User:Philosobot/Source code/afd/afd.cgi

1. !/usr/bin/perl

use POSIX; # the strftime function use CGI::Carp qw(fatalsToBrowser); use strict; undef $/;

use lib $ENV{HOME} . '/public_html/cgi-bin/wp/modules'; # absolute path to perl modules use lib '/home/philosobot/public_html/cgi-bin/wp/modules'; # absolute path to perl modules use lib '../wp/modules'; # relative path to perl modules

require 'bin/perlwikipedia_utils.pl'; require 'bin/get_html.pl';

1. Count and list the pages containing Wikipedia articles for deletion discussions (AfD).
2. Archive the pages on which all deletion discussions are closed.
3. Initialize pages for the upcoming days.

1. Discussions more than this number of days in the past are considered old and must be listed at AfD

my $afd_cutoff = 7;

my $gEditor;

MAIN: {

1. This line must be the first to print in a cgi script

print "Content-type: text/html\n\n";

$| = 1; # flush the buffer each line

print "Please be patient, this script can take a minute or two " . "if Wikipedia is slow ...

\n";

1. If the full path to the script is known (such as when running this
2. script from crontab), go to that directory first

my $cur_dir = $0; # $0 stands for the executable, with or without full path if ($cur_dir =~ /^\//){ $cur_dir =~ s/^(.*)\/.*?$/$1/g; chdir $cur_dir; }else{ }

1. The log in process must happen after we switched to the right directory as done above

$gEditor=wikipedia_login();

my $attempts = 10; my $sleep = 1;

my $summary_file = "Wikipedia:Articles_for_deletion/Old.wiki"; my $detailed_file = "Wikipedia:Articles_for_deletion/Old/Open AfDs.wiki";

1. Display the number of open afd discussions in the pages listed in
2. in $summary_file and put links to those those discussions in $detailed_file.
3. Return the list of pages in $combined_stats, we'll need that to decide which
4. pages to archive.

my $combined_stats = &count_and_list_open_AfDs($summary_file, $detailed_file, $attempts, $sleep);

1. Update the list of archived discussions, the ones that are no longer at AfD/Old,
2. which is the text in $combined_stats

my $archive_file = "Wikipedia:Archived deletion discussions.wiki"; &update_archived_discussions($archive_file, $combined_stats, $attempts, $sleep);

1. Initialize afd pages for the next several days

&initialize_new_afd_days($attempts, $sleep);

print "
Finished! One may now go back to " . "<a href=\"http://en.wikipedia.org/w/index.php?title=" . "Wikipedia:Articles_for_deletion/Philosophy/Old&action=purge\">" . "Wikipedia:Articles for deletion/Philosophy/Old</a>.
\n"; }

sub count_and_list_open_AfDs {

1. Display the number of open afd discussions in $summary_file and list them in $detailed_file.

my $summary_file = shift; my $detailed_file = shift; my $attempts = shift; my $sleep = shift;

my ($stats, $detailed_stats, $detailed_combined_stats);

$detailed_combined_stats = "

\n";

my ($text, $edit_summary, $error);

1. Fetch the summary file

$text = wikipedia_fetch($gEditor, $summary_file, $attempts, $sleep);

1. add the discussion from $afd_cutoff+1 days ago, if it is not already in $text

($text, $edit_summary) = &add_another_day ($text);

1. Find the number of open discussions for each listed day

my @lines = split("\n", $text); my $line; my ($brief_afd_link, $link, %stats_hash); foreach $line (@lines) {

1. extract the links to the open discussions

next unless ($line =~ /^\*\s*\[\[(Wikipedia:(?:Pages|Votes|Articles) for deletion\/Log\/\d+.*?)\]\]/); $link=$1; if ($link =~ /^(.*?)\|(.*?)$/){ $link = $1; $brief_afd_link = $2; }else { $brief_afd_link = $link; } print "Now doing $link ... ";

my $full_link = $link; $full_link =~ s/ /_/g; $full_link = 'http://en.wikipedia.org/wiki/' . $full_link;

($text, $error) = &get_html ($full_link);

1. see which AfD debates are not closed yet, and put that info in the link.
2. Get both a brief and a complete list, to put in different places.

