/**
* Creates a test description in Openurl10 format
* @return MetadataDescription
*/
protected function getTestOpenurl10Description()
{
$citationData = array('aulast' => 'von Surname1', 'aufirst' => 'Given1 P', 'auinit1' => 'G', 'auinitm' => 'P', 'auinit' => 'GP', 'ausuffix' => 'suff', 'au' => array(0 => 'Surname1 suff, P. (Given1) von', 1 => 'Surname2, (Given2)'), 'genre' => 'article', 'jtitle' => 'Some Journal Title', 'atitle' => 'Some Article Title', 'date' => '2005-07-03', 'issn' => '0694760949645', 'spage' => 17, 'epage' => 33, 'volume' => '7', 'issue' => '5', 'eissn' => '3049674960475', 'artnum' => '45', 'coden' => 'coden', 'sici' => 'sici');
$openurl10Description = new MetadataDescription('lib.pkp.plugins.metadata.openurl10.schema.Openurl10JournalSchema', ASSOC_TYPE_CITATION);
self::assertTrue($openurl10Description->setStatements($citationData));
return $openurl10Description;
}
/**
* This implementation of the CrosswalkFilter
* simply removes statements from the incoming meta-data
* description that are not in the target description's schema.
* @see Filter::process()
* @param $input MetadataDescription
* @return MetadataDescription
*/
function &process(&$input)
{
// Create the target description
$output = new MetadataDescription($this->_toSchema);
// Compare the property names of the incoming description with
// the property names allowed in the target schema.
$sourceProperties = $input->getSetPropertyNames();
$targetProperties = $output->getPropertyNames();
$propertiesToBeRemoved = array_diff($sourceProperties, $targetProperties);
// Remove statements for properties that are not in the target schema.
$statements =& $input->getStatements();
foreach ($propertiesToBeRemoved as $propertyToBeRemoved) {
assert(isset($statements[$propertyToBeRemoved]));
unset($statements[$propertyToBeRemoved]);
}
// Set the remaining statements in the target description
$success = $output->setStatements($statements);
assert($success);
return $output;
}
/**
* @copydoc MetadataDataObjectAdapter::extractMetadataFromDataObject()
* @param $dataObject Citation
* @return MetadataDescription
*/
function extractMetadataFromDataObject(&$dataObject)
{
$metadataDescription = $this->instantiateMetadataDescription();
// Establish the association between the meta-data description
// and the citation object.
$metadataDescription->setAssocId($dataObject->getId());
// Identify the length of the name space prefix
$namespacePrefixLength = strlen($this->getMetadataNamespace()) + 1;
// Get all meta-data field names
$fieldNames = array_merge($this->getDataObjectMetadataFieldNames(false), $this->getDataObjectMetadataFieldNames(true));
// Retrieve the statements from the data object
$statements = array();
foreach ($fieldNames as $fieldName) {
if ($dataObject->hasData($fieldName)) {
// Remove the name space prefix
$propertyName = substr($fieldName, $namespacePrefixLength);
if (in_array($propertyName, array('person-group[@person-group-type="author"]', 'person-group[@person-group-type="editor"]'))) {
// Retrieve the names array (must not be by-ref
// to protect the original citation object!)
$names = $dataObject->getData($fieldName);
// Convert key/value arrays to MetadataDescription objects.
foreach ($names as $key => $name) {
if (is_array($name)) {
// Construct a meta-data description from
// this name array.
switch ($propertyName) {
case 'person-group[@person-group-type="author"]':
$assocType = ASSOC_TYPE_AUTHOR;
break;
case 'person-group[@person-group-type="editor"]':
$assocType = ASSOC_TYPE_EDITOR;
break;
}
$nameDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30NameSchema', $assocType);
$nameDescription->setStatements($name);
$names[$key] =& $nameDescription;
unset($nameDescription);
} else {
// The only non-structured data allowed here
// is the et-al string.
import('lib.pkp.plugins.metadata.nlm30.filter.Nlm30PersonStringFilter');
assert($name == PERSON_STRING_FILTER_ETAL);
}
}
$statements[$propertyName] =& $names;
unset($names);
} else {
$statements[$propertyName] =& $dataObject->getData($fieldName);
}
}
}
// Set the statements in the meta-data description
$success = $metadataDescription->setStatements($statements);
assert($success);
return $metadataDescription;
}
/**
* Take an array of citation parse/lookup results and derive a citation
* with one "best" set of values.
