/**
* DataCleaner (community edition)
* Copyright (C) 2014 Neopost - Customer Information Management
*
* This copyrighted material is made available to anyone wishing to use, modify,
* copy, or redistribute it subject to the terms and conditions of the GNU
* Lesser General Public License, as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
* or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License
* for more details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this distribution; if not, write to:
* Free Software Foundation, Inc.
* 51 Franklin Street, Fifth Floor
* Boston, MA 02110-1301 USA
*/
package org.datacleaner.spark;
import java.io.ByteArrayInputStream;
import java.io.InputStream;
import java.io.Serializable;
import java.io.UnsupportedEncodingException;
import java.net.URI;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.List;
import java.util.Map;
import java.util.concurrent.atomic.AtomicInteger;
import org.apache.metamodel.util.CollectionUtils;
import org.apache.metamodel.util.HdfsResource;
import org.apache.spark.api.java.JavaSparkContext;
import org.datacleaner.configuration.DataCleanerConfiguration;
import org.datacleaner.configuration.JaxbConfigurationReader;
import org.datacleaner.job.AnalysisJob;
import org.datacleaner.job.ComponentJob;
import org.datacleaner.job.JaxbJobReader;
import org.datacleaner.job.OutputDataStreamJob;
import org.datacleaner.job.builder.AnalysisJobBuilder;
import org.datacleaner.job.builder.AnalyzerComponentBuilder;
import org.datacleaner.job.builder.ComponentBuilder;
import org.datacleaner.spark.utils.HdfsHelper;
import org.datacleaner.util.InputStreamToPropertiesMapFunc;
import org.datacleaner.util.SystemProperties;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.google.common.base.Strings;
/**
* A container for for values that need to be passed between Spark workers. All
* the values need to be {@link Serializable} or loadable from those
* {@link Serializable} properties.
*/
public class SparkJobContext implements Serializable {
private static final String METADATA_PROPERTY_COMPONENT_INDEX = "org.datacleaner.spark.component.index";
private static final Logger logger = LoggerFactory.getLogger(SparkJobContext.class);
private static final String PROPERTY_RESULT_PATH = "datacleaner.result.hdfs.path";
private static final String PROPERTY_RESULT_ENABLED = "datacleaner.result.hdfs.enabled";
private static final long serialVersionUID = 1L;
private final String _jobName;
private final String _configurationXml;
private final String _analysisJobXml;
private final Map<String, String> _customProperties;
private final List<SparkJobLifeCycleListener> _sparkJobLifeCycleListeners = new ArrayList<>();
// cached/transient state
private transient DataCleanerConfiguration _dataCleanerConfiguration;
private transient AnalysisJobBuilder _analysisJobBuilder;
public SparkJobContext(final URI dataCleanerConfigurationPath, final URI analysisJobXmlPath,
final URI customPropertiesPath, final JavaSparkContext sparkContext) {
final HdfsHelper hdfsHelper = new HdfsHelper(sparkContext);
_jobName = getAnalysisJobName(analysisJobXmlPath);
logger.info("Loading SparkJobContext for {} - job name '{}'", analysisJobXmlPath, _jobName);
_configurationXml = hdfsHelper.readFile(dataCleanerConfigurationPath, true);
if (Strings.isNullOrEmpty(_configurationXml)) {
throw new IllegalArgumentException(
"Failed to read content from configuration file: " + dataCleanerConfigurationPath);
}
_analysisJobXml = hdfsHelper.readFile(analysisJobXmlPath, true);
if (Strings.isNullOrEmpty(_analysisJobXml)) {
throw new IllegalArgumentException("Failed to read content from job file: " + analysisJobXmlPath);
}
final String propertiesString = hdfsHelper.readFile(customPropertiesPath);
if (propertiesString == null) {
_customProperties = Collections.emptyMap();
} else {
// this is a pretty ugly way to go back to the bytes to read the
// properties - but works and is quick
_customProperties =
new InputStreamToPropertiesMapFunc().eval(new ByteArrayInputStream(propertiesString.getBytes()));
}
validateCustomProperties();
}
public SparkJobContext(final String jobName, final String dataCleanerConfiguration, final String analysisJobXml,
final Map<String, String> customProperties) {
_jobName = jobName;
_customProperties = customProperties;
_configurationXml = dataCleanerConfiguration;
_analysisJobXml = analysisJobXml;
validateCustomProperties();
}
/**
* Gets the job name (removing the extension '.analysis.xml')
*
* @return
*/
private static String getAnalysisJobName(final URI uri) {
final String filename = uri.getPath();
final int lastIndexOfSlash = filename.lastIndexOf("/");
final int lastIndexOfFileExtension = filename.lastIndexOf(".analysis.xml");
return filename.substring(lastIndexOfSlash + 1, lastIndexOfFileExtension);
}
private void validateCustomProperties() {
if (isResultEnabled()) {
// ensure parsability of result path
getResultPath();
}
}
public DataCleanerConfiguration getConfiguration() {
if (_dataCleanerConfiguration == null) {
final JaxbConfigurationReader confReader =
new JaxbConfigurationReader(new SparkConfigurationReaderInterceptor(_customProperties));
_dataCleanerConfiguration = confReader.read(createInputStream(_configurationXml));
}
return _dataCleanerConfiguration;
}
public String getJobName() {
return _jobName;
}
private InputStream createInputStream(final String fileContents) {
try {
final byte[] bytes = fileContents.getBytes("UTF-8");
return new ByteArrayInputStream(bytes);
} catch (final UnsupportedEncodingException e) {
throw new IllegalStateException(e);
}
}
public AnalysisJob getAnalysisJob() {
return getAnalysisJobBuilder().toAnalysisJob();
}
public AnalysisJobBuilder getAnalysisJobBuilder() {
if (_analysisJobBuilder == null) {
// set HDFS as default scheme to avoid file resources
SystemProperties.setIfNotSpecified(SystemProperties.DEFAULT_RESOURCE_SCHEME, HdfsResource.SCHEME_HDFS);
final DataCleanerConfiguration configuration = getConfiguration();
final JaxbJobReader jobReader = new JaxbJobReader(configuration);
_analysisJobBuilder = jobReader.create(createInputStream(_analysisJobXml), _customProperties);
}
applyComponentIndexForKeyLookups(_analysisJobBuilder, new AtomicInteger(0));
return _analysisJobBuilder;
}
/**
* Appliesc compopnent indices via the component metadata to enable proper
* functioning of the {@link #getComponentByKey(String)} and
* {@link #getComponentKey(ComponentJob)} methods.
