/**
* Split a string into a clean array of keywords
* @param $text string
* @param $allowWildcards boolean
* @return array of keywords
*/
static function filterKeywords($text, $allowWildcards = false)
{
$minLength = Config::getVar('search', 'min_word_length');
$stopwords = self::_loadStopwords();
// Join multiple lines into a single string
if (is_array($text)) {
$text = join("\n", $text);
}
$cleanText = Core::cleanVar($text);
// Remove punctuation
$cleanText = PKPString::regexp_replace('/[!"\\#\\$%\'\\(\\)\\.\\?@\\[\\]\\^`\\{\\}~]/', '', $cleanText);
$cleanText = PKPString::regexp_replace('/[\\+,:;&\\/<=>\\|\\\\]/', ' ', $cleanText);
$cleanText = PKPString::regexp_replace('/[\\*]/', $allowWildcards ? '%' : ' ', $cleanText);
$cleanText = PKPString::strtolower($cleanText);
// Split into words
$words = PKPString::regexp_split('/\\s+/', $cleanText);
// FIXME Do not perform further filtering for some fields, e.g., author names?
// Remove stopwords
$keywords = array();
foreach ($words as $k) {
if (!isset($stopwords[$k]) && PKPString::strlen($k) >= $minLength && !is_numeric($k)) {
$keywords[] = PKPString::substr($k, 0, SEARCH_KEYWORD_MAX_LENGTH);
}
}
return $keywords;
}
/**
* Generate a filename for a library file.
* @param $type int LIBRARY_FILE_TYPE_...
* @param $originalFileName string
* @return string
*/
function generateFileName($type, $originalFileName)
{
$libraryFileDao = DAORegistry::getDAO('LibraryFileDAO');
$suffix = $this->getFileSuffixFromType($type);
$ext = $this->getExtension($originalFileName);
$truncated = $this->truncateFileName($originalFileName, 127 - PKPString::strlen($suffix) - 1);
$baseName = PKPString::substr($truncated, 0, PKPString::strpos($originalFileName, $ext) - 1);
// Try a simple syntax first
$fileName = $baseName . '-' . $suffix . '.' . $ext;
if (!$libraryFileDao->filenameExists($this->contextId, $fileName)) {
return $fileName;
}
for ($i = 1;; $i++) {
$fullSuffix = $suffix . '-' . $i;
//truncate more if necessary
$truncated = $this->truncateFileName($originalFileName, 127 - PKPString::strlen($fullSuffix) - 1);
// get the base name and append the suffix
$baseName = PKPString::substr($truncated, 0, PKPString::strpos($originalFileName, $ext) - 1);
//try the following
$fileName = $baseName . '-' . $fullSuffix . '.' . $ext;
if (!$libraryFileDao->filenameExists($this->contextId, $fileName)) {
return $fileName;
}
}
}
/**
* @see FormValidator::isValid()
* Value is valid if it is empty and optional or meets the specified length requirements.
* @return boolean
*/
function isValid()
{
if ($this->isEmptyAndOptional()) {
return true;
} else {
$length = PKPString::strlen($this->getFieldValue());
switch ($this->_comparator) {
case '==':
return $length == $this->_length;
case '!=':
return $length != $this->_length;
case '<':
return $length < $this->_length;
case '>':
return $length > $this->_length;
case '<=':
return $length <= $this->_length;
case '>=':
return $length >= $this->_length;
}
return false;
}
}
/**
* Derive a confidence score calculated as the similarity of the
* original raw citation and the citation text generated from the
* citation description.
* @param $metadataDescription MetadataDescription
* @return integer filter confidence score
*/
function _filterConfidenceScore(&$metadataDescription)
{
// Retrieve the original plain text citation.
$originalCitation = $this->getOriginalRawCitation();
// Generate the formatted citation output from the description.
$citationOutputFilter =& $this->getCitationOutputFilter();
$generatedCitation = $citationOutputFilter->execute($metadataDescription);
// Strip formatting and the Google Scholar tag so that we get a plain
// text string that is comparable with the raw citation.
$generatedCitation = trim(str_replace(GOOGLE_SCHOLAR_TAG, '', strip_tags($generatedCitation)));
// Compare the original to the generated citation.
