[GraphAnnotation]Extention of bulktagging to include the easiest graph annotation patterns. Fixed issue and add test

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/SparkBulkTagJob.java

@@ -123,8 +123,8 @@ public class SparkBulkTagJob {
 					TaggingConstants.CLASS_NAME_BULKTAG_ORGANIZATION);
 				execEntityTag(
 					spark, inputPath + "project", outputPath + "project",
-						Utils.getCommunityProjects(baseURL), Project.class, TaggingConstants.CLASS_ID_PROJECT,
+					Utils.getCommunityProjects(baseURL), Project.class, TaggingConstants.CLASS_ID_PROJECT,
-						TaggingConstants.CLASS_NAME_BULKTAG_PROJECT);
+					TaggingConstants.CLASS_NAME_BULKTAG_PROJECT);
 				execDatasourceTag(spark, inputPath, outputPath, Utils.getDatasourceCommunities(baseURL));
 			});
 	}
@@ -290,6 +290,7 @@ public class SparkBulkTagJob {
 			.parallelStream()
 			.filter(ModelSupport::isResult)
 			.forEach(e -> {
+
 				removeOutputDir(spark, outputPath + e.name());
 				ResultTagger resultTagger = new ResultTagger();
 				Class<R> resultClazz = ModelSupport.entityTypes.get(e);

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/ResultTagger.java

@@ -10,6 +10,7 @@ import java.lang.reflect.Method;
 import java.util.*;
 import java.util.stream.Collectors;
 
+import com.jayway.jsonpath.Criteria;
 import org.apache.commons.lang3.StringUtils;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -19,13 +20,11 @@ import com.jayway.jsonpath.DocumentContext;
 import com.jayway.jsonpath.JsonPath;
 import com.jayway.jsonpath.PathNotFoundException;
 
-import eu.dnetlib.dhp.bulktag.Tagging;
 import eu.dnetlib.dhp.bulktag.actions.MapModel;
 import eu.dnetlib.dhp.bulktag.actions.Parameters;
 import eu.dnetlib.dhp.bulktag.eosc.EoscIFTag;
 import eu.dnetlib.dhp.schema.oaf.*;
 import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
-import scala.Tuple2;
 
 /** Created by miriam on 02/08/2018. */
 public class ResultTagger implements Serializable {
@@ -123,9 +122,11 @@ public class ResultTagger implements Serializable {
 
 //adding code for tagging of results searching supplementaryMaterial
 		final Set<String> tags = new HashSet<>();
+
+
 		taggingConstraints.getTags().forEach(t -> {
 			if (t.getCriteria().stream().anyMatch(crit -> crit.verifyCriteria(param)))
-				tags.add(t.getTagId());
+				tags.add(t.getId());
 		});
 
 		// communities contains all the communities to be not added to the context
@@ -262,7 +263,7 @@ public class ResultTagger implements Serializable {
 		tags.forEach(t -> {
 			Context con = new Context();
 			con.setId(t);
-			List<DataInfo> dataInfoList = Arrays
+			con.setDataInfo(Arrays
 				.asList(
 					OafMapperUtils
 						.dataInfo(
@@ -271,7 +272,7 @@ public class ResultTagger implements Serializable {
 								.qualifier(
 									CLASS_ID_ANNOTATION, CLASS_NAME_ANNOTATION, DNET_PROVENANCE_ACTIONS,
 									DNET_PROVENANCE_ACTIONS),
-							TAGGING_TRUST));
+							TAGGING_TRUST)));
 			result.getContext().add(con);
 		});
 

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/TaggingConstraint.java

@@ -1,14 +1,16 @@
 
 package eu.dnetlib.dhp.bulktag.community;
 
-public class TaggingConstraint extends SelectionConstraints {
+import java.io.Serializable;
-	private String tagId;
 
-	public String getTagId() {
+public class TaggingConstraint extends SelectionConstraints implements Serializable {
-		return tagId;
+	private String id;
+
+	public String getId() {
+		return id;
 	}
 
-	public void setTagId(String tagId) {
+	public void setId(String id) {
-		this.tagId = tagId;
+		this.id = id;
 	}
 }

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/TaggingConstraints.java

@@ -1,9 +1,10 @@
 
 package eu.dnetlib.dhp.bulktag.community;
 
+import java.io.Serializable;
 import java.util.List;
 
-public class TaggingConstraints {
+public class TaggingConstraints implements Serializable {
 	private List<TaggingConstraint> tags;
 
 	public List<TaggingConstraint> getTags() {

dhp-workflows/dhp-enrichment/src/main/resources/eu/dnetlib/dhp/wf/subworkflows/bulktag/job.properties

@@ -1,4 +1,4 @@
 sourcePath=/tmp/miriam/12_graph_copy
 pathMap=/data/bulktagging/pathMap
 baseURL=https://services.openaire.eu/openaire/community/
-taggingCriteria={"tags":[{"id":"SM","criteria":[{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary material for"},{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary document for"},{"verb":"starts_with_caseinsensitive","field":"title","value":"figure"},{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary figure"},{"verb":"starts_with_caseinsensitive","field":"title","value":"supplemental figure"},{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary table"},{"verb":"starts_with_caseinsensitive","field":"title","value":"table for"}]}]}]}
+taggingCriteria={"tags":[{"id":"SM","criteria":[{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary material for"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary document for"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"figure"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary figure"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplemental figure"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"supplementary table"}]},{"constraint":[{"verb":"starts_with_caseinsensitive","field":"title","value":"table for"}]}]}]}

dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/bulktag/BulkTagJobTest.java

@@ -33,6 +33,7 @@ import com.fasterxml.jackson.databind.ObjectMapper;
 import com.google.gson.Gson;
 
 import eu.dnetlib.dhp.bulktag.community.ProtoMap;
+import eu.dnetlib.dhp.bulktag.community.TaggingConstraints;
 import eu.dnetlib.dhp.schema.oaf.*;
 
 public class BulkTagJobTest {
@@ -68,7 +69,7 @@ public class BulkTagJobTest {
 
 	private static String taggingConf = "";
 
-	private static String taggingCriteria = "{\"criteria\":[{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplementary material\"}]}]}";
+	private static String taggingCriteria = "{\"tags\":[{\"id\":\"SM\",\"criteria\":[{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplementary material for\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplementary document for\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"figure\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplementary figure\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplemental figure\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"supplementary table\"}]},{\"constraint\":[{\"verb\":\"starts_with_caseinsensitive\",\"field\":\"title\",\"value\":\"table for\"}]}]}]}";
 
 	static {
 		try {
@@ -2011,6 +2012,10 @@ public class BulkTagJobTest {
 					"-nameNode", "local"
 				});
 
+		System.out.println("prrr");
+
 	}
 
+
+
 }