*
* @param analysisJobBuilder
* @param currentComponentIndex
*/
private void applyComponentIndexForKeyLookups(final AnalysisJobBuilder analysisJobBuilder,
final AtomicInteger currentComponentIndex) {
final Collection<ComponentBuilder> componentBuilders = analysisJobBuilder.getComponentBuilders();
for (final ComponentBuilder componentBuilder : componentBuilders) {
componentBuilder.setMetadataProperty(METADATA_PROPERTY_COMPONENT_INDEX,
Integer.toString(currentComponentIndex.getAndIncrement()));
}
final List<AnalysisJobBuilder> childJobBuilders = analysisJobBuilder.getConsumedOutputDataStreamsJobBuilders();
for (final AnalysisJobBuilder childJobBuilder : childJobBuilders) {
applyComponentIndexForKeyLookups(childJobBuilder, currentComponentIndex);
}
}
public String getComponentKey(final ComponentJob componentJob) {
final String key = componentJob.getMetadataProperties().get(METADATA_PROPERTY_COMPONENT_INDEX);
if (key == null) {
throw new IllegalStateException("No key registered for component: " + componentJob);
}
final String partitionKey = componentJob.getMetadataProperties()
.get(AnalyzerComponentBuilder.METADATA_PROPERTY_BUILDER_PARTITION_INDEX);
if (partitionKey != null) {
return key + "." + partitionKey;
} else {
return key;
}
}
public ComponentJob getComponentByKey(final String key) {
final AnalysisJob job = getAnalysisJob();
final ComponentJob result = getComponentByKey(job, key);
if (result == null) {
throw new IllegalArgumentException("Cannot resolve component with key: " + key);
}
return result;
}
private ComponentJob getComponentByKey(final AnalysisJob job, final String queriedKey) {
final List<ComponentJob> componentJobs = CollectionUtils
.concat(false, job.getTransformerJobs(), job.getTransformerJobs(), job.getAnalyzerJobs());
for (final ComponentJob componentJob : componentJobs) {
final String componentKey = getComponentKey(componentJob);
if (queriedKey.equals(componentKey)) {
return componentJob;
}
final OutputDataStreamJob[] outputDataStreamJobs = componentJob.getOutputDataStreamJobs();
for (final OutputDataStreamJob outputDataStreamJob : outputDataStreamJobs) {
final AnalysisJob childJob = outputDataStreamJob.getJob();
if (childJob != null) {
final ComponentJob result = getComponentByKey(childJob, queriedKey);
if (result != null) {
return result;
}
}
}
}
return null;
}
/**
* Gets the path defined in the job properties file
*
* @return
*/
public URI getResultPath() {
final String str = _customProperties.get(PROPERTY_RESULT_PATH);
if (Strings.isNullOrEmpty(str)) {
return null;
}
return URI.create(str);
}
public boolean isResultEnabled() {
final String enabledString = _customProperties.get(PROPERTY_RESULT_ENABLED);
return !"false".equalsIgnoreCase(enabledString);
}
/**
* Adds a listener for the job life cycle.
*
* @param sparkJobLifeCycleListener
* The listener to add. Must be serializable.
*/
public void addSparkJobLifeCycleListener(final SparkJobLifeCycleListener sparkJobLifeCycleListener) {
_sparkJobLifeCycleListeners.add(sparkJobLifeCycleListener);
}
/**
* Removes a life cycle listener. Please note that this will _not_ work
* globally after job start. If you remove it on a node, it will only be
* removed on that node.
*/
public void removeSparkJobLifeCycleListener(final SparkJobLifeCycleListener sparkJobLifeCycleListener) {
_sparkJobLifeCycleListeners.remove(sparkJobLifeCycleListener);
}
public void triggerOnPartitionProcessingEnd() {
for (final SparkJobLifeCycleListener listener : _sparkJobLifeCycleListeners) {
try {
listener.onPartitionProcessingEnd(this);
} catch (final Throwable e) {
logger.warn("onPartitionProcessingEnd: Listener {} threw exception", listener, e);
}
}
}
public void triggerOnPartitionProcessingStart() {
for (final SparkJobLifeCycleListener listener : _sparkJobLifeCycleListeners) {
try {
listener.onPartitionProcessingStart(this);
} catch (final Throwable e) {
logger.warn("onPartitionProcessingStart: Listener {} threw exception", listener, e);
}
}
}
public void triggerOnJobStart() {
for (final SparkJobLifeCycleListener listener : _sparkJobLifeCycleListeners) {
try {
listener.onJobStart(this);
} catch (final Throwable e) {
logger.warn("onJobStart: Listener {} threw exception", listener, e);
}
}
}
public void triggerOnJobEnd() {
for (final SparkJobLifeCycleListener listener : _sparkJobLifeCycleListeners) {
try {
listener.onJobEnd(this);
} catch (final Throwable e) {
logger.warn("onJobEnd: Listener {} threw exception", listener, e);
}
}
}
}