$citationDiff = PKPString::diff($originalCitation, $generatedCitation);
// Calculate similarity as the number of deleted characters in relation to the
// number of characters in the original citation. This intentionally excludes
// additions as these can represent useful data like a DOI or an external link.
$deletedCharacters = 0;
foreach ($citationDiff as $diffPart) {
// Identify deletions.
if (key($diffPart) == -1) {
$deletedCharacters += PKPString::strlen(current($diffPart));
}
}
$originalCharacters = PKPString::strlen($originalCitation);
$partOfCommonCharacters = ($originalCharacters - $deletedCharacters) / $originalCharacters;
$filterConfidenceScore = (int) round(min($partOfCommonCharacters * 100, 100));
return $filterConfidenceScore;
}
/**
* Converts a string with a single person
* to an NLM name description.
*
* TODO: add initials from all given names to initials
* element
*
* @param $personString string
* @param $title boolean true to parse for title
* @param $degrees boolean true to parse for degrees
* @return MetadataDescription an NLM name description or null
* if the string could not be converted
*/
function &_parsePersonString($personString, $title, $degrees)
{
// Expressions to parse person strings, ported from CiteULike person
// plugin, see http://svn.citeulike.org/svn/plugins/person.tcl
static $personRegex = array('title' => '(?:His (?:Excellency|Honou?r)\\s+|Her (?:Excellency|Honou?r)\\s+|The Right Honou?rable\\s+|The Honou?rable\\s+|Right Honou?rable\\s+|The Rt\\.? Hon\\.?\\s+|The Hon\\.?\\s+|Rt\\.? Hon\\.?\\s+|Mr\\.?\\s+|Ms\\.?\\s+|M\\/s\\.?\\s+|Mrs\\.?\\s+|Miss\\.?\\s+|Dr\\.?\\s+|Sir\\s+|Dame\\s+|Prof\\.?\\s+|Professor\\s+|Doctor\\s+|Mister\\s+|Mme\\.?\\s+|Mast(?:\\.|er)?\\s+|Lord\\s+|Lady\\s+|Madam(?:e)?\\s+|Priv\\.-Doz\\.\\s+)+', 'degrees' => '(,\\s+(?:[A-Z\\.]+))+', 'initials' => '(?:(?:[A-Z]\\.){1,3}[A-Z]\\.?)|(?:(?:[A-Z]\\.\\s){1,3}[A-Z]\\.?)|(?:[A-Z]{1,4})|(?:(?:[A-Z]\\.-?){1,4})|(?:(?:[A-Z]\\.-?){1,3}[A-Z]\\.?)|(?:(?:[A-Z]-){1,3}[A-Z])|(?:(?:[A-Z]\\s){1,3}[A-Z]\\.?)|(?:(?:[A-Z]-){1,3}[A-Z]\\.?)', 'prefix' => 'Dell(?:[a|e])?(?:\\s|$)|Dalle(?:\\s|$)|D[a|e]ll\'(?:\\s|$)|Dela(?:\\s|$)|Del(?:\\s|$)|[Dd]e(?:\\s|$)(?:La(?:\\s|$)|Los(?:\\s|$))?|[Dd]e(?:\\s|$)|[Dd][a|i|u](?:\\s|$)|L[a|e|o](?:\\s|$)|[D|L|O]\'|St\\.?(?:\\s|$)|San(?:\\s|$)|[Dd]en(?:\\s|$)|[Vv]on(?:\\s|$)(?:[Dd]er(?:\\s|$))?|(?:[Ll][ea](?:\\s|$))?[Vv]an(?:\\s|$)(?:[Dd]e(?:n|r)?(?:\\s|$))?', 'givenName' => '(?:[^ \\t\\n\\r\\f\\v,.;()]{2,}|[^ \\t\\n\\r\\f\\v,.;()]{2,}\\-[^ \\t\\n\\r\\f\\v,.;()]{2,})');
// The expressions for given name, suffix and surname are the same
$personRegex['surname'] = $personRegex['suffix'] = $personRegex['givenName'];
$personRegex['double-surname'] = "(?:" . $personRegex['surname'] . "\\s)*" . $personRegex['surname'];
// Shortcut for prefixed surname
$personRegexPrefixedSurname = "(?P<prefix>(?:" . $personRegex['prefix'] . ")?)(?P<surname>" . $personRegex['surname'] . ")";
$personRegexPrefixedDoubleSurname = "(?P<prefix>(?:" . $personRegex['prefix'] . ")?)(?P<surname>" . $personRegex['double-surname'] . ")";
// Instantiate the target person description
$personDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30NameSchema', $this->_assocType);
// Clean the person string
$personString = trim($personString);
// 1. Extract title and degree from the person string and use this as suffix
$suffixString = '';
$results = array();
if ($title && PKPString::regexp_match_get('/^(' . $personRegex['title'] . ')/i', $personString, $results)) {