($stats, $detailed_stats) = &see_open_afd_discussions ($link, $text, $detailed_file); $detailed_combined_stats = $detailed_combined_stats . $detailed_stats;

$line = '* ' . $brief_afd_link . ' ' . $stats; $stats_hash{$link} = "$line"; }

1. The file might have changed while we were doing the calculation above.
2. Get a new copy.

$text = wikipedia_fetch($gEditor, $summary_file, $attempts, $sleep); ($text, $edit_summary) = &add_another_day ($text); @lines = split("\n", $text);

1. gather all the info in $text

my $num_days = 0; # cout how many days are listed, ... my $num_open_disc = 0; # ... and how open many discussions my $combined_stats=""; foreach $line (@lines){ if ($line =~ /^\*\s*\[\[(Wikipedia:(?:Pages|Votes|Articles) for deletion\/Log\/\d+.*?)\s*(?:\||\]\])/) { $link=$1; $num_days++; if (exists $stats_hash{$link}) { $line = $stats_hash{$link}; # Overwite this line with the stats we found above } if ( $line =~ /\((\d+) open/ ){ $num_open_disc = $num_open_disc + $1; } } $combined_stats = $combined_stats . "$line\n"; }

my $utc_time=strftime("%H:%M, %B %d, %Y (UTC)", gmtime(time)); $combined_stats =~ s/(\/afd\/afd\.cgi.*?\]).*?\n/$1 \(last update at $utc_time\)\n/;

$edit_summary = "There are $num_open_disc open discussion(s) in $num_days day(s)." . $edit_summary; if ($num_open_disc > 200){ $edit_summary = "Big Backlog: " . $edit_summary; }

wikipedia_submit($gEditor, $summary_file, $edit_summary, $combined_stats, $attempts, $sleep); wikipedia_submit($gEditor, $detailed_file, $edit_summary, $detailed_combined_stats, $attempts, $sleep);

return $combined_stats; }

sub add_another_day{

my ($text, $afd_link, $hour_now, $thresh, $edit_summary, $SECONDS_PER_DAY, $brief_afd_link, $seconds);

$text = shift;

1. If beyond certain hour of the day (midnight GMT time),
2. add a link for the Afd discussion six days ago if not here yet

$hour_now=strftime("%H", localtime(time)); $thresh = 0; # midnight on gmt

if ($hour_now < $thresh){ return ($text, ""); }

($afd_link, $brief_afd_link) = &get_afd_link(-$afd_cutoff-1); # Older than $afd_cutoff

my $tag=; $edit_summary=""; if ($text !~ /\n\*\s*\[\[\Q$afd_link\E/){ $text =~ s/$tag/$tag\n\* \[\[$afd_link\|$brief_afd_link\]\]/g; $edit_summary=" Link to \[\[$afd_link\]\]."; }

return ($text, $edit_summary); }

sub get_afd_link {

my $days_from_now = shift;

my $SECONDS_PER_DAY = 60 * 60 * 24;

my $seconds = time() + $days_from_now * $SECONDS_PER_DAY; my $afd_link = strftime("Wikipedia:Articles for deletion/Log/%Y %B %d", localtime($seconds)); $afd_link =~ s/ 0(\d)$/ $1/g; # make 2005 April 01 into 2005 April 1

my $brief_afd_link = strftime("%d %B (%A)", localtime($seconds)); $brief_afd_link =~ s/^0//g;

return ($afd_link, $brief_afd_link); }

sub fmt_date {

my ($link, $date); $link = shift;

1. 'Wikipedia:Articles for deletion/Log/2006 December 16' --> '16 December'

if ($link =~ /^.*\/(\d+)\s+(\w+)\s+(\d+)/){ return "$3 $2"; }else{ return ""; }

}

sub extract_links{

1. Extract links to afd discussions and put them into a hash

my $text = shift; my @links_arr = ($text =~ /\[\[(Wikipedia:Articles for deletion\/Log\/\d+.*?)(?:\||\]\])/g);

my ($link, %links_hash); foreach $link(@links_arr){ $links_hash{$link} = 1; }

return %links_hash; }

sub update_archived_discussions {

1. Daily Afd pages that are at least six days old and that are no longer
2. at Afd/Old (where they are closed) are considered archived, and
3. should be added to the list of archived Afd pages.

my $archive_file = shift; # The name of the file containing the archives my $afd_text = shift; # What is at AfD/Old, those things should not yet be archived my $attempts = shift; my $sleep = shift;

1. The current text on the archive. We'll add to it.

my $archived_text = wikipedia_fetch($gEditor, $archive_file, $attempts, $sleep);

my ($curr_year, $prev_year);

1. Identify the discussions in AfD/Old, which won't be added to the archive

my %skip_archive = &extract_links($afd_text);

