D-Net

package eu.dnetlib.msro.openaireplus.workflows.nodes.index;

import java.io.IOException;

import java.io.InputStream;

import java.io.StringReader;

import java.io.StringWriter;

import javax.annotation.Resource;

import javax.xml.transform.Transformer;

import javax.xml.transform.TransformerException;

import javax.xml.transform.TransformerFactory;

import javax.xml.transform.stream.StreamResult;

import javax.xml.transform.stream.StreamSource;

import com.googlecode.sarasvati.Arc;

import com.googlecode.sarasvati.NodeToken;

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.miscutils.datetime.DateUtils;

import eu.dnetlib.miscutils.functional.hash.Hashing;

import eu.dnetlib.msro.workflows.nodes.SimpleJobNode;

import org.apache.commons.io.IOUtils;

import org.apache.commons.lang.StringUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

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

import org.springframework.core.io.ClassPathResource;

public class PrepareIndexDataJobNode extends SimpleJobNode {

	/**

	 * logger.

	 */

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

	@Resource

	private UniqueServiceLocator serviceLocator;

	/**

	 * Stylesheet which transforms a layout to another stylesheet which converts a input record to a index record.

	 */

	private String layoutToRecordStylesheet;

	private String outputRecordsPathParam;

	private String rottenRecordsPathParam;

	private String hbaseTable;

	private String oafSchemaLocationProperty;

	@Override

	protected String execute(final NodeToken token) throws Exception {

		log.info("start preparing job");

		final String xslt = prepareXslt(env("format", token), env("layout", token));

		token.getEnv().setAttribute("index.xslt", xslt);

		if (!StringUtils.isBlank(getOutputRecordsPathParam())) {

			token.getEnv().setAttribute(getOutputRecordsPathParam(), "/tmp" + getFileName(token, "indexrecords"));

		}

		if (!StringUtils.isBlank(getRottenRecordsPathParam())) {

			token.getEnv().setAttribute(getRottenRecordsPathParam(), "/tmp" + getFileName(token, "rottenrecords"));

		}

		token.getEnv().setAttribute("index.solr.url", getIndexSolrUrlZk());

		token.getEnv().setAttribute("index.solr.collection", getCollectionName(token));

		token.getEnv().setAttribute("index.shutdown.wait.time", getIndexSolrShutdownWait());

		token.getEnv().setAttribute("index.buffer.flush.threshold", getIndexBufferFlushTreshold());

		token.getEnv().setAttribute("index.solr.sim.mode", isFeedingSimulationMode());

		token.getEnv().setAttribute("index.feed.timestamp", DateUtils.now_ISO8601());

		token.getEnv().setAttribute(getOafSchemaLocationProperty(), getPropertyFetcher().getProperty(getOafSchemaLocationProperty()));

		return Arc.DEFAULT_ARC;

	}

	protected String tableName(final NodeToken token) {

		if (token.getEnv().hasAttribute("hbaseTable")) {

			String table = token.getEnv().getAttribute("hbaseTable");

			log.debug("found override value in wfEnv for 'hbaseTable' param: " + table);

			return table;

		}

		return getHbaseTable();

	}

	public String getIndexSolrUrlZk() throws ISLookUpException {

		return getServiceConfigValue(

				"for $x in /RESOURCE_PROFILE[.//RESOURCE_TYPE/@value='IndexServiceResourceType'] return $x//PROTOCOL[./@name='solr']/@address/string()");

	}

	public String getIndexSolrShutdownWait() throws ISLookUpException {

		return queryForServiceProperty("solr:feedingShutdownTolerance");

	}

	public String getIndexBufferFlushTreshold() throws ISLookUpException {

		return queryForServiceProperty("solr:feedingBufferFlushThreshold");

	}

	public String isFeedingSimulationMode() throws ISLookUpException {

		return queryForServiceProperty("solr:feedingSimulationMode");

