D-Net

[

{

  "type_source": "SVN", 

  "goal": "package -U -T 4C source:jar", 

  "url": "http://svn-public.driver.research-infrastructures.eu/driver/dnet40/modules/icm-iis-transformers/trunk/", 

  "deploy_repository": "dnet4-snapshots", 

  "version": "4",

  "mail": "m.horst@icm.edu.pl,d.tkaczyk@icm.edu.pl",

  "deploy_repository_url": "http://maven.research-infrastructures.eu/nexus/content/repositories/dnet4-snapshots", 

  "name": "icm-iis-transformers"

},

{

  "type_source": "SVN",

  "goal": "clean verify -U -e -X",

  "url": "http://svn-public.driver.research-infrastructures.eu/driver/dnet40/modules/icm-iis-transformers/trunk/",

  "nightly" : "true",

  "cron" : "H H * * *",

  "version": "4",

  "mail": "m.horst@icm.edu.pl",

  "name": "icm-iis-transformers-embedded-integration-test"

}

]

package eu.dnetlib.iis.transformers.referenceextraction;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author mhorst

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.export.documentmetadata;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author mhorst

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/export/documentmetadata/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.export.documenttodataset_without_imported_data;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/export/documenttodataset_without_imported_data/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.export.documenttoproject_without_imported_data;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/export/documenttoproject_without_imported_data/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.importer.documentmetadata.idextractor;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author mhorst

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

    public void testIdExtraction() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/importer/documentmetadata/idextractor/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.metadataextraction;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author mhorst

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

    public void testChecksumPreprocessing() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/metadataextraction/checksum/preprocessing/sampledataproducer/oozie_app");

    }

    @Test

    public void testChecksumPostprocessingText() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/metadataextraction/checksum/postprocessing/text/sampledataproducer/oozie_app");

    }

    @Test

    public void testChecksumPostprocessingMeta() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/metadataextraction/checksum/postprocessing/meta/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class IntegerFirstNotEmptyTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        IntegerFirstNotEmpty udf = new IntegerFirstNotEmpty();

        TupleFactory factory = TupleFactory.getInstance();

        assertNull(udf.exec(null));

        assertNull(udf.exec(factory.newTuple()));

        assertNull(udf.exec(factory.newTuple((Integer) null)));

        assertEquals(125, (int)udf.exec(factory.newTuple(Integer.valueOf(125))));

        assertNull(udf.exec(factory.newTuple(Lists.newArrayList())));

        assertNull(udf.exec(factory.newTuple(Lists.newArrayList(null, null))));

        assertEquals(23, (int)udf.exec(factory.newTuple(Lists.newArrayList(23, null, 256, 90))));

        assertEquals(256, (int)udf.exec(factory.newTuple(Lists.newArrayList(null, null, null, 256, 567))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.BagFactory;

import org.apache.pig.data.DataBag;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class StringBagsMergerTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        StringBagsMerger udf = new StringBagsMerger();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        BagFactory bagFactory = BagFactory.getInstance();

        DataBag emptyBag = bagFactory.newDefaultBag();

        DataBag bag1 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup2")));

        DataBag bag2 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup3"), 

                                                                   tupleFactory.newTuple("tup4")));

        DataBag bag3 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup4"),

                                                                   tupleFactory.newTuple("tup5")));

        DataBag bag4 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup2"),

                                                                   tupleFactory.newTuple("tup3"),

                                                                   tupleFactory.newTuple("tup4"),

                                                                   tupleFactory.newTuple("tup5")));

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertNull(udf.exec(tupleFactory.newTuple((DataBag)null)));

        assertNull(udf.exec(tupleFactory.newTuple(emptyBag)));

        assertEquals(bag1, udf.exec(tupleFactory.newTuple(bag1)));

        assertEquals(bag4, udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag1, emptyBag, bag2, bag3))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.BagFactory;

import org.apache.pig.data.DataBag;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class EmptyBagToNullTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        EmptyBagToNull udf = new EmptyBagToNull();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        BagFactory bagFactory = BagFactory.getInstance();

        DataBag emptyBag = bagFactory.newDefaultBag();

        DataBag bag = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), tupleFactory.newTuple()));

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertNull(udf.exec(tupleFactory.newTuple((DataBag)null)));

        assertNull(udf.exec(tupleFactory.newTuple(emptyBag)));

        assertEquals(bag, udf.exec(tupleFactory.newTuple(bag)));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag, emptyBag))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.BagFactory;

import org.apache.pig.data.DataBag;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class NullToEmptyBagTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        NullToEmptyBag udf = new NullToEmptyBag();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        BagFactory bagFactory = BagFactory.getInstance();

        DataBag emptyBag = bagFactory.newDefaultBag();

