User:Wherebot/Source
From Wikipedia, the free encyclopedia
Here is the latest code as of 3/5/2007. There are no major problems with it at as of the time of writing.
Here is the source code. This has only been tested on UNIX-like systems, but it should theoretically also work on Windows. Note that the code was not intended for wide distribution, so it is not well-commented. Sorry! Also note that the code requires wget, pywikipediabot ,Yahoo's python search plugin, perl , and the Bot::BasicBot and IPC::Open2 perl modules. You may use the code under the GNU General Public License.
If you want to modify Wherebot to run on a different wiki or language, there are some modifications that need to be made. I have marked where people may want to do so on lines containing the text "#CONFIG."
Please go into edit mode to see the source of the program with proper linebreaks.
Here is the main file, cv-watch.pl. Place it where you wish:
#!/usr/bin/perl
use strict;
#some of the IRC parts of this bot are based off of the Bot::BasicBot sample code
Wherebot->new(channels => ["#en.wikipedia", "#en.wikiversity"], nick=>"Wherebot4", server => "irc.wikimedia.org")->run(); #CONFIG: change Wherebot4 to something unique
package Wherebot;
use base qw/Bot::BasicBot/;
use IPC::Open2;
sub said {
shift(); #don't care about the first parameter
our %hash = %{shift()};
our $rawMessage = $hash{"body"};
our $channel = $hash{"channel"};
our $site = $channel;
$site =~ s&#&&;
$rawMessage =~ m#02(http://$site.org[^ ]+)#;
our $url = $1;
#CONFIG: the next four lines are to ignore certain pages. Customize if you like
if ($url =~ /[Tt]alk:/) {return;}
if ($url =~ /Sandbox/) {return;}
if ($url =~ /Articles for deletion/) {return;}
if ($url =~ /Wikipedia:Introduction/) {return;}
chop $rawMessage;
if ($rawMessage =~ /N\x{03}10/) {
#CONFIG: the next four lines are to ignore certain namespaces. Customize if you like.
if ($url =~ /User:/) {return;}
if ($url =~ /Wikipedia:/) {return;}
if ($url =~ /Portal:/) {return;}
if ($url =~ /Help:/) {return;}
if ($url =~ /Template:/) {return;}
if ($url =~ /Category:/) {return;}
if ($url =~ /Image:/) {return;}
&act($channel, $url);
}
}
sub act {
our $misc = "/home/where/misc";
our $channel = shift;
our $url = shift;
$url =~ s#'##g; #just in case, although this would never be necessary
chop $url;
our $term = `wget '$url?action=raw' -q -O - | head -n 1`;
chomp $term;
our $origUrl = $url;
$url =~ m#/wiki/(.*)#;
our $page = $1;
$url .= "?action=raw";
$url =~ s#'##g; #shouldn't be a problem, but hey, I'm paranoid
chomp $term;
$term = &trim($term); #get it to <100 words so yahoo doesn't go crazy
if ($term =~ /#redirect/i) {
return;
}
if ($term =~ /^\{/) {
return;
}
if ($term =~ /^</) {
return;
}
$term =~ s#'''##g;
$term =~ s#''##g;
$term =~ s#\[\[##g;
$term =~ s#\]\]##g;
$term =~ s#\*##g;
$term =~ s#"##g; #Yahoo chokes on quotes; yes, this will probably return false matches, but it is better than the alternative
$term =~ s#\(##g;
$term =~ s#\)##g;
# if (m#([^\(\)]+)[\(\)]#) { #same thing with parenthesis
# $term = $1;
# }
if (length($term) < 75) {
return;
}
our $firstLine;
our $n=0;
while (1) {
our $pid = open2(*Reader, *Writer, "python", "$misc/search.py", "-t", "web", '"' . $term . '"'); #CONFIG: CHANGE $misc/search2.py to the path to search.py from the Yahoo search API
$firstLine = <Reader>;
# print "($url): FL: $firstLine\n";
if ($firstLine =~ /Internal WebService error, temporarily unavailable/ || $firstLine =~ /^Got an error/) {
warn "Search failed; retrying\n";
sleep 60;
waitpid $pid, 0;
++$n;
if ($n < 3) {
next;
}
else {
last;
}
}
else {
waitpid $pid, 0;
last;
}
}
if (!($firstLine =~ /^No results\s*/)) {
<Reader>;<Reader>; #skip some lines
our $from = <Reader>;
$from =~ s#\s##g;
if ($from =~ m#^http://en\.wikipedia\.org# || $from =~ m#\.gov# || $from =~ m#^http://en.wikibooks#) {
return;
}
#Get the page in the proper format
while ($page =~ /\%([0-9A-F][0-9A-F])/) {
my $hex = eval("0x" . $1);
my $char = chr($hex);
my $replace = "\%$1";
$page =~ s/$replace/$char/;
#$page =~ s/'/\\'/g;
}
$page =~ s#_# #g;
our $strippedUrl = $from;
$strippedUrl =~ s#^http://##;
#print "($page) copyvio from $from\n";
if ($channel eq "#en.wikipedia") { #CONFIG: change this line according to your language and version
chdir "$misc/pywikipedia"; #CONFIG: change this line according to where your pywikipedia directory is
}
print "Writing\n";
system "nice", "-n", "10", "python", "append.py", "* [[$page]] -- [$from $strippedUrl]. Reported at ~~~~~"; #CONFIG: change wording of how Wherebot reports if you like
}
}
sub trim { #cut parameter to <100 words
our $in = shift;
our @in = split / /, $in;
our $out = "";
our $i = 1;
for (@in) {
$out .= $_ . " ";
++$i;
if ($i == 99) {
last;
}
}
chop $out; #get rid of last space
return $out;
}
For reasons unknown the me, after long periods of time, the bot may shut down. I thus recommend running it using persist.pl:
#!/bin/perl
while (1) {
system "perl cv-watch.pl";
}
The following file, append.py, should go in the pywikipediabot directory.
#!/usr/bin/python
import wikipedia
import sys
site = wikipedia.getSite()
page = wikipedia.Page(site, "Wikipedia:Suspected copyright violations") #CONFIG: change this for what page to commit to
text = page.get()
text = text + "\n" + sys.argv[1]
wikipedia.setAction("Adding a suspected copyright violation") #CONFIG: change edit summary
page.put(text,minorEdit=False)
You need a user-config.py file in the pywikipediabot dir. Here's mine:
mylang='en' #CONFIG: change for your wiki language usernames['wikipedia']['en']='Wherebot' #CONFIG: change for your wiki, wiki language and username maxthrottle=2 put_throttle=3
Now run login.py in the pywikipediabot dir.
Finally, run persist.pl.
