/**
* {@inheritdoc}
*/
public function transform($value, MigrateExecutableInterface $migrate_executable, Row $row, $destination_property)
{
$new_value = $this->transliteration->transliterate($value, LanguageInterface::LANGCODE_DEFAULT, '_');
$new_value = strtolower($new_value);
$new_value = preg_replace('/[^a-z0-9_]+/', '_', $new_value);
return preg_replace('/_+/', '_', $new_value);
}
/**
* {@inheritdoc}
*/
public function preprocessIndexValue($value, $type = 'text')
{
if ($type == 'text') {
return $value;
}
return Unicode::strtolower($this->transliterator->transliterate($value));
}
/**
* Transliterates a string in given language. Various postprocessing possible.
*
* @param \Symfony\Component\HttpFoundation\Request $request
* The input string and language for the transliteration.
* Optionally may contain the replace_pattern, replace, lowercase parameters.
*
* @return \Symfony\Component\HttpFoundation\JsonResponse
* The transliterated string.
*/
public function transliterate(Request $request) {
$text = $request->query->get('text');
$langcode = $request->query->get('langcode');
$replace_pattern = $request->query->get('replace_pattern');
$replace_token = $request->query->get('replace_token');
$replace = $request->query->get('replace');
$lowercase = $request->query->get('lowercase');
$transliterated = $this->transliteration->transliterate($text, $langcode, '_');
if ($lowercase) {
$transliterated = Unicode::strtolower($transliterated);
}
if (isset($replace_pattern) && isset($replace)) {
if (!isset($replace_token)) {
throw new AccessDeniedException("Missing 'replace_token' query parameter.");
}
elseif (!$this->tokenGenerator->validate($replace_token, $replace_pattern)) {
throw new AccessDeniedException("Invalid 'replace_token' query parameter.");
}
// Quote the pattern delimiter and remove null characters to avoid the e
// or other modifiers being injected.
$transliterated = preg_replace('@' . strtr($replace_pattern, ['@' => '\@', chr(0) => '']) . '@', $replace, $transliterated);
}
return new JsonResponse($transliterated);
}
/**
* {@inheritdoc}
*/
public function sortResults(Result $a, Result $b)
{
$a = $this->transliteration->removeDiacritics($a->getDisplayValue());
$b = $this->transliteration->removeDiacritics($b->getDisplayValue());
if ($a == $b) {
return 0;
}
return strnatcasecmp($a, $b);
}
/**
* Tests machine name transformation of non-alphanumeric characters.
*/
public function testMachineNames()
{
// Tests the following transformations:
// - non-alphanumeric character (including spaces) -> underscore,
// - Uppercase -> lowercase,
// - Multiple consecutive underscore -> single underscore.
$human_name_ascii = 'foo2, the.bar;2*&the%baz!YEE____HaW ';
$human_name = $human_name_ascii . 'áéő';
$expected_result = 'foo2_the_bar_2_the_baz_yee_haw_aeo';
// Test for calling transliterate on mock object.
$this->transliteration->expects($this->once())->method('transliterate')->with($human_name)->will($this->returnValue($human_name_ascii . 'aeo'));
$plugin = new MachineName(array(), 'machine_name', array(), $this->transliteration);
$value = $plugin->transform($human_name, $this->migrateExecutable, $this->row, 'destinationproperty');
$this->assertEquals($expected_result, $value);
}
/**
* Transliterates a string in given language. Various postprocessing possible.
*
* @param \Symfony\Component\HttpFoundation\Request $request
* The input string and language for the transliteration.
* Optionally may contain the replace_pattern, replace, lowercase parameters.
*
* @return \Symfony\Component\HttpFoundation\JsonResponse
* The transliterated string.
*/
public function transliterate(Request $request)
{
$text = $request->query->get('text');
$langcode = $request->query->get('langcode');
$replace_pattern = $request->query->get('replace_pattern');
$replace = $request->query->get('replace');
$lowercase = $request->query->get('lowercase');
$transliterated = $this->transliteration->transliterate($text, $langcode, '_');
if ($lowercase) {
$transliterated = Unicode::strtolower($transliterated);
}
if (isset($replace_pattern) && isset($replace)) {
$transliterated = preg_replace('@' . $replace_pattern . '@', $replace, $transliterated);
}
return new JsonResponse($transliterated);
}
/**
* Transliterates a string in given language. Various postprocessing possible.
*
* @param \Symfony\Component\HttpFoundation\Request $request
* The input string and language for the transliteration.
* Optionally may contain the replace_pattern, replace, lowercase parameters.
*
* @return \Symfony\Component\HttpFoundation\JsonResponse
* The transliterated string.
*/
public function transliterate(Request $request)
{
$text = $request->query->get('text');
$langcode = $request->query->get('langcode');
$replace_pattern = $request->query->get('replace_pattern');
$replace = $request->query->get('replace');
$lowercase = $request->query->get('lowercase');
$transliterated = $this->transliteration->transliterate($text, $langcode, '_');
if ($lowercase) {
$transliterated = Unicode::strtolower($transliterated);
}
if (isset($replace_pattern) && isset($replace)) {
// Quote the pattern delimiter and remove null characters to avoid the e
// or other modifiers being injected.