        DataBag bag = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), tupleFactory.newTuple()));

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertEquals(emptyBag, udf.exec(tupleFactory.newTuple((DataBag)null)));

        assertEquals(emptyBag, udf.exec(tupleFactory.newTuple(emptyBag)));

        assertEquals(bag, udf.exec(tupleFactory.newTuple(bag)));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag, emptyBag))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.BagFactory;

import org.apache.pig.data.DataBag;

import org.apache.pig.data.Tuple;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class NullTupleFieldsToNullTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        NullTupleFieldsToNull udf = new NullTupleFieldsToNull();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        BagFactory bagFactory = BagFactory.getInstance();

        DataBag emptyBag = bagFactory.newDefaultBag();

        DataBag bag = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), null));

        Tuple nullTuple = tupleFactory.newTuple(Lists.newArrayList(null, null, null));

        Tuple tuple = tupleFactory.newTuple(Lists.newArrayList(null, null, "tup1"));

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertNull(udf.exec(tupleFactory.newTuple((Tuple)null)));

        assertNull(udf.exec(tupleFactory.newTuple(tupleFactory.newTuple())));

        assertNull(udf.exec(tupleFactory.newTuple(nullTuple)));

        assertEquals(tuple, udf.exec(tupleFactory.newTuple(tuple)));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag, emptyBag))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import java.util.HashMap;

import java.util.Map;

import junit.framework.TestCase;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class StringMapsMergerTest extends TestCase {

    @SuppressWarnings({ "unchecked", "rawtypes" })

	@Test

	public void testUDF() throws IOException {

        StringMapsMerger udf = new StringMapsMerger();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        Map<String, String> emptyMap = new HashMap<String, String>();

        Map<String, String> map1 = new HashMap<String, String>();

        map1.put("12", "12");

        map1.put("3", "bag");

        map1.put("1", "");

        Map<String, String> map2 = new HashMap<String, String>();

        map2.put("12", "null");

        map2.put("3", "data");

        map2.put("35", "empty");

        Map<String, String> map3 = new HashMap<String, String>();

        map3.put("1", "notempty");

        Map<String, String> map4 = new HashMap<String, String>();

        map4.put("1", "");

        map4.put("3", "bag");

        map4.put("12", "12");

        map4.put("35", "empty");

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertNull(udf.exec(tupleFactory.newTuple((Map)null)));

        assertNull(udf.exec(tupleFactory.newTuple(emptyMap)));

        assertEquals(map1, udf.exec(tupleFactory.newTuple(map1)));

        assertEquals(map4, udf.exec(tupleFactory.newTuple(Lists.newArrayList(map1, emptyMap, map2, null, map3))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.BagFactory;

import org.apache.pig.data.DataBag;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class StringBagsDifferenceTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        StringBagsDifference udf = new StringBagsDifference();

        TupleFactory tupleFactory = TupleFactory.getInstance();

        BagFactory bagFactory = BagFactory.getInstance();

        DataBag nullBag = null;

        DataBag emptyBag = bagFactory.newDefaultBag();

        DataBag bag1 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup2")));

        DataBag bag2 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup3"), 

                                                                   tupleFactory.newTuple("tup4")));

        DataBag bag3 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup4"),

                                                                   tupleFactory.newTuple("tup5")));

        DataBag bag4 = bagFactory.newDefaultBag(Lists.newArrayList(tupleFactory.newTuple("tup1"), 

                                                                   tupleFactory.newTuple("tup5")));

        assertNull(udf.exec(null));

        assertNull(udf.exec(tupleFactory.newTuple()));

        assertNull(udf.exec(tupleFactory.newTuple(nullBag)));

        assertNull(udf.exec(tupleFactory.newTuple(emptyBag)));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag1, emptyBag, bag2, bag3))));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(nullBag, bag3, bag2))));

        assertEquals(bag1, udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag1, bag2))));

        assertEquals(bag4, udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag3, bag2))));

        assertNull(udf.exec(tupleFactory.newTuple(Lists.newArrayList(bag4, bag3))));

    }

}

package eu.dnetlib.iis.transformers.udfs;

import com.google.common.collect.Lists;

import java.io.IOException;

import junit.framework.TestCase;

import org.apache.pig.data.TupleFactory;

import org.junit.Test;

/**

 *

 * @author Dominika Tkaczyk

 */

public class StringFirstNotEmptyTest extends TestCase {

    @Test

	public void testUDF() throws IOException {

        StringFirstNotEmpty udf = new StringFirstNotEmpty();