*
* We determine the best values within the citations that have a score above
* the given threshold. Citations with a score below the threshold will be
* ignored.
*
* For these citations we count the frequency of values per meta-data property.
* The most frequent value will be chosen as "best" value.
*
* If two values have the same frequency then decide based on the score. If
* this is still ambivalent then return the first of the remaining values.
*
* @param $scoredCitations
* @param $scoreThreshold integer a number between 0 (=no threshold) and 100
* @return Citation one citation with the "best" values set
*/
function &_guessValues(&$scoredCitations, $scoreThreshold)
{
assert($scoreThreshold >= 0 && $scoreThreshold <= 100);
// Create the target citation description.
$targetDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30CitationSchema', ASSOC_TYPE_CITATION);
// Step 1: List all values and max scores that have been identified for a given element
// but only include values from results above a given scoring threshold
// Initialize variables for the first step.
$valuesByPropertyName = array();
$maxScoresByPropertyNameAndValue = array();
// Sort the scored citations by score with the highest score first.
krsort($scoredCitations);
foreach ($scoredCitations as $currentScore => $citationsForCurrentScore) {
// Check whether the current score is below the threshold, if so
// stop the loop. We've sorted our citations by score so the remaining
// citations all have scores below the threshold and we can forget
// about them.
if ($currentScore < $scoreThreshold) {
break;
}
foreach ($citationsForCurrentScore as $citationForCurrentScore) {
$statements = $citationForCurrentScore->getStatements();
// Add the property values and scores of this citation
// to the overall property lists
foreach ($statements as $propertyName => $value) {
// Initialize sub-arrays if necessary
if (!isset($valuesByPropertyName[$propertyName])) {
$valuesByPropertyName[$propertyName] = array();
}
if (!isset($maxScoresByPropertyNameAndValue[$propertyName])) {
$maxScoresByPropertyNameAndValue[$propertyName] = array();
}
// Add the value for the given property, as we want to count
// value frequencies later, we explicitly allow duplicates.
$serializedValue = serialize($value);
$valuesByPropertyName[$propertyName][] = $serializedValue;
// As we have ordered our citations descending by score, the
// first score found for a value is also the maximum score.
if (!isset($maxScoresByPropertyNameAndValue[$propertyName][$serializedValue])) {
$maxScoresByPropertyNameAndValue[$propertyName][$serializedValue] = $currentScore;
}
}
}
}
// Step 2: Find out the values that occur most frequently for each element
// and order these by score.
foreach ($valuesByPropertyName as $propertyName => $values) {
// Count the occurrences of each value within the given element
$valueFrequencies = array_count_values($values);
// Order the most frequent values to the beginning of the array
arsort($valueFrequencies);
// Get the most frequent values (may be several if there are more than one
// with the same frequency).
$scoresOfMostFrequentValues = array();
$previousValueFrequency = 0;
foreach ($valueFrequencies as $value => $valueFrequency) {
// Only extract the most frequent values, jump out of the
// loop when less frequent values start.
if ($previousValueFrequency > $valueFrequency) {
break;
}
$previousValueFrequency = $valueFrequency;
$scoresOfMostFrequentValues[$value] = $maxScoresByPropertyNameAndValue[$propertyName][$value];
}
// Now we can order the most frequent values by score, starting
// with the highest score.
arsort($scoresOfMostFrequentValues);
// Now get the first key which represents the value with the
// highest frequency and the highest score.
reset($scoresOfMostFrequentValues);
$bestValue = unserialize(key($scoresOfMostFrequentValues));
// Set the found "best" element value in the result citation.
$statements = array($propertyName => $bestValue);
$success = $targetDescription->setStatements($statements);
assert($success);
}
// Instantiate the target citation
$targetCitation = new Citation();
$targetCitation->injectMetadata($targetDescription);
return $targetCitation;
}
/**
* Creates a new NLM citation description and adds the data
* of an array of property/value pairs as statements.
* @param $metadataArray array
* @return MetadataDescription
*/
function &getNlm30CitationDescriptionFromMetadataArray(&$metadataArray)
{
// Create a new citation description
$citationDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30CitationSchema', ASSOC_TYPE_CITATION);
// Add the meta-data to the description
$metadataArray = arrayClean($metadataArray);
if (!$citationDescription->setStatements($metadataArray)) {
$translationParams = array('filterName' => $this->getDisplayName());
$this->addError(__('submission.citations.filter.invalidMetadata', $translationParams));
$nullVar = null;
return $nullVar;
}
// Set display name in the meta-data description
// to the corresponding value from the filter. This is important
// so that we later know which result came from which filter.
