User:Adodge/WLWP

#!/usr/bin/perl
# Watchlist Wikipedia v0.1b
# Alexdodge.net
# Licensed under Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported
# http://creativecommons.org/licenses/by-nc-sa/3.0/
use strict;
use warnings;
use XML::FeedPP;
use URI::Escape;
use Getopt::Std;
 
my $version = "WLWP v0.1b (http://alexdodge.net/)";
 
my %opts = (
        u => undef,  #-u <username>
        p => undef,  #-p <password>
        o => undef,  #-o <FILE> Select output file.
        t => undef,  #-t Don't list talk pages.
        s => undef,  #-s Output to STDOUT
        q => undef,  #-q Quiet mode
        d => undef,  #-d <FILE> Output subscription list to file. (Doesn't create RSS Feed.)
        l => undef,  #-l <FILE> Load subscription list from file.
        n => undef,  #-n <name> Name of feed
        h => undef,  #-h Show Help
        v => undef,
        n => undef,
);
getopts('n:u:p:o:tsqd:l:hv', \%opts);
 
if(defined $opts{'h'}){
        print <<_HELP_;
USAGE: wlwp [-u <USERNAME> -p <PASSWORD> [-d <FILE>]] [-l <FILE>] [-o <FILE>] [-n <NAME>] -tsqhv
wlwp creates an rss feed from a wikipedia watchlist or a file with a list of wikipedia articles.
 
$version
 
Licensed under Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported
http://creativecommons.org/licenses/by-nc-sa/3.0/
 
EXAMPLE OF USE
wlwp -u adodge -o wlwp.rss
 
OPTIONS
        -u <USERNAME>
USERNAME is the username of the user on wikipedia.  If this is not specified and -l is not set, the user will be prompted interactively.
 
        -p <PASSWORD>
PASSWORD is the password of the user on wikipedia.  If this is not specified and -l is not set, the user will be prompted interactively.
 
        -d <FILE>
If set, wlwp will dump the watchlist into FILE.  No rss feed will be created.
 
        -l <FILE>
If set, wlwp will load the watchlist from FILE.  If set, the username and password need not be provided.
 
        -o <FILE>
wlwp will output the rss feed to FILE.  If this is not specified and -s is not set, the user will be prompted interactively.
 
        -n <NAME>
NAME is the title of the RSS feed.  If this is not set, the title will be either the username or the watchlist filename.
 
SWITCHES
        -t
Do not subscribe automatically to article talk pages.
 
        -s
If -o is not set, do not prompt the user interactively.  Instead, output to STDOUT.
 
        -q
Quiet mode.  Interactive prompts will still be printed.
 
        -h
Show this help and quit.
 
        -v
Print version information and quit.
 
_HELP_
        exit;
}
 
if(defined $opts{'v'}){
        print $version."\n";
        exit;
}
 
 
my ($username, $password, $outfile);
 
if(!defined $opts{'l'}){
        if(defined($opts{'u'})){
                $username = $opts{'u'};
        }else{
                print STDERR "Enter username: ";
                $username = <STDIN>;
                chomp $username;
        }
 
        if(defined($opts{'p'})){
                $password = $opts{'p'};
        }else{
                print STDERR "Enter password: ";
                $password = <STDIN>;
                chomp $password;
        }
}
 
my $title;
 
if(defined $opts{'n'}){
        $title = $opts{'n'};
}elsif(defined $opts{'l'}){
        $title = $opts{'l'};
}else{
        $title = $username;
}
 
if(!defined $opts{'d'}){
        if(defined($opts{'o'})){
                $outfile = $opts{'o'};
        }else{
                if(!defined($opts{'s'})){
                        print STDERR "Enter file for output (blank for STDOUT): ";
                        $outfile = <STDIN>;
                        chomp $outfile;
                }
        }
}
 
my @list;
 
if(!defined $opts{'l'}){
        `curl -c ~/.wlwpcookies -A "$version" -d wpName=$username -d wpPassword=$password -d wpRemember=1 -d wpLoginattempt=Log+in "http://en.wikipedia.org/w/index.php?title=Special:Userlogin&action=submitlogin&type=login&returnto=Special:Watchlist" 2> /dev/null`;
        my $data = `curl -b ~/.wlwpcookies -A "$version" "http://en.wikipedia.org/w/index.php?title=Special:Watchlist&action=raw" 2> /dev/null`;
        die("Failure") if !($data =~ /<textarea id="titles" name="titles" rows="25" cols="80">(.*)<\/textarea>/s);
        @list = split("\n", $1);
}else{
        open (FD, '<', $opts{'l'});
        while(<FD>){
                chomp;
                push @list, $_;
        }
        close FD;
}
 
if(defined $opts{'d'}){
        open (FD, '+>', $opts{'d'});
        foreach my $item (@list){
                print FD $item."\n";
        }
        exit;
}
 
my @talkpages;
 
if(!defined $opts{'t'}){
        foreach my $o (@list){
                my $item = $o;
                if($item =~ /\w+:.*/){
                        $item =~ s/(\w*):/$1_talk:/;
                }else{
                        $item = "Talk:$item";
                }
                push(@talkpages, $item);
        }
}
 
my $feed = XML::FeedPP::RSS->new();
$feed->title("WLWP - $title");
$feed->description("$version");
$feed->link("http://en.wikipedia.org/w/index.php?title=Special:Watchlist");
 
foreach my $item_orig (@list, @talkpages){
        print STDERR "$item_orig ... " if(!defined $opts{'q'});
        my $item = uri_escape($item_orig);
        my $tmpfeed = XML::FeedPP::RSS->new("http://en.wikipedia.org/w/index.php?title=$item&action=history&feed=rss");
        if( uri_unescape(($tmpfeed->get_item(0))->description()) =~ /<p><\/p>\n<p>Can\x27t load revision \d+<\/p>/){
                print STDERR "DNE" if(!defined $opts{'q'});
        }else{
                print STDERR "ok" if(!defined $opts{'q'});
                my $i = 0;
                while(my $itemobj = $tmpfeed->get_item($i)){
                        $itemobj->title($item_orig." - ".$itemobj->title());
                        $i ++;
                }
                $feed->merge($tmpfeed);
        }
        print STDERR "\n" if(!defined $opts{'q'});
}
 
my $rss = $feed->to_string();
 
if(!defined $outfile || $outfile eq ""){
        print $rss;
}else{
        open(FD,"+>",$outfile);
        print FD $rss;
        close FD;
}