        TupleFactory factory = TupleFactory.getInstance();

        assertNull(udf.exec(null));

        assertNull(udf.exec(factory.newTuple()));

        assertNull(udf.exec(factory.newTuple((String) null)));

        assertEquals("tup", udf.exec(factory.newTuple("tup")));

        assertNull(udf.exec(factory.newTuple(Lists.newArrayList())));

        assertNull(udf.exec(factory.newTuple(Lists.newArrayList(null, null))));

        assertEquals("val1", udf.exec(factory.newTuple(Lists.newArrayList("val1", null, "256", "90"))));

        assertEquals("k256", udf.exec(factory.newTuple(Lists.newArrayList(null, null, null, "k256", "567"))));

        assertEquals("k256", udf.exec(factory.newTuple(Lists.newArrayList(null, "", null, "k256", "567"))));

    }

}

package eu.dnetlib.iis.transformers.metadatamerger;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testJoin() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/metadatamerger/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.citationmatching;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testJoin() throws Exception {

    	runWorkflow("eu/dnetlib/iis/transformers/citationmatching/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.common.existencefilter.sampledataproducer;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Mateusz Fedoryszak

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/common/existencefilter/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.documentssimilarity;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Michal Oniszczuk (m.oniszczuk@icm.edu.pl)

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/documentssimilarity/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.ingest.pmc.metadata;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author mhorst

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/ingest/pmc/metadata/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.documentssimilarity_with_fulltext;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import eu.dnetlib.iis.core.WorkflowConfiguration;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Michal Oniszczuk (m.oniszczuk@icm.edu.pl)

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        WorkflowConfiguration wf = new WorkflowConfiguration();

        wf.setTimeoutInSeconds(720);

        runWorkflow("eu/dnetlib/iis/transformers/documentssimilarity_with_fulltext/sampledataproducer/oozie_app", wf);

    }

}

package eu.dnetlib.iis.transformers.idreplacer;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 * @author Michal Oniszczuk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

    public void testReplacer1Field() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/idreplacer/replacer_1_field/oozie_app");

    }

    @Test

	public void testReplacer2Fields() throws Exception {

    	runWorkflow("eu/dnetlib/iis/transformers/idreplacer/replacer_2_fields/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.metricsprimary;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/metricsprimary/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.statistics;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/statistics/sampledataproducer/oozie_app");

    }

}

package eu.dnetlib.iis.transformers.documentsclassification;

import eu.dnetlib.iis.IntegrationTest;

import eu.dnetlib.iis.core.AbstractWorkflowTestCase;

import org.junit.Test;

import org.junit.experimental.categories.Category;

/**

 * 

 * @author Dominika Tkaczyk

 *

 */

@Category(IntegrationTest.class)

public class WorkflowTest extends AbstractWorkflowTestCase {

    @Test

	public void testWorkflow() throws Exception {

        runWorkflow("eu/dnetlib/iis/transformers/documentsclassification/sampledataproducer/oozie_app");

    }

}

## This is a classpath-based import file (this header is required)

transformer_project_toconcept classpath eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/oozie_app

<workflow-app xmlns="uri:oozie:workflow:0.4" name="test-transformers_referenceextraction_project_toconcept">

    <start to="producer"/>

    <action name="producer">

        <java>

            <job-tracker>${jobTracker}</job-tracker>

            <name-node>${nameNode}</name-node>

			<!-- The data generated by this node is deleted in this section -->

			<prepare>

				<delete path="${nameNode}${workingDir}/producer" />

				<mkdir path="${nameNode}${workingDir}/producer" />

			</prepare>

            <configuration>

                <property>

                    <name>mapred.job.queue.name</name>

                    <value>${queueName}</value>

                </property>

            </configuration>

            <!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.Producer</arg>

			<!-- Specification of the output ports -->

            <arg>-C{document_to_project,

				eu.dnetlib.iis.referenceextraction.project.schemas.DocumentToProject,

				eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/sampledataproducer/data/document_to_project.json}</arg>

			<arg>-C{project,

				eu.dnetlib.iis.importer.schemas.Project,

				eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/sampledataproducer/data/project.json}</arg>

			<arg>-C{concept,

				eu.dnetlib.iis.importer.schemas.Concept,

				eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/sampledataproducer/data/concept.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