$suffixString = trim($results[1], ',:; ');
$personString = PKPString::regexp_replace('/^(' . $personRegex['title'] . ')/i', '', $personString);
}
if ($degrees && PKPString::regexp_match_get('/(' . $personRegex['degrees'] . ')$/i', $personString, $results)) {
$degreesArray = explode(',', trim($results[1], ','));
foreach ($degreesArray as $key => $degree) {
$degreesArray[$key] = PKPString::trimPunctuation($degree);
}
$suffixString .= ' - ' . implode('; ', $degreesArray);
$personString = PKPString::regexp_replace('/(' . $personRegex['degrees'] . ')$/i', '', $personString);
}
if (!empty($suffixString)) {
$personDescription->addStatement('suffix', $suffixString);
}
// Space initials when followed by a given name or last name.
$personString = PKPString::regexp_replace('/([A-Z])\\.([A-Z][a-z])/', '\\1. \\2', $personString);
// 2. Extract names and initials from the person string
// The parser expressions are ordered by specificity. The most specific expressions
// come first. Only if these specific expressions don't work will we turn to less
// specific ones. This avoids parsing errors. It also explains why we don't use the
// ?-quantifier for optional elements like initials or middle name where they could
// be misinterpreted.
$personExpressions = array('/^' . $personRegexPrefixedSurname . '$/i', '/^(?P<initials>' . $personRegex['initials'] . ')\\s' . $personRegexPrefixedSurname . '$/', '/^' . $personRegexPrefixedSurname . ',?\\s(?P<initials>' . $personRegex['initials'] . ')$/', '/^' . $personRegexPrefixedDoubleSurname . ',\\s(?P<givenName>' . $personRegex['givenName'] . ')\\s(?P<initials>' . $personRegex['initials'] . ')$/', '/^(?P<givenName>' . $personRegex['givenName'] . ')\\s(?P<initials>' . $personRegex['initials'] . ')\\s' . $personRegexPrefixedSurname . '$/', '/^' . $personRegexPrefixedDoubleSurname . ',\\s(?P<givenName>(?:' . $personRegex['givenName'] . '\\s)+)(?P<initials>' . $personRegex['initials'] . ')$/', '/^(?P<givenName>(?:' . $personRegex['givenName'] . '\\s)+)(?P<initials>' . $personRegex['initials'] . ')\\s' . $personRegexPrefixedSurname . '$/', '/^' . $personRegexPrefixedDoubleSurname . ',(?P<givenName>(?:\\s' . $personRegex['givenName'] . ')+)$/', '/^(?P<givenName>(?:' . $personRegex['givenName'] . '\\s)+)' . $personRegexPrefixedSurname . '$/', '/^\\s*(?P<surname>' . $personRegex['surname'] . ')(?P<suffix>(?:\\s+' . $personRegex['suffix'] . ')?)\\s*,\\s*(?P<initials>(?:' . $personRegex['initials'] . ')?)\\s*\\((?P<givenName>(?:\\s*' . $personRegex['givenName'] . ')+)\\s*\\)\\s*(?P<prefix>(?:' . $personRegex['prefix'] . ')?)$/', '/^(?P<givenName>' . $personRegex['givenName'] . ')\\.(?P<surname>' . $personRegex['double-surname'] . ')$/', '/^(?P<surname>.*)$/');
$results = array();
foreach ($personExpressions as $expressionId => $personExpression) {
if ($nameFound = PKPString::regexp_match_get($personExpression, $personString, $results)) {
// Given names
if (!empty($results['givenName'])) {
// Split given names
$givenNames = explode(' ', trim($results['givenName']));
foreach ($givenNames as $givenName) {
$personDescription->addStatement('given-names', $givenName);
unset($givenName);
}
}
// Initials (will also be saved as given names)
if (!empty($results['initials'])) {
$results['initials'] = str_replace(array('.', '-', ' '), array('', '', ''), $results['initials']);
for ($initialNum = 0; $initialNum < PKPString::strlen($results['initials']); $initialNum++) {
$initial = $results['initials'][$initialNum];
