public static function parserBeforeStrip(&$parser, &$text, &$stripState)
{
#wfDebug( 'HIPP: ' . __METHOD__ . "\n" );
# if revisionid is 0 this is not an article chunk
if (isset($parser->mDymFirstChunk) || !$parser->getVariableValue('revisionid') || $parser->getVariableValue('namespace')) {
return true;
}
$parser->mDymFirstChunk = 'done';
$title = $parser->getTitle();
$parser->mDymSees = DidYouMean::lookup($title->getArticleID(), $title->getText());
if (preg_match("/{{[sS]ee\\|([^}]*)}}/", $text, $see)) {
wfDebug("HIPP: see Hit\n");
$hasTemplate = true;
$sees = explode("|", $see[1]);
} elseif (preg_match("/{{[xX]see(\\|[^}]*)}}/", $text, $see)) {
wfDebug("HIPP: xsee Hit\n");
$hasTemplate = true;
preg_match_all("/\\|\\[\\[([^]|]*)(?:\\|([^|]*))?\\]\\](?: \\(([^)]*)\\))?/", $see[1], $ma);
$sees = $ma[1];
} else {
wfDebug("HIPP: (x)see Miss\n");
# there's no {{see}} in this chunk of wikitext
# if this is the 1st chunk of the article itself we can put an empty {{see}} there.
$hasTemplate = false;
$sees = array();
}
# normalize entities and urlencoding to pure utf-8
foreach ($sees as &$value) {
$value = rawurldecode(html_entity_decode($value, ENT_QUOTES, 'UTF-8'));
}
wfDebug('HIPP: Parser: ' . implode(', ', $sees) . "\n");
wfDebug('HIPP: DBase: ' . implode(', ', $parser->mDymSees) . "\n");
# add in the stuff from the database lookup
$sees = array_unique(array_merge($sees, $parser->mDymSees));
sort($sees);
wfDebug('HIPP: Merged: ' . implode(', ', $sees) . "\n");
# TODO is it better to use $parser->insertStripItem() ?
if (count($sees)) {
if (!$hasTemplate) {
// We need to squish in a fresh copy of the template...
$text = "{{see|}}\n" . $text;
}
$built_sees = DidYouMean::build_sees($sees);
} else {
$built_sees = '';
}
$text = preg_replace('/{{[xX]?[sS]ee\\|[^}]*}}/', $built_sees, $text);
return true;
}