            <arg>-SworkingDir=${workingDir}/producer/working_dir</arg>

            <arg>-Odocument_to_project=${workingDir}/producer/document_to_project</arg>

            <arg>-Oproject=${workingDir}/producer/project</arg>

            <arg>-Oconcept=${workingDir}/producer/concept</arg>

        </java>

        <ok to="transformer_project_toconcept"/>

        <error to="fail"/>

    </action>

    <action name="transformer_project_toconcept">

        <sub-workflow>

            <app-path>${wf:appPath()}/transformer_project_toconcept</app-path>

            <configuration>

                <property>

                    <name>jobTracker</name>

                    <value>${jobTracker}</value>

                </property>

                <property>

                    <name>nameNode</name>

                    <value>${nameNode}</value>

                </property>

                <property>

                    <name>queueName</name>

                    <value>${queueName}</value>

                </property>

                <!-- Working directory of the subworkflow -->

                <property>

                    <name>workingDir</name>

                    <value>${workingDir}/transformer_project_toconcept/working_dir</value>

                </property>

                <property>

                    <name>grant_id_param_name</name>

                    <value>CD_PROJECT_NUMBER</value>

                </property>

                <!-- Input ports. -->

                <property>

                    <name>input_document_to_project</name>

                    <value>${workingDir}/producer/document_to_project</value>

                </property>

                <property>

                    <name>input_project</name>

                    <value>${workingDir}/producer/project</value>

                </property>

                <property>

                    <name>input_concept</name>

                    <value>${workingDir}/producer/concept</value>

                </property>

                <!-- Output port bound to given path -->

                <property>

                    <name>output</name>

                    <value>${workingDir}/transformer_project_toconcept/output</value>

                </property>

            </configuration>

        </sub-workflow>

        <ok to="consumer"/>

        <error to="fail"/>

    </action>

    <action name="consumer">

		<java>

			<job-tracker>${jobTracker}</job-tracker>

			<name-node>${nameNode}</name-node>

			<configuration>

				<property>

					<name>mapred.job.queue.name</name>

					<value>${queueName}</value>

				</property>

			</configuration>

			<!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.TestingConsumer</arg>

			<!-- Specification of the input ports -->

			<arg>-C{output,

				eu.dnetlib.iis.referenceextraction.researchinitiative.schemas.DocumentToConceptId,

				eu/dnetlib/iis/transformers/referenceextraction/project/toconcept/sampledataproducer/data/output.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

			<arg>-SworkingDir=${workingDir}/consumer/working_dir</arg>

			<arg>-Ioutput=${workingDir}/transformer_project_toconcept/output</arg>

		</java>

		<ok to="end" />

		<error to="fail" />

	</action>    

    <kill name="fail">

		<message>Unfortunately, the workflow failed -- error message:

			[${wf:errorMessage(wf:lastErrorNode())}]</message>

    </kill>

    <end name="end"/>

</workflow-app>

{"id": "concept-id1", "label": "concept id1", "params": {"CD_PROJECT_NUMBER": "123456", "CD_FRAMEWORK": "FP7"}}

{"id": "concept-id2", "label": "concept id2", "params": {"CD_PROJECT_NUMBER": "654321", "CD_FRAMEWORK": "FP7"}}

{"id": "concept-id3", "label": "concept id3", "params": {"CD_PROJECT_NUMBER": "999999", "CD_FRAMEWORK": "FP7"}}

{"documentId": "docId1", "conceptId": "concept-id1", "confidenceLevel": 0.5}

{"documentId": "docId1", "conceptId": "concept-id2", "confidenceLevel": 0.1}

{"documentId": "docId2", "conceptId": "concept-id3", "confidenceLevel": 0.3}

{"documentId": "docId1", "projectId": "projId1", "confidenceLevel": 0.5}

{"documentId": "docId1", "projectId": "projId2", "confidenceLevel": 0.1}

{"documentId": "docId2", "projectId": "projId3", "confidenceLevel": 0.3}

{"documentId": "docId2", "projectId": "nonexistingProjId", "confidenceLevel": 0.1}

{"documentId": "docId2", "projectId": "noFetProjId", "confidenceLevel": 0.1}

{"documentId": "docId3", "projectId": "nonexistingProjId", "confidenceLevel": 0.1}

{"id":"projId1","projectAcronym":"project 1","projectGrantId":"123456","fundingClass":"FP7"}

{"id":"projId2","projectAcronym":"project 2","projectGrantId":"654321","fundingClass":"FP7"}

{"id":"projId3","projectAcronym":"project 3","projectGrantId":"999999","fundingClass":"FP7"}

{"id":"noFetProjId","projectAcronym":"non fet project","projectGrantId":"000000","fundingClass":"FP7"}

## This is a classpath-based import file (this header is required)

transformer_export_documentmetadata classpath eu/dnetlib/iis/transformers/export/documentmetadata/oozie_app

<workflow-app xmlns="uri:oozie:workflow:0.2" name="test-transformers_export_documentmetadata_sampledataproducer">

    <start to="producer"/>

    <action name="producer">

        <java>

            <job-tracker>${jobTracker}</job-tracker>

            <name-node>${nameNode}</name-node>

			<!-- The data generated by this node is deleted in this section -->

			<prepare>

				<delete path="${nameNode}${workingDir}/producer" />

				<mkdir path="${nameNode}${workingDir}/producer" />

			</prepare>

            <configuration>

                <property>

                    <name>mapred.job.queue.name</name>

                    <value>${queueName}</value>

                </property>

            </configuration>

            <!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.Producer</arg>

			<!-- Specification of the output ports -->

            <arg>-C{extracted_metadata,

				eu.dnetlib.iis.metadataextraction.schemas.ExtractedDocumentMetadata,

				eu/dnetlib/iis/transformers/export/documentmetadata/sampledataproducer/data/extr_metadata.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