	}

	private String queryForServiceProperty(final String key) throws ISLookUpException {

		return getServiceConfigValue(

				"for $x in /RESOURCE_PROFILE[.//RESOURCE_TYPE/@value='IndexServiceResourceType'] return $x//SERVICE_PROPERTIES/PROPERTY[./@ key='"

						+ key + "']/@value/string()");

	}

	private String getServiceConfigValue(final String xquery) throws ISLookUpException {

		log.debug("quering for service property: " + xquery);

		final String res = serviceLocator.getService(ISLookUpService.class).getResourceProfileByQuery(xquery);

		if (StringUtils.isBlank(res)) { throw new IllegalStateException("unable to find unique service property, xquery: " + xquery); }

		return res;

	}

	private String getFileName(final NodeToken token, final String fileNamePrefix) {

		return "/" + fileNamePrefix + "_" + tableName(token) + "_" + token.getEnv().getAttribute("format") + ".seq";

	}

	private String getCollectionName(final NodeToken token) {

		return env("format", token) + "-" + env("layout", token) + "-" + env("interpretation", token);

	}

	private String env(final String s, final NodeToken token) {

		return token.getEnv().getAttribute(s);

	}

	/**

	 * Transforms each OAF record into a index record.

	 *

	 * @param format         format

	 * @param layout         layout

	 * @return resultset with transformed records

	 * @throws ISLookUpException    could happen

	 * @throws IOException          could happen

	 * @throws TransformerException could happen

	 */

	protected String prepareXslt(final String format, final String layout) throws ISLookUpException, IOException, TransformerException {

		final TransformerFactory factory = TransformerFactory.newInstance();

		final Transformer layoutTransformer = factory.newTransformer(new StreamSource(new StringReader(readXslt(getLayoutToRecordStylesheet()))));

		final StreamResult layoutToXsltXslt = new StreamResult(new StringWriter());

		layoutTransformer.setParameter("format", format);

		layoutTransformer.transform(new StreamSource(new StringReader(getLayoutSource(format, layout))), layoutToXsltXslt);

		return new String(Hashing.encodeBase64(layoutToXsltXslt.getWriter().toString()));

	}

	private String readXslt(final String s) throws IOException {

		ClassPathResource resource = new ClassPathResource(s);

		InputStream inputStream = resource.getInputStream();

		return IOUtils.toString(inputStream);

	}

	private String getLayoutSource(final String format, final String layout) throws ISLookUpDocumentNotFoundException, ISLookUpException {

		return serviceLocator.getService(ISLookUpService.class).getResourceProfileByQuery(

				"collection('')//RESOURCE_PROFILE[.//RESOURCE_TYPE/@value = 'MDFormatDSResourceType' and .//NAME='" + format + "']//LAYOUT[@name='" + layout

						+ "']");

	}

	public String getLayoutToRecordStylesheet() {

		return layoutToRecordStylesheet;

	}

	public void setLayoutToRecordStylesheet(final String layoutToRecordStylesheet) {

		this.layoutToRecordStylesheet = layoutToRecordStylesheet;

	}

	public String getHbaseTable() {

		return hbaseTable;

	}

	@Required

	public void setHbaseTable(final String hbaseTable) {

		this.hbaseTable = hbaseTable;

	}

	public String getOutputRecordsPathParam() {

		return outputRecordsPathParam;

	}

	public void setOutputRecordsPathParam(final String outputRecordsPathParam) {

		this.outputRecordsPathParam = outputRecordsPathParam;

	}

	public String getRottenRecordsPathParam() {

		return rottenRecordsPathParam;

	}

	public void setRottenRecordsPathParam(final String rottenRecordsPathParam) {

		this.rottenRecordsPathParam = rottenRecordsPathParam;

	}

	public String getOafSchemaLocationProperty() {

		return oafSchemaLocationProperty;

	}

	public void setOafSchemaLocationProperty(final String oafSchemaLocationProperty) {

		this.oafSchemaLocationProperty = oafSchemaLocationProperty;

	}

}