/**
* @since 0.4
*
* @param string $text
*
* @return string
*/
public function getSearchKey($text)
{
if ($text === null) {
return null;
}
if ($text === '') {
return '';
}
// composed normal form
$nfcText = $this->stringNormalizer->cleanupToNFC($text);
if (!is_string($nfcText) || $nfcText === '') {
wfWarn("Unicode normalization failed for `{$text}`");
}
// WARNING: *any* invalid UTF8 sequence causes preg_replace to return an empty string.
// Control character classes excluding private use areas.
$strippedText = preg_replace('/[\\p{Cc}\\p{Cf}\\p{Cn}\\p{Cs}]+/u', ' ', $nfcText);
// \p{Z} includes all whitespace characters and invisible separators.
$strippedText = preg_replace('/^\\p{Z}+|\\p{Z}+$/u', '', $strippedText);
if ($strippedText === '') {
// NOTE: This happens when there is only whitespace in the string.
// However, preg_replace will also return an empty string if it
// encounters any invalid utf-8 sequence.
return '';
}
//TODO: Use Language::lc to convert to lower case.
// But that requires us to load ALL the language objects,
// which loads ALL the messages, which makes us run out
// of RAM (see bug T43103).
$normalized = mb_strtolower($strippedText, 'UTF-8');
if (!is_string($normalized) || $normalized === '') {
wfWarn("mb_strtolower normalization failed for `{$strippedText}`");
}
return $normalized;
}
protected function addToOutput(Entity $entity, Status $status, $oldRevId = null)
{
$this->getResultBuilder()->addBasicEntityInformation($entity->getId(), 'entity');
$this->getResultBuilder()->addRevisionIdFromStatusToResult($status, 'entity', $oldRevId);
$params = $this->extractRequestParams();
if (isset($params['site']) && isset($params['title'])) {
$normTitle = $this->stringNormalizer->trimToNFC($params['title']);
if ($normTitle !== $params['title']) {
$this->getResultBuilder()->addNormalizedTitle($params['title'], $normTitle, 'normalized');
}
}
$this->getResultBuilder()->markSuccess(1);
}
/**
* Tries to find item id for given siteId and title combination
*
* @param string $siteId
* @param string $title
* @param bool $normalize
*
* @return ItemId|null
*/
private function getItemId($siteId, $title, $normalize)
{
// FIXME: This code is duplicated in SpecialItemByTitle::execute!
$title = $this->stringNormalizer->trimToNFC($title);
$id = $this->siteLinkLookup->getItemIdForLink($siteId, $title);
// Try harder by requesting normalization on the external site.
if ($id === null && $normalize === true) {
$siteObj = $this->siteStore->getSite($siteId);
//XXX: this passes the normalized title back into $title by reference...
$this->normalizeTitle($title, $siteObj);
$id = $this->siteLinkLookup->getItemIdForLink($siteObj->getGlobalId(), $title);
}
return $id;
}
/**
* Trims leading and trailing whitespace and performs unicode normalization
* by calling Wikibase\StringNormalizer::trimToNFC().
*
* @see StringNormalizer::normalize()
* @see Wikibase\StringNormalizer::trimToNFC()
*
* @param string $value the value to normalize
*
* @throws InvalidArgumentException if $value is not a string
* @return string the normalized value
*/
public function normalize($value)
{
return $this->normalizer->trimToNFC($value);
}