$citationDescription->setDisplayName($this->getDisplayName());
return $citationDescription;
}
/**
* Instantiate an NLM name description from an array.
* @param $personArray array
* @param $assocType integer
* @return MetadataDescription
*/
private function &instantiateNlm30NameDescriptions(&$personArray, $assocType)
{
$personDescriptions = array();
foreach ($personArray as $key => $person) {
if ($person == PERSON_STRING_FILTER_ETAL) {
$personDescription = 'et-al';
} else {
// Create a new NLM name description and fill it
// with the values from the test array.
$personDescription = new MetadataDescription('lib.pkp.plugins.metadata.nlm30.schema.Nlm30NameSchema', $assocType);
self::assertTrue($personDescription->setStatements($person));
}
// Add the result to the descriptions list
$personDescriptions[$key] = $personDescription;
}
return $personDescriptions;
}
/**
* Adds the data of an array of property/value pairs
* as statements to an NLM citation description.
* If no citation description is given, a new one will
* be instantiated.
* @param $metadataArray array
* @param $citationDescription MetadataDescription
* @return MetadataDescription
*/
function &addMetadataArrayToNlmCitationDescription(&$metadataArray, $citationDescription = null)
{
// Create a new citation description if no one was given
if (is_null($citationDescription)) {
$metadataSchema = new NlmCitationSchema();
$citationDescription = new MetadataDescription($metadataSchema, ASSOC_TYPE_CITATION);
}
// Add the meta-data to the description
if (!$citationDescription->setStatements($metadataArray)) {
$nullVar = null;
return $nullVar;
}
return $citationDescription;
}
/**
* @see MetadataDataObjectAdapter::extractMetadataFromDataObject()
* @param $dataObject Citation
* @return MetadataDescription
*/
function &extractMetadataFromDataObject(&$dataObject)
{
$metadataDescription =& $this->instantiateMetadataDescription();
// Identify the length of the name space prefix
$namespacePrefixLength = strlen($this->getMetadataNamespace()) + 1;
// Get all meta-data field names
$fieldNames = array_merge($this->getDataObjectMetadataFieldNames(false), $this->getDataObjectMetadataFieldNames(true));
// Retrieve the statements from the data object
$statements = array();
$nameSchema = new NlmNameSchema();
foreach ($fieldNames as $fieldName) {
if ($dataObject->hasData($fieldName)) {
// Remove the name space prefix
$propertyName = substr($fieldName, $namespacePrefixLength);
if (in_array($propertyName, array('person-group[@person-group-type="author"]', 'person-group[@person-group-type="editor"]'))) {
// Convert key/value arrays to MetadataDescription objects.
$names =& $dataObject->getData($fieldName);
foreach ($names as $key => $name) {
switch ($propertyName) {
case 'person-group[@person-group-type="author"]':
$assocType = ASSOC_TYPE_AUTHOR;
break;
case 'person-group[@person-group-type="editor"]':
$assocType = ASSOC_TYPE_EDITOR;
break;
}
$nameDescription = new MetadataDescription($nameSchema, $assocType);
$nameDescription->setStatements($name);
$names[$key] =& $nameDescription;
unset($nameDescription);
}
$statements[$propertyName] =& $names;
} else {
$statements[$propertyName] =& $dataObject->getData($fieldName);
}
}
}
// Set the statements in the meta-data description
$metadataDescription->setStatements($statements);
return $metadataDescription;
}
/**
* Instantiate an NLM name description from an array.
* @param $personArray array
* @param $assocType integer
* @return MetadataDescription
*/
private function &instantiateNlmNameDescriptions(&$personArray, $assocType)
{
$nlmNameSchema = new NlmNameSchema();
$personDescriptions = array();
foreach ($personArray as $key => $person) {
// Create a new NLM name description and fill it
// with the values from the test array.
$personDescription = new MetadataDescription($nlmNameSchema, $assocType);
self::assertTrue($personDescription->setStatements($person));
// Add the result to the descriptions list
$personDescriptions[$key] = $personDescription;
}
return $personDescriptions;
}