            <arg>-SworkingDir=${workingDir}/producer/working_dir</arg>

            <arg>-Oextracted_metadata=${workingDir}/producer/extr_metadata</arg>

        </java>

        <ok to="transformer_export_documentmetadata"/>

        <error to="fail"/>

    </action>

    <action name="transformer_export_documentmetadata">

        <sub-workflow>

            <app-path>${wf:appPath()}/transformer_export_documentmetadata</app-path>

            <configuration>

                <property>

                    <name>jobTracker</name>

                    <value>${jobTracker}</value>

                </property>

                <property>

                    <name>nameNode</name>

                    <value>${nameNode}</value>

                </property>

                <property>

                    <name>queueName</name>

                    <value>${queueName}</value>

                </property>

                <!-- Working directory of the subworkflow -->

                <property>

                    <name>workingDir</name>

                    <value>${workingDir}/transformer_export_documentmetadata/working_dir</value>

                </property>

                <!-- Input ports. -->

                <property>

                    <name>input_extracted_metadata</name>

                    <value>${workingDir}/producer/extr_metadata</value>

                </property>

                <!-- Output port bound to given path -->

                <property>

                    <name>output_metadata</name>

                    <value>${workingDir}/transformer_export_documentmetadata/output_metadata</value>

                </property>

            </configuration>

        </sub-workflow>

        <ok to="consumer"/>

        <error to="fail"/>

    </action>

    <action name="consumer">

		<java>

			<job-tracker>${jobTracker}</job-tracker>

			<name-node>${nameNode}</name-node>

			<configuration>

				<property>

					<name>mapred.job.queue.name</name>

					<value>${queueName}</value>

				</property>

			</configuration>

			<!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.TestingConsumer</arg>

			<!-- Specification of the input ports -->

			<arg>-C{output_metadata,

				eu.dnetlib.iis.export.schemas.DocumentMetadata,

				eu/dnetlib/iis/transformers/export/documentmetadata/sampledataproducer/data/output_metadata.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

			<arg>-SworkingDir=${workingDir}/consumer/working_dir</arg>

			<arg>-Ioutput_metadata=${workingDir}/transformer_export_documentmetadata/output_metadata</arg>

		</java>

		<ok to="end" />

		<error to="fail" />

	</action>    

    <kill name="fail">

		<message>Unfortunately, the workflow failed -- error message:

			[${wf:errorMessage(wf:lastErrorNode())}]</message>

    </kill>

    <end name="end"/>

</workflow-app>

{"id": "id-2", "affiliations": []}

{"id": "id-3", "affiliations": [{"organization": "some-org", "countryName": "pl", "countryCode": "PL", "address": null, "rawText": "some raw text"}]}

{"publisher": "Tor Science Fiction", "affiliations": null, "authors": [{"authorFullName": "author-1", "affiliationPositions": null}, {"authorFullName": "author-2", "affiliationPositions": null}, {"authorFullName": "author-3", "affiliationPositions": null}], "language": "eng", "title": "Enders Game", "externalIdentifiers": null, "journal": "Journal-2", "id": "id-1", "pages": {"start": "123", "end": "128"}, "volume": null, "references": null, "year": null, "keywords": null, "issue": null, "abstract": null}

{"publisher": null, "affiliations": [], "authors": [{"authorFullName": "author-1", "affiliationPositions": null}, {"authorFullName": "author-2", "affiliationPositions": null}], "language": null, "title": null, "externalIdentifiers": {"id-1": "val-extr-1", "id-3": "val-extr-3"}, "journal": null, "id": "id-2", "pages": null, "volume": "124", "references": [{"position": 1, "basicMetadata": {"publisher": null, "title": "The Lord of the Rings", "url": null, "series": null, "authors": null, "volume": null, "edition": null, "source": null, "year": "2012", "issue": null, "pages": {"start": "1", "end": "236"}, "location": null}, "text": "J.R.R. Tolkien, The Lord of the Rings, 2012"}, {"position": 2, "basicMetadata": {"publisher": null, "title": "The Other Wind", "url": null, "series": null, "authors": null, "volume": "vol.23", "edition": null, "source": null, "year": "2003", "issue": null, "pages": null, "location": null}, "text": "Ursula K. Le Guin, The Other Wind, 2003"}], "year": 1970, "keywords": ["kwd_1", "kwd_3", "kwd_5"], "issue": null, "abstract": "The tales"}

