D-Net

package eu.dnetlib.data.hadoop.action;

import java.io.StringReader;

import java.util.Map;

import java.util.Map.Entry;

import java.util.Set;

import java.util.concurrent.Executor;

import java.util.concurrent.Executors;

import javax.annotation.Resource;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.apache.hadoop.conf.Configuration;

import org.dom4j.Document;

import org.dom4j.DocumentException;

import org.dom4j.Node;

import org.dom4j.io.SAXReader;

import org.springframework.beans.factory.annotation.Autowired;

import com.google.common.collect.Sets;

import eu.dnetlib.data.hadoop.config.ClusterName;

import eu.dnetlib.data.hadoop.config.ConfigurationEnumerator;

import eu.dnetlib.data.hadoop.rmi.HadoopBlackboardActions;

import eu.dnetlib.data.hadoop.rmi.HadoopJobType;

import eu.dnetlib.data.hadoop.rmi.HadoopServiceException;

import eu.dnetlib.data.hadoop.utils.JobProfile;

import eu.dnetlib.data.hadoop.utils.ScanFactory;

import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpDocumentNotFoundException;

import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;

import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;

import eu.dnetlib.enabling.locators.UniqueServiceLocator;

import eu.dnetlib.enabling.tools.blackboard.BlackboardJob;

import eu.dnetlib.enabling.tools.blackboard.BlackboardServerAction;

import eu.dnetlib.enabling.tools.blackboard.BlackboardServerHandler;

public abstract class AbstractHadoopAction implements BlackboardServerAction<HadoopBlackboardActions> {

	private static final Log log = LogFactory.getLog(AbstractHadoopAction.class); // NOPMD by marko on 11/24/08 5:02 PM

	/**

	 * Special I/O hdfs property names for which we support relative paths.

	 */

	public final static Set<String> HDFS_SPECIAL_PROPERTIES = Sets.newHashSet("mapred.input.dir", "mapred.output.dir");

	@Resource

	protected UniqueServiceLocator serviceLocator;

	@Autowired

	protected ConfigurationEnumerator configurationEnumerator;

	private final Executor executor = Executors.newCachedThreadPool();

	protected abstract void executeAsync(final BlackboardServerHandler handler, final BlackboardJob job) throws HadoopServiceException;

	@Override

	public void execute(final BlackboardServerHandler handler, final BlackboardJob job) {

		executor.execute(new Runnable() {

			@Override

			public void run() {

				try {

					handler.ongoing(job);

					executeAsync(handler, job);

				} catch (Throwable e) {

					log.error("exception: " + e);

					log.error("cause: " + e.getCause());

					handler.failed(job, e);

				}

			}

		});

	}

	protected JobProfile loadISJobConfiguration(final String jobName, final Map<String, String> bbParams) throws HadoopServiceException {

		log.info("reading job configuration profile: " + jobName);

		try {

			String profile = serviceLocator.getService(ISLookUpService.class).getResourceProfileByQuery(

					"/RESOURCE_PROFILE[.//RESOURCE_TYPE/@value = 'HadoopJobConfigurationDSResourceType' and .//HADOOP_JOB/@name='" + jobName + "']");

			return parseJobProfile(profile, bbParams);

		} catch (ISLookUpDocumentNotFoundException e) {

			throw new HadoopServiceException("cannot find job profile: " + jobName, e);

		} catch (ISLookUpException e) {

			throw new HadoopServiceException("unable to read job profile: " + jobName, e);

		}

	}

	private JobProfile parseJobProfile(final String profile, final Map<String, String> bbParams) throws HadoopServiceException {

		JobProfile jobProfile = new JobProfile();

		try {

			Document doc = new SAXReader().read(new StringReader(profile));

			log.debug("setting job description");

			jobProfile.setDescription(doc.valueOf("//DESCRIPTION"));

			log.debug("setting job name");

			jobProfile.setName(doc.valueOf("//HADOOP_JOB/@name"));

			log.debug("setting job type");

			jobProfile.setJobType(HadoopJobType.valueOf(doc.valueOf("//HADOOP_JOB/@type")));

			log.debug("setting job static configuration");

			for (Object o : doc.selectNodes("//STATIC_CONFIGURATION/PROPERTY")) {

				Node node = (Node) o;

				jobProfile.getJobDefinition().put(node.valueOf("./@key"), node.valueOf("./@value"));

			}

			log.debug("setting job required parameters");

			for (Object o : doc.selectNodes("//JOB_INTERFACE/PARAM[./@required = 'true']")) {

				Node node = (Node) o;

				jobProfile.getRequiredParams().add(node.valueOf("./@name"));

			}

			if (doc.selectSingleNode("//SCAN") != null) {

				jobProfile.setScanProperties(ScanFactory.parseScanProperties(doc, bbParams));

			}

		} catch (DocumentException e) {

			throw new HadoopServiceException("cannot parse job profile");

		}

		if (jobProfile.isEmpty()) { throw new HadoopServiceException("job configuration is empty"); }

		return jobProfile;

	}

	protected Configuration getConf(final ClusterName clusterName) {

		return configurationEnumerator.get(clusterName);

	}

	protected boolean checkHdfsProperty(final Entry<String, String> e) {

		return HDFS_SPECIAL_PROPERTIES.contains(e.getKey()) && !e.getValue().isEmpty() && !e.getValue().startsWith("hdfs://");

	}

	public UniqueServiceLocator getServiceLocator() {

		return serviceLocator;

	}

}