$personDescription->addStatement('given-names', $initial);
unset($initial);
}
}
// Surname
if (!empty($results['surname'])) {
// Correct all-upper surname
if (strtoupper($results['surname']) == $results['surname']) {
$results['surname'] = ucwords(strtolower($results['surname']));
}
$personDescription->addStatement('surname', $results['surname']);
}
// Prefix/Suffix
foreach (array('prefix', 'suffix') as $propertyName) {
if (!empty($results[$propertyName])) {
$results[$propertyName] = trim($results[$propertyName]);
$personDescription->addStatement($propertyName, $results[$propertyName]);
}
}
break;
}
}
return $personDescription;
}
/**
* Checks whether the given string is an ISBN.
* @param $isbn
* @return boolean
*/
function isValidIsbn($isbn)
{
return is_string($isbn) && is_numeric($isbn) && PKPString::strlen($isbn) == 13;
}
/**
* Construct an array of search strings from a citation
* description and an array of search templates.
* The templates may contain the placeholders
* %aulast%: the first author's surname
* %au%: the first author full name
* %title%: the article-title (if it exists),
* otherwise the source
* %date%: the publication year
* %isbn%: ISBN
* @param $searchTemplates an array of templates
* @param $citationDescription MetadataDescription
* @return array
*/
function constructSearchStrings(&$searchTemplates, &$citationDescription)
{
// Convert first authors' name description to a string
import('lib.pkp.plugins.metadata.nlm30.filter.Nlm30NameSchemaPersonStringFilter');
$personStringFilter = new Nlm30NameSchemaPersonStringFilter();
// Retrieve the authors
$firstAuthorSurname = $firstAuthor = '';
$authors = $citationDescription->getStatement('person-group[@person-group-type="author"]');
if (is_array($authors) && count($authors)) {
$firstAuthorSurname = (string) $authors[0]->getStatement('surname');
$firstAuthor = $personStringFilter->execute($authors[0]);
}
// Retrieve the editors
$firstEditorSurname = $firstEditor = '';
$editors = $citationDescription->getStatement('person-group[@person-group-type="editor"]');
if (is_array($editors) && count($editors)) {
$firstEditorSurname = (string) $editors[0]->getStatement('surname');
$firstEditor = $personStringFilter->execute($editors[0]);
}
// Retrieve (default language) title
$title = (string) ($citationDescription->hasStatement('article-title') ? $citationDescription->getStatement('article-title') : $citationDescription->getStatement('source'));
// Extract the year from the publication date
$year = (string) $citationDescription->getStatement('date');
$year = PKPString::strlen($year) > 4 ? PKPString::substr($year, 0, 4) : $year;
// Retrieve ISBN
$isbn = (string) $citationDescription->getStatement('isbn');
// Replace the placeholders in the templates
$searchStrings = array();
foreach ($searchTemplates as $searchTemplate) {
// Try editors and authors separately
$searchStrings[] = str_replace(array('%aulast%', '%au%', '%title%', '%date%', '%isbn%'), array($firstAuthorSurname, $firstAuthor, $title, $year, $isbn), $searchTemplate);
$searchStrings[] = str_replace(array('%aulast%', '%au%', '%title%', '%date%', '%isbn%'), array($firstEditorSurname, $firstEditor, $title, $year, $isbn), $searchTemplate);
}
// Remove empty or duplicate searches
$searchStrings = array_map(array('PKPString', 'trimPunctuation'), $searchStrings);
$searchStrings = array_unique($searchStrings);
$searchStrings = arrayClean($searchStrings);
return $searchStrings;
}
/**
* Transform a single NLM name description to a person string.
