $url ";
		my $res = $ua->request($req);
		if ($res->is_success) {
			if ($res->content =~ /<title>(.*?)<\/title>/) {
				my $title = $1;
				my $decoder = guess_encoding($res->content, qw/shiftjis euc-jp iso-2022-jp 7bit-jis ascii/);
				if (ref($decoder)) {
					$title = $decoder->decode($title);
					$title = encode('utf-8', $title);
				} else {
					$decoder = guess_encoding($res->content, qw/cp932/);
					if (ref($decoder)) {
						$title = $decoder->decode($title);
						$title = encode('utf-8', $title);
					}
				}
				print "[$title]\n";
				print $OutputHandle "$mae<A HREF=\"$url\"$ushiro>$title</A>\n";
			} else {
				print "notitle\n";
				print $OutputHandle "$line\n";
			}
		}
		else {
			print $res->status_line, "\n";
			print $OutputHandle "$line\n";
		}
		$n++;
	} else {
		print $OutputHandle "$line\n";
	}
}

close($InputHandle);
close($OutputHandle);
exit;