User:AnomieBOT/source/tasks/WatchlistUpdater.pm: Difference between revisions

Content deleted Content added
AnomieBOT (talk | contribs)
Updating published sources: General: * Add support for the AssertEdit extension. * BREAKING CHANGE: Assert "bot" by default. This means any non-flagged bot on a wiki with AssertEdit enabled will not be able to edit without
AnomieBOT (talk | contribs)
Updating published sources: General: * Update for the addition of 'rvslots'. DatedCategoryDeleterTest: * Disable. It's clear that task won't be needed. BrokenRedirectDeleter: * Handle pages with newlines before the <code>#REDIRECT</code>.
 
(19 intermediate revisions by 3 users not shown)
Line 1:
{{ombox|type=notice|text= Per [[WP:BOT#Approval]], any bot or automated editing process that only affects only the operators' user and talk pages (or subpages thereof), and which are not otherwise disruptive, may be run without prior approval.}}
<sourcesyntaxhighlight lang="perl">
package tasks::WatchlistUpdater;
 
Line 7:
=begin metadata
 
Bot: AnomieBOT
Task: WatchlistUpdater
BRFA: N/A
BRFA: N/A
Status: Begun 2008-08-15
Created: 2008-08-16
Rate: As needed, at most every 6 hours
 
Updates algorithmically-defined "watchlists" (like [[User:Anomie/uw-templates]])
Line 27 ⟶ 28:
@ISA=qw/AnomieBOT::Task/;
 
use POSIX qw/strftime/;
use Data::Dumper;
 
Line 37:
frequency => 6*60*60,
maxrows => 10000,
query => [{
list => 'allpages',
apprefix => 'Uw-',
apnamespace => '10',
aplimit => 'max'
}],
gcontinue => 'apcontinueallpages',
result => 'allpages',
match => {},
Line 73:
frequency => 6*60*60,
maxrows => 10000,
query => [{
list => 'allpages',
apprefix => 'AnomieBOT/',
apnamespace => '2',
aplimit => 'max'
},{
gcontinue list => 'apcontinueallpages',
apprefix => 'AnomieBOT II/',
apnamespace => '2',
aplimit => 'max'
},{
list => 'allpages',
apprefix => 'AnomieBOT III/',
apnamespace => '2',
aplimit => 'max'
},{
list => 'allpages',
apprefix => 'MediationBot/',
apnamespace => '2',
aplimit => 'max'
},{
list => 'allpages',
apprefix => 'MedcabBot/',
apnamespace => '2',
aplimit => 'max'
}],
gcontinue => 'allpages',
result => 'allpages',
match => {},
Line 102 ⟶ 122:
outerror => sub { "|-\n|colspan=\"2\"|".$_[1]."\n" },
outsuffix => sub { "|}" }
},
{
page => 'User:Anomie/index',
beginmarker => "\n<!-- SNIP HERE -->\n",
endmarker => '',
frequency => 6*60*60,
maxrows => 10000,
query => [{
list => 'allpages',
apprefix => 'Anomie/',
apnamespace => '2',
aplimit => 'max'
}],
gcontinue => 'allpages',
result => 'allpages',
match => {},
summary => 'Automatically updating userspace index',
botflag => 1,
outprefix => sub { "{| class=\"wikitable\"\n" },
outformat => sub {
my ($main, $talk);
if($_[1]{'ns'}==14 || $_[1]{'ns'}==6){
$main=':'.$_[1]{'title'};
} else {
$main=$_[1]{'title'};
}
if($_[1]{'ns'}==0){
$talk="Talk:".$_[1]{'title'};
} else {
$talk=$_[1]{'title'};
substr($talk, index($talk, ':'), 0)=' talk';
}
return "|-\n|[[$main]]||[[$talk]]\n",
},
outerror => sub { "|-\n|colspan=\"2\"|".$_[1]."\n" },
outsuffix => sub { "|}" }
},
{
page => 'User:AnomieBOT/nobots tests',
beginmarker => "\n<!-- SNIP HERE -->\n",
endmarker => '',
frequency => 6*60*60,
maxrows => 10000,
query => [{
list => 'allpages',
apprefix => 'AnomieBOT/nobots test ',
apnamespace => '2',
aplimit => 'max'
}],
gcontinue => 'allpages',
result => 'allpages',
match => {},
summary => 'Automatically updating list of bot exclusion tests',
botflag => 1,
outprefix => sub { "{{div col}}\n" },
keyforpage => sub { my $t=$_[0]{'title'}; return $t unless $t=~/ (\d+)$/; return sprintf("%08d", $1)."|$t"; },
outformat => sub {
my $t=$_[1]{'title'};
$t=~s/^[^|]*\|//;
return "* [[$t]]\n";
},
outerror => sub { "* <strong class=\"error\">".$_[1]."</strong>\n" },
outsuffix => sub { "\n{{div col end}}" }
}
);
Line 130 ⟶ 213:
my ($self, $api)=@_;
 