{"publisher": "Harp3r T0rch", "affiliations": [{"organization": "some-org", "countryName": "pl", "countryCode": "PL", "address": null, "rawText": "some raw text"}], "authors": null, "language": "en", "title": "Small Gods", "externalIdentifiers": null, "journal": "Journal", "id": "id-3", "pages": null, "volume": "32", "references": [{"position": 1, "basicMetadata": {"publisher": null, "title": "The Lord of the Rings", "url": null, "series": null, "authors": null, "volume": null, "edition": null, "source": null, "year": "2012", "issue": null, "pages": {"start": "1", "end": "236"}, "location": null}, "text": "J.R.R. Tolkien, The Lord of the Rings, 2012"}], "year": null, "keywords": null, "issue": "4", "abstract": null}

## This is a classpath-based import file (this header is required)

transformer_export_researchinitiatives classpath eu/dnetlib/iis/transformers/export/researchinitiatives/oozie_app

<workflow-app xmlns="uri:oozie:workflow:0.2" name="test-transformers_export_researchinitiatives_sampledataproducer">

    <start to="producer"/>

    <action name="producer">

        <java>

            <job-tracker>${jobTracker}</job-tracker>

            <name-node>${nameNode}</name-node>

			<!-- The data generated by this node is deleted in this section -->

			<prepare>

				<delete path="${nameNode}${workingDir}/producer" />

				<mkdir path="${nameNode}${workingDir}/producer" />

			</prepare>

            <configuration>

                <property>

                    <name>mapred.job.queue.name</name>

                    <value>${queueName}</value>

                </property>

            </configuration>

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.Producer</arg>

            <arg>-C{document_to_concept_id,

				eu.dnetlib.iis.referenceextraction.researchinitiative.schemas.DocumentToConceptId,

				eu/dnetlib/iis/transformers/export/researchinitiatives/sampledataproducer/data/document_to_concept_id.json}</arg>

            <arg>-SworkingDir=${workingDir}/producer/working_dir</arg>

            <arg>-Odocument_to_concept_id=${workingDir}/producer/output</arg>

        </java>

        <ok to="transformer_export_researchinitiatives"/>

        <error to="fail"/>

    </action>

    <action name="transformer_export_researchinitiatives">

        <sub-workflow>

            <app-path>${wf:appPath()}/transformer_export_researchinitiatives</app-path>

            <configuration>

                <property>

                    <name>jobTracker</name>

                    <value>${jobTracker}</value>

                </property>

                <property>

                    <name>nameNode</name>

                    <value>${nameNode}</value>

                </property>

                <property>

                    <name>queueName</name>

                    <value>${queueName}</value>

                </property>

                <property>

                    <name>workingDir</name>

                    <value>${workingDir}/transformer_export_researchinitiatives/working_dir</value>

                </property>

                <property>

                    <name>input_document_to_research_initiative</name>

                    <value>${workingDir}/producer/output</value>

                </property>

                <property>

                    <name>output_document_to_research_initiatives</name>

                    <value>${workingDir}/transformer_export_researchinitiatives/output</value>

                </property>

            </configuration>

        </sub-workflow>

        <ok to="consumer"/>

        <error to="fail"/>

    </action>

    <action name="consumer">

		<java>

			<job-tracker>${jobTracker}</job-tracker>

			<name-node>${nameNode}</name-node>

			<configuration>

				<property>

					<name>mapred.job.queue.name</name>

					<value>${queueName}</value>

				</property>

			</configuration>

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.TestingConsumer</arg>

			<arg>-C{output,

				eu.dnetlib.iis.export.schemas.DocumentToConceptIds,

				eu/dnetlib/iis/transformers/export/researchinitiatives/sampledataproducer/data/document_to_concept_ids.json}</arg>

			<arg>-SworkingDir=${workingDir}/consumer/working_dir</arg>

			<arg>-Ioutput=${workingDir}/transformer_export_researchinitiatives/output</arg>

		</java>

		<ok to="end" />

		<error to="fail" />

	</action>    

    <kill name="fail">

		<message>Unfortunately, the workflow failed -- error message:

			[${wf:errorMessage(wf:lastErrorNode())}]</message>

    </kill>

    <end name="end"/>

</workflow-app>

{"documentId": "docId1", "conceptId": "concept1", "confidenceLevel": 0.9}

{"documentId": "docId1", "conceptId": "concept1", "confidenceLevel": null}

{"documentId": "docId2", "conceptId": "concept1", "confidenceLevel": 0.9}

{"documentId": "docId2", "conceptId": "concept2", "confidenceLevel": 0.8}

{"documentId": "docId3", "conceptId": "concept3", "confidenceLevel": 0.7}

{"documentId": "docId3", "conceptId": "concept2", "confidenceLevel": null}

{"documentId": "docId3", "conceptId": "concept1", "confidenceLevel": 0.7}

{"documentId": "docId1", "conceptIds": ["concept1"]}

{"documentId": "docId2", "conceptIds": ["concept1", "concept2"]}

{"documentId": "docId3", "conceptIds": ["concept1", "concept2", "concept3"]}

{"id": "id-2", "affiliations": []}

{"id": "id-3", "affiliations": [{"organization": "some-org", "countryName": "pl", "countryCode": "PL", "address": null, "rawText": "some raw text"}]}

{"publisher": "Tor Science Fiction", "affiliations": null, "authors": [{"authorFullName": "author-1", "affiliationPositions": null}, {"authorFullName": "author-2", "affiliationPositions": null}, {"authorFullName": "author-3", "affiliationPositions": null}], "language": "eng", "title": "Enders Game", "externalIdentifiers": null, "journal": "Journal-2", "id": "id-1", "pages": {"start": "123", "end": "128"}, "volume": null, "references": null, "year": null, "keywords": null, "issue": null, "abstract": null}

{"publisher": null, "affiliations": [], "authors": [{"authorFullName": "author-1", "affiliationPositions": null}, {"authorFullName": "author-2", "affiliationPositions": null}], "language": null, "title": null, "externalIdentifiers": {"id-1": "val-extr-1", "id-3": "val-extr-3"}, "journal": null, "id": "id-2", "pages": null, "volume": "124", "references": [{"position": 1, "basicMetadata": {"publisher": null, "title": "The Lord of the Rings", "url": null, "series": null, "authors": null, "volume": null, "edition": null, "source": null, "year": "2012", "issue": null, "pages": {"start": "1", "end": "236"}, "location": null}, "text": "J.R.R. Tolkien, The Lord of the Rings, 2012"}, {"position": 2, "basicMetadata": {"publisher": null, "title": "The Other Wind", "url": null, "series": null, "authors": null, "volume": "vol.23", "edition": null, "source": null, "year": "2003", "issue": null, "pages": null, "location": null}, "text": "Ursula K. Le Guin, The Other Wind, 2003"}], "year": 1970, "keywords": ["kwd_1", "kwd_3", "kwd_5"], "issue": null, "abstract": "The tales"}

{"publisher": "Harp3r T0rch", "affiliations": [{"organization": "some-org", "countryName": "pl", "countryCode": "PL", "address": null, "rawText": "some raw text"}], "authors": null, "language": "en", "title": "Small Gods", "externalIdentifiers": null, "journal": "Journal", "id": "id-3", "pages": null, "volume": "32", "references": [{"position": 1, "basicMetadata": {"publisher": null, "title": "The Lord of the Rings", "url": null, "series": null, "authors": null, "volume": null, "edition": null, "source": null, "year": "2012", "issue": null, "pages": {"start": "1", "end": "236"}, "location": null}, "text": "J.R.R. Tolkien, The Lord of the Rings, 2012"}], "year": null, "keywords": null, "issue": "4", "abstract": null}

## This is a classpath-based import file (this header is required)

transformer_export_documenttodataset_without_imported_data classpath eu/dnetlib/iis/transformers/export/documenttodataset_without_imported_data/oozie_app

<workflow-app xmlns="uri:oozie:workflow:0.2" name="test-transformers_export_documenttodataset_without_imported_data_sampledataproducer">

    <start to="producer"/>

    <action name="producer">

        <java>

            <job-tracker>${jobTracker}</job-tracker>

            <name-node>${nameNode}</name-node>

			<!-- The data generated by this node is deleted in this section -->

			<prepare>

				<delete path="${nameNode}${workingDir}/producer" />

				<mkdir path="${nameNode}${workingDir}/producer" />

			</prepare>

            <configuration>

                <property>

                    <name>mapred.job.queue.name</name>

                    <value>${queueName}</value>

                </property>

            </configuration>

            <!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.Producer</arg>

			<!-- Specification of the output ports -->

            <arg>-C{document_to_dataset,

				eu.dnetlib.iis.referenceextraction.dataset.schemas.DocumentToDataSet,

				eu/dnetlib/iis/transformers/export/documenttodataset_without_imported_data/sampledataproducer/data/document_to_dataset.json}</arg>

            <arg>-C{document_relation,

				eu.dnetlib.iis.importer.schemas.DocumentRelation,

				eu/dnetlib/iis/transformers/export/documenttodataset_without_imported_data/sampledataproducer/data/document_relation.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

