dnet-applications/libs/dnet-data-services/src/main/java/eu/dnetlib/data/mapping/cleaner/Cleaner.java

64 lines
1.8 KiB
Java

package eu.dnetlib.data.mapping.cleaner;
import java.io.StringReader;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.dom4j.Document;
import org.dom4j.Element;
import org.dom4j.Namespace;
import org.dom4j.QName;
import org.dom4j.io.SAXReader;
import eu.dnetlib.data.mapping.RecordTransformer;
public class Cleaner extends RecordTransformer<String, String> {
private static final Log log = LogFactory.getLog(Cleaner.class);
private final List<VocabularyRule> rules = new ArrayList<>();
@Override
public String transform(final String source) {
try {
final List<Map<String, String>> errors = new ArrayList<>();
final Document doc = new SAXReader().read(new StringReader(source));
for (final VocabularyRule r : rules) {
errors.addAll(r.applyXpathRule(doc));
}
if (errors.size() > 0) {
markAsInvalid(doc, errors);
}
return doc.asXML();
} catch (final Exception e) {
log.error("Error evaluating rule", e);
}
return null;
}
private void markAsInvalid(final Document doc, final List<Map<String, String>> errors) {
final Element element = (Element) doc.selectSingleNode("//*[local-name()='header']");
if (element != null) {
final Element inv = element.addElement(new QName("invalid", new Namespace("dri", "http://www.driver-repository.eu/namespace/dri")));
for (final Map<String, String> e : errors) {
final Element err = inv.addElement(new QName("error", new Namespace("dri", "http://www.driver-repository.eu/namespace/dri")));
for (final Map.Entry<String, String> entry : e.entrySet()) {
err.addAttribute(entry.getKey(), entry.getValue());
}
}
inv.addAttribute("value", "true");
}
}
public List<VocabularyRule> getRules() {
return rules;
}
}