$transliterated = preg_replace('@' . strtr($replace_pattern, ['@' => '\\@', chr(0) => '']) . '@', $replace, $transliterated);
}
return new JsonResponse($transliterated);
}
/**
* {@inheritdoc}
*/
public function cleanString($string, array $options = array())
{
if (empty($this->cleanStringCache)) {
// Generate and cache variables used in this method.
$config = $this->configFactory->get('pathauto.settings');
$this->cleanStringCache = array('separator' => $config->get('separator'), 'strings' => array(), 'transliterate' => $config->get('transliterate'), 'punctuation' => array(), 'reduce_ascii' => (bool) $config->get('reduce_ascii'), 'ignore_words_regex' => FALSE, 'lowercase' => (bool) $config->get('case'), 'maxlength' => min($config->get('max_component_length'), $this->aliasStorageHelper->getAliasSchemaMaxLength()));
// Generate and cache the punctuation replacements for strtr().
$punctuation = $this->getPunctuationCharacters();
foreach ($punctuation as $name => $details) {
$action = $config->get('punctuation.' . $name);
switch ($action) {
case PathautoManagerInterface::PUNCTUATION_REMOVE:
$cache['punctuation'][$details['value']] = '';
$this->cleanStringCache;
case PathautoManagerInterface::PUNCTUATION_REPLACE:
$this->cleanStringCache['punctuation'][$details['value']] = $this->cleanStringCache['separator'];
break;
case PathautoManagerInterface::PUNCTUATION_DO_NOTHING:
// Literally do nothing.
break;
}
}
// Generate and cache the ignored words regular expression.
$ignore_words = $config->get('ignore_words');
$ignore_words_regex = preg_replace(array('/^[,\\s]+|[,\\s]+$/', '/[,\\s]+/'), array('', '\\b|\\b'), $ignore_words);
if ($ignore_words_regex) {
$this->cleanStringCache['ignore_words_regex'] = '\\b' . $ignore_words_regex . '\\b';
if (function_exists('mb_eregi_replace')) {
mb_regex_encoding('UTF-8');
$this->cleanStringCache['ignore_words_callback'] = 'mb_eregi_replace';
} else {
$this->cleanStringCache['ignore_words_callback'] = 'preg_replace';
$this->cleanStringCache['ignore_words_regex'] = '/' . $this->cleanStringCache['ignore_words_regex'] . '/i';
}
}
}
// Empty strings do not need any processing.
if ($string === '' || $string === NULL) {
return '';
}
$langcode = NULL;
if (!empty($options['language'])) {
$langcode = $options['language']->getId();
} elseif (!empty($options['langcode'])) {
$langcode = $options['langcode'];
}
// Check if the string has already been processed, and if so return the
// cached result.
if (isset($this->cleanStringCache['strings'][$langcode][(string) $string])) {
return $this->cleanStringCache['strings'][$langcode][(string) $string];
}
// Remove all HTML tags from the string.
$output = Html::decodeEntities($string);
$output = PlainTextOutput::renderFromHtml($output);
// Optionally transliterate.
if ($this->cleanStringCache['transliterate']) {
// If the reduce strings to letters and numbers is enabled, don't bother
// replacing unknown characters with a question mark. Use an empty string
// instead.
$output = $this->transliteration->transliterate($output, $langcode, $this->cleanStringCache['reduce_ascii'] ? '' : '?');
}
// Replace or drop punctuation based on user settings.
$output = strtr($output, $this->cleanStringCache['punctuation']);
// Reduce strings to letters and numbers.
if ($this->cleanStringCache['reduce_ascii']) {
$output = preg_replace('/[^a-zA-Z0-9\\/]+/', $this->cleanStringCache['separator'], $output);
}
// Get rid of words that are on the ignore list.
if ($this->cleanStringCache['ignore_words_regex']) {
$words_removed = $this->cleanStringCache['ignore_words_callback']($this->cleanStringCache['ignore_words_regex'], '', $output);
if (Unicode::strlen(trim($words_removed)) > 0) {
$output = $words_removed;
}
}
// Always replace whitespace with the separator.
$output = preg_replace('/\\s+/', $this->cleanStringCache['separator'], $output);
// Trim duplicates and remove trailing and leading separators.
$output = $this->getCleanSeparators($this->getCleanSeparators($output, $this->cleanStringCache['separator']));
// Optionally convert to lower case.
if ($this->cleanStringCache['lowercase']) {
$output = Unicode::strtolower($output);
}
// Shorten to a logical place based on word boundaries.
$output = Unicode::truncate($output, $this->cleanStringCache['maxlength'], TRUE);
// Cache this result in the static array.
$this->cleanStringCache['strings'][$langcode][(string) $string] = $output;
return $output;
}