* NB: We use the style: surname suffix, initials (first-name) prefix
* which is relatively easy to parse back.
* @param $personDescription MetadataDescription|'et-al'
* @return string
*/
function _flattenPersonDescription(&$personDescription)
{
// Handle et-al
if (is_string($personDescription) && $personDescription == PERSON_STRING_FILTER_ETAL) {
return 'et al';
}
$nameVars['%surname%'] = (string) $personDescription->getStatement('surname');
$givenNames = $personDescription->getStatement('given-names');
$nameVars['%firstname%'] = $nameVars['%initials%'] = '';
if (is_array($givenNames) && count($givenNames)) {
if (PKPString::strlen($givenNames[0]) > 1) {
$nameVars['%firstname%'] = array_shift($givenNames);
}
foreach ($givenNames as $givenName) {
$nameVars['%initials%'] .= PKPString::substr($givenName, 0, 1) . '.';
}
}
if (!empty($nameVars['%initials%'])) {
$nameVars['%initials%'] = ' ' . $nameVars['%initials%'];
}
$nameVars['%prefix%'] = (string) $personDescription->getStatement('prefix');
if (!empty($nameVars['%prefix%'])) {
$nameVars['%prefix%'] = ' ' . $nameVars['%prefix%'];
}
$nameVars['%suffix%'] = (string) $personDescription->getStatement('suffix');
if (!empty($nameVars['%suffix%'])) {
$nameVars['%suffix%'] = ' ' . $nameVars['%suffix%'];
}
// Fill placeholders in person template.
$personString = str_replace(array_keys($nameVars), array_values($nameVars), $this->getTemplate());
// Remove empty brackets and trailing/leading whitespace
$personString = trim(str_replace('()', '', $personString));
return $personString;
}
/**
* @function abntDateFormatWithDay Format date taking in consideration ABNT month abbreviations
* @param $string string
* @return string
*/
function abntDateFormatWithDay($string)
{
if (is_numeric($string)) {
// it is a numeric string, we handle it as timestamp
$timestamp = (int) $string;
} else {
$timestamp = strtotime($string);
}
$format = "%d %B %Y";
if (PKPString::strlen(strftime("%B", $timestamp)) > 4) {
$format = "%d %b. %Y";
}
return PKPString::strtolower(strftime($format, $timestamp));
}
/**
* Fills the given citation object with
* meta-data retrieved from PubMed.
* @param $pmid string
* @return MetadataDescription
*/
function &_lookup($pmid)
{
$nullVar = null;
// Use eFetch to get XML metadata for the given PMID
$lookupParams = array('db' => 'pubmed', 'mode' => 'xml', 'tool' => 'pkp-wal', 'id' => $pmid);
if (!is_null($this->getEmail())) {
$lookupParams['email'] = $this->getEmail();
}
// Call the eFetch URL and get an XML result
if (is_null($resultDOM = $this->callWebService(PUBMED_WEBSERVICE_EFETCH, $lookupParams))) {
return $nullVar;
}
$articleTitleNodes = $resultDOM->getElementsByTagName('ArticleTitle');
$articleTitleFirstNode = $articleTitleNodes->item(0);
$medlineTaNodes = $resultDOM->getElementsByTagName('MedlineTA');
$medlineTaFirstNode = $medlineTaNodes->item(0);
$metadata = array('pub-id[@pub-id-type="pmid"]' => $pmid, 'article-title' => $articleTitleFirstNode->textContent, 'source' => $medlineTaFirstNode->textContent);
$volumeNodes = $resultDOM->getElementsByTagName('Volume');
$issueNodes = $resultDOM->getElementsByTagName('Issue');
if ($volumeNodes->length > 0) {
$volumeFirstNode = $volumeNodes->item(0);
}
$metadata['volume'] = $volumeFirstNode->textContent;
if ($issueNodes->length > 0) {
$issueFirstNode = $issueNodes->item(0);
}
$metadata['issue'] = $issueFirstNode->textContent;
// Get list of author full names
