D-Net

package eu.dnetlib.data.resultSet.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 com.google.common.base.Function;

import com.google.common.collect.Lists;

public class Cleaner implements Function<String, String> {

	private static final Log log = LogFactory.getLog(Cleaner.class);

	private List<XPathCleaningRule> xpathRules = Lists.newArrayList();

	public Cleaner(final List<XPathCleaningRule> xpathRules) {

		this.xpathRules = xpathRules;

	}

	@Override

	public String apply(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");

		}

	}

}