Iterate over an array of delimiters and see whether
it exists in the given input string. If so, then use
it to explode the string into an array.
| static public iterativeExplode ( $delimiters, $input ) : array | ||
| $delimiters | array | |
| $input | string | |
| return | array | |
/**
* Converts a string with multiple persons
* to an array of NLM name descriptions.
*
* @param $personsString string
* @param $title boolean true to parse for title
* @param $degrees boolean true to parse for degrees
* @return array an array of NLM name descriptions or null
* if the string could not be converted plus optionally a
* single 'et-al' string.
*/
function &_parsePersonsString($personsString, $title, $degrees)
{
// Check for 'et al'.
$personsStringBeforeEtal = PKPString::strlen($personsString);
$personsString = PKPString::regexp_replace('/et ?al$/', '', $personsString);
$etAl = $personsStringBeforeEtal == PKPString::strlen($personsString) ? false : true;
// Remove punctuation.
$personsString = trim($personsString, ':;, ');
// Cut the authors string into pieces.
$personStrings = PKPString::iterativeExplode(array(':', ';'), $personsString);
// If we did not have success with simple patterns then try more complex
// patterns to tokenize multiple-person strings.
if (count($personStrings) == 1) {
// The first pattern must match the whole string, the second is used
// to extract names.
$complexPersonsPatterns = array(array('/^((([^ \\t\\n\\r\\f\\v,.&]{2,}\\s*)+,\\s*([A-Z]\\.\\s*)+),\\s*)+(\\&|\\.\\s\\.\\s\\.)\\s*([^ \\t\\n\\r\\f\\v,.&]{2,}\\s*,\\s*([A-Z]\\.\\s*)+)$/i', '/(?:[^ \\t\\n\\r\\f\\v,.&]{2,}\\s*)+,\\s*(?:[A-Z]\\.\\s*)+/i'), array('/^((([^ \\t\\n\\r\\f\\v,&]+\\s+)+[^ \\t\\n\\r\\f\\v,&]+\\s*)[,&]\\s*)+(([^ \\t\\n\\r\\f\\v,&]+\\s+)+[^ \\t\\n\\r\\f\\v,&]+)/i', '/(?:(?:[^ \\t\\n\\r\\f\\v,&.]+|[^ \\t\\n\\r\\f\\v,&]{2,})\\s+)+(?:[^ \\t\\n\\r\\f\\v,&.]+|[^ \\t\\n\\r\\f\\v,&]{2,})/i'));
$matched = false;
foreach ($complexPersonsPatterns as $complexPersonsPattern) {
// Break at the first pattern that matches.
if ($matched = PKPString::regexp_match($complexPersonsPattern[0], $personsString)) {
// Retrieve names.
$success = PKPString::regexp_match_all($complexPersonsPattern[1], $personsString, $personStrings);
assert($success && count($personStrings) == 1);
$personStrings = $personStrings[0];
break;
}
}
if (!$matched) {
// If nothing matches then try to parse as a single person.
$personStrings = array($personsString);
}
}
// Parse persons.
$persons = array();
foreach ($personStrings as $personString) {
$persons[] =& $this->_parsePersonString($personString, $title, $degrees);
}
// Add et-al string.
if ($etAl) {
$persons[] = PERSON_STRING_FILTER_ETAL;
}
return $persons;
}
/**
* @copydoc Filter::process()
* @param $input string
* @return MetadataDescription
*/
function &process(&$input)
{
$citationString =& $input;
$nullVar = null;
// Check the availability of perl
$perlCommand = Config::getVar('cli', 'perl');
if (empty($perlCommand) || !file_exists($perlCommand)) {
return $nullVar;
}
// Convert to ASCII - Paracite doesn't handle UTF-8 well
$citationString = PKPString::utf8_to_ascii($citationString);
// Call the paracite parser
$wrapperScript = dirname(__FILE__) . DIRECTORY_SEPARATOR . 'paracite.pl';
$paraciteCommand = $perlCommand . ' ' . escapeshellarg($wrapperScript) . ' ' . $this->getCitationModule() . ' ' . escapeshellarg($citationString);
$xmlResult = shell_exec($paraciteCommand);
if (empty($xmlResult)) {
return $nullVar;
}
if (Config::getVar('i18n', 'charset_normalization') == 'On' && !PKPString::utf8_compliant($xmlResult)) {
$xmlResult = PKPString::utf8_normalize($xmlResult);
}
// Create a temporary DOM document
$resultDOM = new DOMDocument();
$resultDOM->recover = true;
$resultDOM->loadXML($xmlResult);
// Extract the parser results as an array
$xmlHelper = new XMLHelper();
$metadata = $xmlHelper->xmlToArray($resultDOM->documentElement);
// We have to merge subtitle and title as neither OpenURL
// nor NLM can handle subtitles.
if (isset($metadata['subtitle'])) {
$metadata['title'] .= '. ' . $metadata['subtitle'];
unset($metadata['subtitle']);
}
// Break up the authors field
if (isset($metadata['authors'])) {
$metadata['authors'] = PKPString::trimPunctuation($metadata['authors']);
$metadata['authors'] = PKPString::iterativeExplode(array(':', ';'), $metadata['authors']);
}
// Convert pages to integers
foreach (array('spage', 'epage') as $pageProperty) {
if (isset($metadata[$pageProperty])) {
$metadata[$pageProperty] = (int) $metadata[$pageProperty];
}
}
// Convert titles to title case
foreach (array('title', 'chapter', 'publication') as $titleProperty) {
if (isset($metadata[$titleProperty])) {
$metadata[$titleProperty] = PKPString::titleCase($metadata[$titleProperty]);
}
}
// Map ParaCite results to OpenURL - null means
// throw the value away.