foreach ($resultDOM->getElementsByTagName("Author") as $authorNode) {
if (!isset($metadata['person-group[@person-group-type="author"]'])) {
$metadata['person-group[@person-group-type="author"]'] = array();
}
// Instantiate an NLM name description
$authorDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30NameSchema', ASSOC_TYPE_AUTHOR);
// Surname
$lastNameNodes = $authorNode->getElementsByTagName('LastName');
$lastNameFirstNode = $lastNameNodes->item(0);
$authorDescription->addStatement('surname', $lastNameFirstNode->textContent);
// Given names
$givenNamesString = '';
$firstNameNodes = $authorNode->getElementsByTagName('FirstName');
if ($firstNameNodes->length > 0) {
$firstNameFirstNode = $firstNameNodes->item(0);
$givenNamesString = $firstNameFirstNode->textContent;
} else {
$foreNameNodes = $authorNode->getElementsByTagName('ForeName');
if ($foreNameNodes->length > 0) {
$foreNameFirstNode = $foreNameNodes->item(0);
$givenNamesString = $foreNameFirstNode->textContent;
}
}
if (!empty($givenNamesString)) {
foreach (explode(' ', $givenNamesString) as $givenName) {
$authorDescription->addStatement('given-names', PKPString::trimPunctuation($givenName));
}
}
// Suffix
$suffixNodes = $authorNode->getElementsByTagName('Suffix');
if ($suffixNodes->length > 0) {
$suffixFirstNode = $suffixNodes->item(0);
$authorDescription->addStatement('suffix', $suffixFirstNode->textContent);
}
// Include collective names
// FIXME: This corresponds to an NLM-citation <collab> tag and should be part of the Metadata implementation
/*if ($resultDOM->getElementsByTagName("CollectiveName")->length > 0 && $authorNode->getElementsByTagName("CollectiveName")->item(0)->textContent != '') {
}*/
$metadata['person-group[@person-group-type="author"]'][] =& $authorDescription;
unset($authorDescription);
}
// Extract pagination
$medlinePgnNodes = $resultDOM->getElementsByTagName('MedlinePgn');
$medlinePgnFirstNode = $medlinePgnNodes->item(0);
if (PKPString::regexp_match_get("/^[:p\\.\\s]*(?P<fpage>[Ee]?\\d+)(-(?P<lpage>\\d+))?/", $medlinePgnFirstNode->textContent, $pages)) {
$fPage = (int) $pages['fpage'];
$metadata['fpage'] = $fPage;
if (!empty($pages['lpage'])) {
$lPage = (int) $pages['lpage'];
// Deal with shortcuts like '382-7'
if ($lPage < $fPage) {
$lPage = (int) (PKPString::substr($pages['fpage'], 0, -PKPString::strlen($pages['lpage'])) . $pages['lpage']);
}
$metadata['lpage'] = $lPage;
}
}
// Get publication date (can be in several places in PubMed).
$dateNode = null;
$articleDateNodes = $resultDOM->getElementsByTagName('ArticleDate');
if ($articleDateNodes->length > 0) {
$dateNode = $articleDateNodes->item(0);
} else {
$pubDateNodes = $resultDOM->getElementsByTagName('PubDate');
if ($pubDateNodes->length > 0) {
$dateNode = $pubDateNodes->item(0);
}
}
// Retrieve the data parts and assemble date.
if (!is_null($dateNode)) {
$publicationDate = '';
$requiresNormalization = false;
foreach (array('Year' => 4, 'Month' => 2, 'Day' => 2) as $dateElement => $padding) {
$dateElementNodes = $dateNode->getElementsByTagName($dateElement);
if ($dateElementNodes->length > 0) {
if (!empty($publicationDate)) {
$publicationDate .= '-';
}
$dateElementFirstNode = $dateElementNodes->item(0);
$datePart = str_pad($dateElementFirstNode->textContent, $padding, '0', STR_PAD_LEFT);
if (!is_numeric($datePart)) {
$requiresNormalization = true;
}
$publicationDate .= $datePart;
} else {
break;
}
}
// Normalize the date to NLM standard if necessary.
