Difference between revisions of "Import CSV data into a wiki"

From Organic Design wiki
m ({{lowercase}})
(remove protection)
Line 1: Line 1:
 
#!/usr/bin/perl
 
#!/usr/bin/perl
# {{perl}}{{#security:*|sysop}}{{Category:Robots}}{{lowercase}}
+
# {{perl}}{{Category:Robots}}{{lowercase}}
 
# - Licenced under LGPL (http://www.gnu.org/copyleft/lesser.html)
 
# - Licenced under LGPL (http://www.gnu.org/copyleft/lesser.html)
 
# - Authors:  [http://www.organicdesign.co.nz/Nad Nad] [http://www.organicdesign.co.nz/Sven Sven]
 
# - Authors:  [http://www.organicdesign.co.nz/Nad Nad] [http://www.organicdesign.co.nz/Sven Sven]

Revision as of 07:22, 4 September 2008

  1. !/usr/bin/perl
  2. Our Perl scripts.Automated scripts to perform batch automation.
  3. - Licenced under LGPL (http://www.gnu.org/copyleft/lesser.html)
  4. - Authors: Nad Sven
  5. - Source: http://www.organicdesign.co.nz/scraper.pl
  6. - Started: 2008-03-21
  7. - API: http://en.wikipedia.org/w/api.php
  1. Todo
  2. Make it so that if there is no title then it increments
  3. $hashref = { $wikitext =~ /\{{3}(.+?)(\|.*?)?\}{3}/g }

require('wiki.pl');

  1. Job, log and error files

$ARGV[0] or die "No job file specified!"; $ARGV[0] =~ /^(.+?)(\..+?)?$/;

  1. Set a debug conditional

$::debug = 0;

$::log = "$1.log"; $::err = "$1.err"; $::sep = ','; $::title = 0; $::template = 'Record'; $::prefix = ""; $::append = 0;

  1. Parse the job file

if (open JOB,'<',$ARGV[0]) { for (<JOB>) { if (/^\*?\s*csv\s*:\s*(.+?)\s*$/i) { $::csv = $1 } if (/^\*?\s*wiki\s*:\s*(.+?)\s*$/i) { $::wiki = $1 } if (/^\*?\s*user\s*:\s*(.+?)\s*$/i) { $::user = $1 } if (/^\*?\s*pass\s*:\s*(.+?)\s*$/i) { $::pass = $1 } if (/^\*?\s*separator\s*:\s*"(.+?)"\s*$/i) { $::sep = $1 } if (/^\*?\s*title\s*:\s*(.+?)\s*$/i) { $::title = $1 } if (/^\*?\s*template\s*:\s*(.+?)\s*$/i) { $::template = $1 } if (/^\*?\s*prefix\s*:\s*(.+?)\s*$/i) { $::prefix = $1 } if (/^\*?\s*append\s*:\s*(.+?)\s*$/i) { $::append = $1 } } close JOB; } else { die "Couldn't parse job file!" }


  1. Open CSV file and read in headings line

if (open CSV,'<',$::csv) { $_ = <CSV>; /^\s*(.+?)\s*$/; @headings = split /$::sep/i, $1; } else { die "Could not open CSV file!" }

  1. Log in to the wiki

wikiLogin($::wiki,$::user,$::pass) or exit;

  1. fetch the template if it exists

$response = $client->get("$wiki?title=Template:$template&action=raw"); if( $response->is_success ) {

 $wikitext = $response->content;
 # Remove noinclude areas
 $wikitext =~ s/.+?<\/noinclude>//gs;
 # Find all unique {{{parameters}}}
 # http://en.wikipedia.org/wiki/Help:Templates#Parameters
 while ($wikitext =~ /\{{3}(.+?)(\|.*?)?\}{3}/g ) {
   $params{$1} = undef;
 }
 # Create %{param=index} hash
 foreach( $i = 0; $i <= $#headings; $i++ ) {
     if(exists($params{$headings[$i]})) {
     $params{$headings[$i]} = $i;
     }
 }
 if($::debug) {
   print "\@headings: @headings\n";
   print "%params: @{[%params]}\n";
 }

}

  1. Get batch size and current number (also later account for n-bots)
  1. todo: log batch start
  1. Process the records

$n = 1; while (<CSV>) { /^\s*(.+?)\s*$/; @record = split /$::sep/, $1; $tmpl = "{{$template\n"; $tmpl .= "|$_ = $record[$params{$_}]\n" foreach (keys %params); $tmpl .= "}}"; print "Processing record ".$n++."\n"; if($::debug) { print "\$tmpl = $tmpl\n"; die "[\$::debug set exiting]\n" ; } # Update the record $text = wikiRawPage($::wiki,$record[$::title],0); $text .= "\n$tmpl" unless $text =~ s/\{\{$template.+?\}\}/$tmpl/is; if($append) { $done = wikiPageAppend($::wiki,$::prefix . $record[$::title],$text,"$::template appended using csv2wiki.pl"); } else { $done = wikiPageEdit($::wiki,$::prefix . $record[$::title],$text,"$::template replacement using csv2wiki.pl");

	}

# log a row error if any }

close CSV;