68 lines
2.1 KiB
Java
68 lines
2.1 KiB
Java
package eu.dnetlib.data.utility.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 org.springframework.beans.factory.annotation.Required;
|
|
|
|
import eu.dnetlib.miscutils.functional.UnaryFunction;
|
|
|
|
public class CleaningRule implements UnaryFunction<String, String> {
|
|
|
|
private static final Log log = LogFactory.getLog(CleaningRule.class); // NOPMD by marko on 11/24/08 5:02 PM
|
|
|
|
private List<XPATHCleaningRule> xpathRules = new ArrayList<XPATHCleaningRule>();
|
|
|
|
@Override
|
|
public String evaluate(final String text) {
|
|
|
|
try {
|
|
final List<Map<String, String>> errors = new ArrayList<Map<String, String>>();
|
|
final Document doc = (new SAXReader()).read(new StringReader(text));
|
|
for (final XPATHCleaningRule r : xpathRules) {
|
|
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 "";
|
|
}
|
|
|
|
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<XPATHCleaningRule> getXpathRules() {
|
|
return xpathRules;
|
|
}
|
|
|
|
@Required
|
|
public void setXpathRules(final List<XPATHCleaningRule> xpathRules) {
|
|
this.xpathRules = xpathRules;
|
|
}
|
|
|
|
}
|