            <arg>-SworkingDir=${workingDir}/producer/working_dir</arg>

            <arg>-Odocument_to_dataset=${workingDir}/producer/document_to_dataset</arg>

            <arg>-Odocument_relation=${workingDir}/producer/document_relation</arg>

        </java>

        <ok to="transformer_export_docdataset_without_imported"/>

        <error to="fail"/>

    </action>

    <action name="transformer_export_docdataset_without_imported">

        <sub-workflow>

            <app-path>${wf:appPath()}/transformer_export_documenttodataset_without_imported_data</app-path>

            <configuration>

                <property>

                    <name>jobTracker</name>

                    <value>${jobTracker}</value>

                </property>

                <property>

                    <name>nameNode</name>

                    <value>${nameNode}</value>

                </property>

                <property>

                    <name>queueName</name>

                    <value>${queueName}</value>

                </property>

                <!-- Working directory of the subworkflow -->

                <property>

                    <name>workingDir</name>

                    <value>${workingDir}/transformer_export_documenttodataset_without_imported_data/working_dir</value>

                </property>

                <!-- Input ports. -->

                <property>

                    <name>input_document_to_dataset</name>

                    <value>${workingDir}/producer/document_to_dataset</value>

                </property>

                <property>

                    <name>input_document_relation</name>

                    <value>${workingDir}/producer/document_relation</value>

                </property>

                <!-- Output port bound to given path -->

                <property>

                    <name>output_document_to_dataset</name>

                    <value>${workingDir}/transformer_export_documenttodataset_without_imported_data/document_to_dataset</value>

                </property>

            </configuration>

        </sub-workflow>

        <ok to="consumer"/>

        <error to="fail"/>

    </action>

    <action name="consumer">

		<java>

			<job-tracker>${jobTracker}</job-tracker>

			<name-node>${nameNode}</name-node>

			<configuration>

				<property>

					<name>mapred.job.queue.name</name>

					<value>${queueName}</value>

				</property>

			</configuration>

			<!-- This is simple wrapper for the Java code -->

			<main-class>eu.dnetlib.iis.core.java.ProcessWrapper</main-class>

			<!-- The business Java code that gets to be executed -->

			<arg>eu.dnetlib.iis.core.java.jsonworkflownodes.TestingConsumer</arg>

			<!-- Specification of the input ports -->

			<arg>-C{document_to_dataset,

				eu.dnetlib.iis.referenceextraction.dataset.schemas.DocumentToDataSet,

				eu/dnetlib/iis/transformers/export/documenttodataset_without_imported_data/sampledataproducer/data/output_document_to_dataset.json}</arg>

			<!-- All input and output ports have to be bound to paths in HDFS, working 

				directory has to be specified as well -->

			<arg>-SworkingDir=${workingDir}/consumer/working_dir</arg>

			<arg>-Idocument_to_dataset=${workingDir}/transformer_export_documenttodataset_without_imported_data/document_to_dataset</arg>

		</java>

		<ok to="end" />

		<error to="fail" />

	</action>    

    <kill name="fail">

		<message>Unfortunately, the workflow failed -- error message:

			[${wf:errorMessage(wf:lastErrorNode())}]</message>

    </kill>

    <end name="end"/>

</workflow-app>

{"id": "id-1", "referencedIds": []}

{"id": "id-3", "referencedIds": ["refid-4", "refid-2"]}

{"id": "id-2", "referencedIds": ["refid-1", "refid-4", "refid-2", "refid-48", "refid-32"]}

{"id": "id-4", "referencedIds": []}

{"id": "id-5", "referencedIds": ["refid-4"]}

{"id": "id-6", "referencedIds": ["refid-4", "refid-48", "refid-32"]}

{"documentId": "id-1", "datasetId": "8095", "confidenceLevel": null}

{"documentId": "id-2", "datasetId": "refid-48", "confidenceLevel": 1}

{"documentId": "id-2", "datasetId": "refid-14", "confidenceLevel": null}

{"documentId": "id-8", "datasetId": "0820", "confidenceLevel": null}

{"documentId": "id-1", "datasetId": "refid-32", "confidenceLevel": 1.1}

{"documentId": "id-2", "datasetId": "refid-32", "confidenceLevel": null}

{"documentId": "id-5", "datasetId": "refid-4", "confidenceLevel": 2}

{"documentId": "id-8", "datasetId": "5103", "confidenceLevel": 2}

{"documentId": "id-1", "datasetId": "5103", "confidenceLevel": null}

{"documentId": "id-2", "datasetId": "refid-148", "confidenceLevel": 0.9}

{"documentId": "id-1", "datasetId": "8095", "confidenceLevel": null}

{"documentId": "id-2", "datasetId": "refid-14", "confidenceLevel": null}

{"documentId": "id-8", "datasetId": "0820", "confidenceLevel": null}

{"documentId": "id-1", "datasetId": "refid-32", "confidenceLevel": 1.1}