$curr_year = strftime("%Y", gmtime(time)); $prev_year = $curr_year - 1;

if ($archived_text !~ /==+\s*$curr_year\s*==+/){

1. Add section for current year if missing

if ($archived_text !~ /^(.*?)(==+\s*)$prev_year(\s*==+)(.*?)$/s){ print "Previous year section is missing, don't know what to do

\n"; return; }

1. Add current year above previous year

$archived_text = $1 . $2 . $curr_year . $3 . "\n" . $2 . $prev_year . $3 . $4; }

1. Any day in the current year up no earlier than
2. $afd_cutoff+2 days ago is a candidate to be in the archive
3. (unless, again, that page is still at AfD/Old). Days 0, -1,
4. -2, -3, -4, -5, , ... -$afd_cutoff are still open, while
5. day -$afd_cutoff-1 is now in the process of being closed.

my $start = -$afd_cutoff-2; my $stop = -366; my $day;

my ($new_links, $afd_link, $prev_afd_link, $link_sans_day, $prev_link_sans_day); my (@new_links_array);

@new_links_array = (); $new_links = ""; $prev_afd_link = "";

1. Add only the days from the current year to the archive.
2. Go in reverse, from the most recent date towards the past.

my $first_day = 1; # mark that this is the first day in the list for ($day = $start ; $day >= $stop ; $day--){

my ($afd_link, $brief_afd_link) = &get_afd_link($day);

1. Pages which are still at Afd/Old should not be archived yet.
2. Eventually after all discussions in such page are closed, the users will
3. remove the page from AfD/Old, and then the bot will get its hand on it.

next if (exists $skip_archive{$afd_link});

next unless ($afd_link =~ /\/$curr_year/); # deal only with the current year

1. See if to add a section separating two months

$link_sans_day = $afd_link; $link_sans_day =~ s/\s*\d+$//g; $prev_link_sans_day = $prev_afd_link; $prev_link_sans_day =~ s/\s*\d+$//g;

1. Add a section heading only if we are between months or we arrived
2. at the most recent day

if ( $first_day || ($link_sans_day ne $prev_link_sans_day && $prev_link_sans_day ne "") ){

$link_sans_day =~ s/^(.*)\/(.*?)$/Deletion discussions\/$2/g; $new_links = $new_links . "\n===$link_sans_day===\n\n";

$first_day = 0; # First day passed }

$new_links = $new_links . "* $afd_link\n"; push(@new_links_array, $afd_link);

1. Prepare for the next loop

$prev_afd_link = $afd_link; }

1. Before updating $archived_text, see what is there currently, so that
2. we can see what changed and put that in the edit summary.

if ($archived_text !~ /^(.*?==+\s*$curr_year\s*==+)(.*?)(==+\s*$prev_year\s*==.*?)$/s) { print "Previous year section is missing, don't know what to do

\n"; return; }

my $p1 = $1; my $existing_text = $2; my $p3 = $3;

1. See what links are in @new_links_array and are not in %existing_links.
2. Put those in the edit summary.

my %existing_links = &extract_links($existing_text);

my $edit_summary = ""; foreach $afd_link (@new_links_array){ if (!exists $existing_links{$afd_link}){

1. This is a link which will be added to the archive now and which was
2. not there before

$edit_summary = $edit_summary . "$afd_link "; } }

if ($edit_summary eq ""){ print "No new pages to archive

\n"; return; }

1. Replace in $archived_text the portion corresponding to the links for this year
2. with $new_links which contains the newly archived links

$archived_text = $p1 . "\n" . $new_links . "\n" . $p3;

$edit_summary = "Archiving " . $edit_summary;

wikipedia_submit($gEditor, $archive_file, $edit_summary, $archived_text, $attempts, $sleep); }

sub see_open_afd_discussions (){ my $link = shift; my $text = shift; my $detailed_file = shift;

my $stats = "";

$text =~ s/\n//g; # rm newlines

1. strip the top part, as otherwise it confuses the parser below

$text =~ s/^.*?\

//sg;

1. some processing to deal with AfD ambiguity recently

$text =~ s/\"boilerplate[_\s]+metadata[_\s+][avp]fd.*?\"/\"boilerplate metadata vfd\"/ig;

