MediaWiki Stubbing

I stubbed about 18000 entries in our StimulusWatch.org wiki so that people would have a little framework to work within when adding data, thus keeping the site relatively uniform in appearance while still providing a wiki. This took far longer than I think it should have. I didn't actually time it, but we're talking on the order of 12-18 hours. I used the MediaWiki perl module. The code is incredibly simple since you get to do something like this:

$c->setup({
  'bot' => {
  'user' => 'accountabot',
  'pass' => 'PASSWORD'},
  'wiki' => {
  'host' => 'stimuluswatch.org',
  'path' => 'mediawiki'
  }}) or die 'Error connecting to wiki';

...

while (($name,$state,$project_id, $cost, $jobs) = $sth->fetchrow_array()) {
  $page_name = "$state:$name:$project_id";

  $page_content = "== General Description == \n{{project-stub}}\n\n== Points in Favor ==\n\n== Points Against ==\n";
  print "$page_name ";
  print "\n";

   do {
       $rv = $c->text($page_name, $page_content);
       if ($rv != 1) {
           print "$rv\n";
       print "error: $c->{error}\n";
       }
   } while ($rv != 1);

}

And you just let that rip. I don't know if this is using a MediaWiki API or screen scraping (hoping on the former but not surprised about that latter). In any case, writes to MediaWiki through this interface are dismal. There must be a quicker way, but one night of processing was just under the threshold where I care because it's done. If I do have to run this again, I'm going to need to find a better way.