if ($requiresNormalization) {
$dateFilter = new DateStringNormalizerFilter();
$publicationDate = $dateFilter->execute($publicationDate);
}
if (!empty($publicationDate)) {
$metadata['date'] = $publicationDate;
}
}
// Get publication type
$publicationTypeNodes = $resultDOM->getElementsByTagName('PublicationType');
if ($publicationTypeNodes->length > 0) {
foreach ($publicationTypeNodes as $publicationType) {
// The vast majority of items on PubMed are articles so catch these...
if (PKPString::strpos(PKPString::strtolower($publicationType->textContent), 'article') !== false) {
$metadata['[@publication-type]'] = NLM30_PUBLICATION_TYPE_JOURNAL;
break;
}
}
}
// Get DOI if it exists
$articleIdNodes = $resultDOM->getElementsByTagName('ArticleId');
foreach ($articleIdNodes as $idNode) {
if ($idNode->getAttribute('IdType') == 'doi') {
$metadata['pub-id[@pub-id-type="doi"]'] = $idNode->textContent;
}
}
// Use eLink utility to find fulltext links
$lookupParams = array('dbfrom' => 'pubmed', 'cmd' => 'llinks', 'tool' => 'pkp-wal', 'id' => $pmid);
if (!is_null($resultDOM = $this->callWebService(PUBMED_WEBSERVICE_ELINK, $lookupParams))) {
// Get a list of possible links
foreach ($resultDOM->getElementsByTagName("ObjUrl") as $linkOut) {
$attributes = '';
foreach ($linkOut->getElementsByTagName("Attribute") as $attribute) {
$attributes .= PKPString::strtolower($attribute->textContent) . ' / ';
}
// Only add links to open access resources
if (PKPString::strpos($attributes, "subscription") === false && PKPString::strpos($attributes, "membership") === false && PKPString::strpos($attributes, "fee") === false && $attributes != "") {
$urlNodes = $linkOut->getElementsByTagName('Url');
$urlFirstNode = $urlNodes->item(0);
$links[] = $urlFirstNode->textContent;
}
}
// Take the first link if we have any left (presumably pubmed returns them in preferential order)
if (isset($links[0])) {
$metadata['uri'] = $links[0];
}
}
return $this->getNlm30CitationDescriptionFromMetadataArray($metadata);
}
/**
* Retrieve auto-suggestions from the faceting service.
* @param $url string
* @param $searchRequest SolrSearchRequest
* @param $userInput string
* @param $fieldName string
* @return array The generated suggestions.
*/
function _getFacetingAutosuggestions($url, $searchRequest, $userInput, $fieldName)
{
// Remove special characters from the user input.
$searchTerms = strtr($userInput, '"()+-|&!', ' ');
// Cut off the last search term.
$searchTerms = explode(' ', $searchTerms);
$facetPrefix = array_pop($searchTerms);
if (empty($facetPrefix)) {
return array();
}
// Use the remaining search query to pre-filter
// facet results. This may be an invalid query
// but edismax will deal gracefully with syntax
// errors.
$userInput = PKPString::substr($userInput, 0, -PKPString::strlen($facetPrefix));
switch ($fieldName) {
case 'query':
// The 'query' filter goes against all fields.
$articleSearch = new ArticleSearch();
$solrFields = array_values($articleSearch->getIndexFieldMap());
break;
case 'indexTerms':
// The 'index terms' filter goes against keyword index fields.
$solrFields = array('discipline', 'subject', 'type', 'coverage');
break;
default:
// All other filters can be used directly.
$solrFields = array($fieldName);
}
$solrFieldString = implode('|', $solrFields);
$searchRequest->addQueryFieldPhrase($solrFieldString, $userInput);
// Construct the main query.
$params = $this->_getSearchQueryParameters($searchRequest);
if (!isset($params['q'])) {
// Use a catch-all query in case we have no limiting
// search.
$params['q'] = '*:*';
}
if ($fieldName == 'query') {
$params['facet.field'] = 'default_spell';
} else {
$params['facet.field'] = $fieldName . '_spell';
}
$facetPrefixLc = PKPString::strtolower($facetPrefix);
$params['facet.prefix'] = $facetPrefixLc;
// Make the request.