$text =~ s/(\<div\s+class\s*=\s*\"boilerplate metadata vfd\".*?\<span\s+class\s*=\s*\"editsectio)(n)(.*?\>)/$1p$3/sgi;

my @all = ($text =~ /\<span\s+class\s*=\s*\"editsectio\w\".*?\>\[\<a href\s*=\s*\"\/w\/index.php\?title\s*=\s*(Wikipedia:\w+[_\s]for[_\s]deletion.*?)\"/g );

my @open = ($text =~ /\<span\s+class\s*=\s*\"editsection\".*?\>\[\<a href\s*=\s*\"\/w\/index.php\?title\s*=\s*(Wikipedia:\w+[_\s]for[_\s]deletion.*?)\"/g );

my @closed = ($text =~ /\<span\s+class\s*=\s*\"editsectiop\".*?\>\[\<a href\s*=\s*\"\/w\/index.php\?title\s*=\s*(Wikipedia:\w+[_\s]for[_\s]deletion.*?)\"/g );

my $openc=0; foreach (@open) {

next if (/Wikipedia:\w+[_\s]for[_\s]deletion\/Log/i); next unless (/\&section=(T-|)1/); s/\&.*?$//g; $openc++;

$stats = "$stats " . "\[\[$_\|$openc]]"; } print "($openc open / ";

my $closedc=0; foreach (@closed) { next if (/Wikipedia:\w+[_\s]for[_\s]deletion\/Log/i); next unless (/\&section=(T-|)1/); s/\&.*?$//g; $closedc++;

1. print "$closedc: $_\n";

} print "$closedc closed / ";

my $allc=0; foreach (@all) { next if (/Wikipedia:\w+[_\s]for[_\s]deletion\/Log/i); next unless (/\&section=(T-|)1/); s/\&.*?$//g; $allc++;

1. print "$allc: $_\n";

} print "$allc total discussions)
\n";

1. some gimmickry, to list to sections in $detailed_file.

my $detailed_stats = $stats; my $short_link = $link; $short_link =~ s/^.*\///g; $detailed_file =~ s/\.wiki$//g;

1. if there are too many open afds, link to the file listing them. Otherwise, list them right here.

if ($openc == 0 ){ $stats = "($openc open / $closedc closed / $allc total discussions)"; }elsif ( $openc > 20 ){ $stats = "($openc open / $closedc closed / $allc total discussions; see open)"; }else{ $stats = "($openc open / $closedc closed / $allc total discussions; open: $stats)"; }

my $http_link = $link; $http_link =~ s/ /_/g; $http_link = '(. $http_link . '&action=edit edit this day\'s list)';

1. text to add to a subpage listing all open discussions

$detailed_stats =~ s/\s*\[\[(.*?)\|\d+\]\]/\* \[\[$1\]\]\n/g; $detailed_stats =~ s/_/ /g; $detailed_stats = "==$short_link==\n" . $http_link . "\n" . $detailed_stats;

return ($stats, $detailed_stats); }

sub initialize_new_afd_days {

1. Initialize afd pages for the next several days by putting in a
2. preamble for each day. When such a future day becomes today, the
3. users will edit that afd page page and will add afd listings below
4. the preamble.

print "\n\n

Initializing AfD pages for the next several days
\n";

1. Parameters related to fetching/submitting data to Wikipedia

my $attempts = shift; my $sleep = shift;

my ($day);

for ($day = 1 ; $day < 5 ; $day++){

my ($prev_afd_link, $prev_afd_name) = &get_afd_link($day - 1); my ($curr_afd_link, $curr_afd_name) = &get_afd_link($day + 0); my ($next_afd_link, $next_afd_name) = &get_afd_link($day + 1);

my $days_page = $curr_afd_link . ".wiki"; my $days_text = wikipedia_fetch($gEditor, $days_page, $attempts, $sleep);

if ($days_text !~ /^\s*$/){

1. This day's page is not empty, so it was already initialized. Skip it.

print "Page exists

\n\n"; next; }

1. Form the page for the current day

$days_text = &get_page_text($prev_afd_link, $prev_afd_name, $next_afd_link, $next_afd_name);

1. Initialize the page for the day

print "\n
Initializing $curr_afd_link
\n"; my $edit_summary = "Initializing a new AfD day"; wikipedia_submit($gEditor, $days_page, $edit_summary, $days_text, $attempts, $sleep);

}

}

sub get_page_text {

my ($prev_afd_link, $prev_afd_name, $next_afd_link, $next_afd_name) = @_;

1. Strip the text in parentheses from the text "1 February (Sunday)"

$prev_afd_name =~ s/\s*\(.*?\)\s*//g; $next_afd_name =~ s/\s*\(.*?\)\s*//g;

return '

\'\'\'Guide to deletion\'\'\'

';

}