$api->task('WatchlistUpdater', 0, 10, qw(d::Timestamp));
$api->read_throttle(0);
$api->edit_throttle(10);
 
my $endtime=time()+300;
Line 152 ⟶ 233:
);
if($res->{'code'} ne 'success'){
$selfapi->warn("Failed to retrieve last edit date for $page: ".$res->{'error'}."\n");
return 60;
}
$res=[values(%{$res->{'query'}{'pages'}})];
if(exists($res->[0]{'revisions'}[0]{'timestamp'})){
$data->{'lastrun'}=$selfapi->ISO2timestamp($res->[0]{'revisions'}[0]{'timestamp'});
} else {
$data->{'lastrun'}=0;
Line 169 ⟶ 250:
my $tok=$api->edittoken($page);
if($tok->{'code'} eq 'shutoff'){
$selfapi->warn("Task disabled: ".$tok->{'content'}."\n");
return 300;
}
if($tok->{'code'} ne 'success'){
$selfapi->warn("Failed to retrieve edit token for $page: ".$tok->{'error'});
return 60;
}
if(exists($tok->{'missing'})){
$selfapi->warn("Page $page does not exist");
$data->{'lastrun'}=time();
next;
}
my $intxt=$tok->{'revisions'}[0]{'slots'}{'main'}{'*'};
 
# Generate new table
Line 187 ⟶ 268:
my $rows=0;
my %cont=();
my @queries=@{$data->{'query'}};
my $query=shift @queries;
do {
my $res=$api->query([$data->{'gcontinue'}], %{$data->{'query'}}, %cont);
if($res->{'code'} ne 'success'){
$selfapi->warn("Failed to retrieve data for $page: ".$res->{'error'});
return 60;
}
Line 199 ⟶ 282:
}
}
$query=shift @queries unless(%cont);
$res=$res->{'query'}{$data->{'result'}};
my @r;
Line 206 ⟶ 290:
@r=values %$res;
} else {
$selfapi->warn("Invalid data for $page: Not an array or hash ref");
return 60;
}
Line 212 ⟶ 296:
next if ($_->{'ns'}&1)==1;
next unless _match($data->{'match'}, $_);
my $outk=exists($data->{'keyforpage'})?$_data->{'titlekeyforpage'}}=($_):$_->{'title'};
$out{$k}=$_;
last if ++$rows>$data->{'maxrows'};
}
} while($rows<=$data->{'maxrows'} && values(%cont)$query);
my $x={};
my $table=$data->{'outprefix'}($x);
Line 237 ⟶ 322:
}
if($begin<0 || $end<0){
$selfapi->warn("Begin/end markers not found, refusing to edit $page\n");
} else {
substr($outtxt,$begin,$end-$begin)=$table;
if($intxt eq $outtxt){
$selfapi->warnlog("No update needed for $page\n");
} else {
my $res=$api->edit($tok, $outtxt, $data->{'summary'}, 0, $data->{'botflag'});
if($res->{'code'} ne 'success'){
$selfapi->warn("Write for $page failed: ".$res->{'error'});
next;
}
$selfapi->warnlog("Updated $page\n");
}
}
Line 293 ⟶ 378:
my $ok=1;
while(my ($k,$v)=each(%$match)){
my $v2=exists($value->{$k})?$value->{$k}: // undef;
$ok=($ok && _match($v,$v2));
}
Line 304 ⟶ 389:
1;
 
</syntaxhighlight>
</source>