$metadataMapping = array('genre' => 'genre', '_class' => null, 'any' => null, 'authors' => 'au', 'aufirst' => 'aufirst', 'aufull' => null, 'auinit' => 'auinit', 'aulast' => 'aulast', 'atitle' => 'atitle', 'cappublication' => null, 'captitle' => null, 'date' => 'date', 'epage' => 'epage', 'featureID' => null, 'id' => null, 'issue' => 'issue', 'jnl_epos' => null, 'jnl_spos' => null, 'match' => null, 'marked' => null, 'num_of_fig' => null, 'pages' => 'pages', 'publisher' => 'pub', 'publoc' => 'place', 'ref' => null, 'rest_text' => null, 'spage' => 'spage', 'targetURL' => 'url', 'text' => null, 'ucpublication' => null, 'uctitle' => null, 'volume' => 'volume', 'year' => 'date');
// Ignore 'year' if 'date' is set
if (isset($metadata['date'])) {
$metadataMapping['year'] = null;
}
// Set default genre
if (empty($metadata['genre'])) {
$metadata['genre'] = OPENURL10_GENRE_ARTICLE;
}
// Handle title, chapter and publication depending on
// the (inferred) genre. Also instantiate the target schema.
switch ($metadata['genre']) {
case OPENURL10_GENRE_BOOK:
case OPENURL10_GENRE_BOOKITEM:
case OPENURL10_GENRE_REPORT:
case OPENURL10_GENRE_DOCUMENT:
$metadataMapping += array('publication' => 'btitle', 'chapter' => 'atitle');
if (isset($metadata['title'])) {
if (!isset($metadata['publication'])) {
$metadata['publication'] = $metadata['title'];
} elseif (!isset($metadata['chapter'])) {
$metadata['chapter'] = $metadata['title'];
}
unset($metadata['title']);
}
$openurl10SchemaName = 'lib.pkp.plugins.metadata.openurl10.schema.Openurl10BookSchema';
$openurl10SchemaClass = 'Openurl10BookSchema';
break;
case OPENURL10_GENRE_ARTICLE:
case OPENURL10_GENRE_JOURNAL:
case OPENURL10_GENRE_ISSUE:
case OPENURL10_GENRE_CONFERENCE:
case OPENURL10_GENRE_PROCEEDING:
case OPENURL10_GENRE_PREPRINT:
default:
$metadataMapping += array('publication' => 'jtitle');
if (isset($metadata['title'])) {
if (!isset($metadata['publication'])) {
$metadata['publication'] = $metadata['title'];
} elseif (!isset($metadata['atitle'])) {
$metadata['atitle'] = $metadata['title'];
}
unset($metadata['title']);
}
$openurl10SchemaName = 'lib.pkp.plugins.metadata.openurl10.schema.Openurl10JournalSchema';
$openurl10SchemaClass = 'Openurl10JournalSchema';
break;
}
// Instantiate an OpenURL description
$openurl10Description = new MetadataDescription($openurl10SchemaName, ASSOC_TYPE_CITATION);
$openurl10Schema = new $openurl10SchemaClass();
// Map the ParaCite result to OpenURL
foreach ($metadata as $paraciteElementName => $paraciteValue) {
if (!empty($paraciteValue)) {
// Trim punctuation
if (is_string($paraciteValue)) {
$paraciteValue = PKPString::trimPunctuation($paraciteValue);
}
// Transfer the value to the OpenURL result array
assert(array_key_exists($paraciteElementName, $metadataMapping));
$openurl10PropertyName = $metadataMapping[$paraciteElementName];
if (!is_null($openurl10PropertyName) && $openurl10Schema->hasProperty($openurl10PropertyName)) {
if (is_array($paraciteValue)) {
foreach ($paraciteValue as $singleValue) {
$success = $openurl10Description->addStatement($openurl10PropertyName, $singleValue);
assert($success);
}
} else {
$success = $openurl10Description->addStatement($openurl10PropertyName, $paraciteValue);
assert($success);
}
}
}
}
// Crosswalk to NLM
$crosswalkFilter = new Openurl10Nlm30CitationSchemaCrosswalkFilter();
$nlm30Description =& $crosswalkFilter->execute($openurl10Description);
assert(is_a($nlm30Description, 'MetadataDescription'));
// Add 'rest_text' as NLM comment (if given)
if (isset($metadata['rest_text'])) {
$nlm30Description->addStatement('comment', PKPString::trimPunctuation($metadata['rest_text']));
}
// Set display name and sequence id in the meta-data description
// to the corresponding values from the filter. This is important
// so that we later know which result came from which filter.
$nlm30Description->setDisplayName($this->getDisplayName());
$nlm30Description->setSequence($this->getSequence());
return $nlm30Description;
}