$response = $this->_makeRequest($url, $params);
if (!is_a($response, 'DOMXPath')) {
return array();
}
// Extract term suggestions.
$nodeList = $response->query('//lst[@name="facet_fields"]/lst/int/@name');
if ($nodeList->length == 0) {
return array();
}
$termSuggestions = array();
foreach ($nodeList as $childNode) {
$termSuggestions[] = $childNode->value;
}
// Add the term suggestion to the remaining user input.
$suggestions = array();
foreach ($termSuggestions as $termSuggestion) {
// Restore case if possible.
if (strpos($termSuggestion, $facetPrefixLc) === 0) {
$termSuggestion = $facetPrefix . PKPString::substr($termSuggestion, PKPString::strlen($facetPrefix));
}
$suggestions[] = $userInput . $termSuggestion;
}
return $suggestions;
}
/**
* Parse an XML file using the specified handler.
* If no handler has been specified, XMLParserDOMHandler is used by default, returning a tree structure representing the document.
* @param $file string full path to the XML file
* @param $dataCallback mixed Optional callback for data handling: function dataCallback($operation, $wrapper, $data = null)
* @return object actual return type depends on the handler
*/
function &parse($file, $dataCallback = null)
{
$parser =& $this->createParser();
if (!isset($this->handler)) {
// Use default handler for parsing
$handler = new XMLParserDOMHandler();
$this->setHandler($handler);
}
xml_set_object($parser, $this->handler);
xml_set_element_handler($parser, "startElement", "endElement");
xml_set_character_data_handler($parser, "characterData");
import('lib.pkp.classes.file.FileWrapper');
$wrapper =& FileWrapper::wrapper($file);
// Handle responses of various types
while (true) {
$newWrapper = $wrapper->open();
if (is_object($newWrapper)) {
// Follow a redirect
unset($wrapper);
$wrapper =& $newWrapper;
unset($newWrapper);
} elseif (!$newWrapper) {
// Could not open resource -- error
$returner = false;
return $returner;
} else {
// OK, we've found the end result
break;
}
}
if (!$wrapper) {
$result = false;
return $result;
}
if ($dataCallback) {
call_user_func($dataCallback, 'open', $wrapper);
}
while (!$wrapper->eof() && ($data = $wrapper->read()) !== false) {
// if the string contains non-UTF8 characters, convert it to UTF-8 for parsing
if (Config::getVar('i18n', 'charset_normalization') == 'On' && !PKPString::utf8_compliant($data)) {
$utf8_last = PKPString::substr($data, PKPString::strlen($data) - 1);
// if the string ends in a "bad" UTF-8 character, maybe it's truncated
while (!$wrapper->eof() && PKPString::utf8_bad_find($utf8_last) === 0) {
// read another chunk of data
$data .= $wrapper->read();
$utf8_last = PKPString::substr($data, PKPString::strlen($data) - 1);
}
$data = PKPString::utf8_normalize($data);
// strip any invalid UTF-8 sequences
$data = PKPString::utf8_bad_strip($data);
// convert named entities to numeric entities
$data = strtr($data, PKPString::getHTMLEntities());
}
// strip any invalid ASCII control characters
$data = PKPString::utf8_strip_ascii_ctrl($data);
if ($dataCallback) {
call_user_func($dataCallback, 'parse', $wrapper, $data);
}
if (!xml_parse($parser, $data, $wrapper->eof())) {
$this->addError(xml_error_string(xml_get_error_code($parser)));
}
}
if ($dataCallback) {
call_user_func($dataCallback, 'close', $wrapper);
}
$wrapper->close();
$result = $this->handler->getResult();
$this->destroyParser($parser);
if (isset($handler)) {
$handler->destroy();
}
return $result;
}
/**
* Truncate a filename to fit in the specified length.
*/
function truncateFileName($fileName, $length = 127)
{
if (PKPString::strlen($fileName) <= $length) {
return $fileName;
}
$ext = $this->getExtension($fileName);
$truncated = PKPString::substr($fileName, 0, $length - 1 - PKPString::strlen($ext)) . '.' . $ext;
return PKPString::substr($truncated, 0, $length);
}