mergin with branch beta

CODE_OF_CONDUCT.md

@@ -0,0 +1,43 @@
+# Contributor Code of Conduct
+
+Openness, transparency and our community-driven participatory approach guide us in our day-to-day interactions and decision-making. Our open source projects are no exception. Trust, respect, collaboration and transparency are core values we believe should live and breathe within our projects. Our community welcomes participants from around the world with different experiences, unique perspectives, and great ideas to share.
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to making participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socio-economic status, nationality, personal appearance, race, religion, or sexual identity and orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment include:
+
+- Using welcoming and inclusive language
+- Being respectful of differing viewpoints and experiences
+- Gracefully accepting constructive criticism
+- Attempting collaboration before conflict
+- Focusing on what is best for the community
+- Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+- Violence, threats of violence, or inciting others to commit self-harm
+- The use of sexualized language or imagery and unwelcome sexual attention or advances
+- Trolling, intentionally spreading misinformation, insulting/derogatory comments, and personal or political attacks
+- Public or private harassment
+- Publishing others' private information, such as a physical or electronic address, without explicit permission
+- Abuse of the reporting process to intentionally harass or exclude others
+- Advocating for, or encouraging, any of the above behavior
+- Other conduct which could reasonably be considered inappropriate in a professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies both within project spaces and in public spaces when an individual is representing the project or its community. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant](https://www.contributor-covenant.org/), [version 1.4](https://www.contributor-covenant.org/version/1/4/code-of-conduct.html).

CONTRIBUTING.md

@@ -0,0 +1,10 @@
+# Contributing to D-Net Hadoop
+
+:+1::tada: First off, thanks for taking the time to contribute! :tada::+1:
+
+This project and everyone participating in it is governed by our [Code of Conduct](CODE_OF_CONDUCT.md). By participating, you are expected to uphold this code. Please report unacceptable behavior to [dnet-team@isti.cnr.it](mailto:dnet-team@isti.cnr.it).
+
+The following is a set of guidelines for contributing to this project and its packages. These are mostly guidelines, not rules, which applies to this project as a while, including all its sub-modules. 
+Use your best judgment, and feel free to propose changes to this document in a pull request.
+
+All contributions are welcome, all contributions will be considered to be contributed under the [project license](LICENSE.md).

README.md

@@ -2,6 +2,11 @@
 
 Dnet-hadoop is the project that defined all the [OOZIE workflows](https://oozie.apache.org/) for the OpenAIRE Graph construction, processing, provisioning.
 
+This project adheres to the Contributor Covenant [code of conduct](CODE_OF_CONDUCT.md).
+By participating, you are expected to uphold this code. Please report unacceptable behavior to [dnet-team@isti.cnr.it](mailto:dnet-team@isti.cnr.it).
+
+This project is licensed under the [AGPL v3 or later version](#LICENSE.md).
+
 How to build, package and run oozie workflows
 ====================
 

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/CategorySummary.java

@@ -0,0 +1,39 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+public class CategorySummary {
+
+	private String id;
+
+	private String label;
+
+	private boolean hasConcept;
+
+	public String getId() {
+		return id;
+	}
+
+	public String getLabel() {
+		return label;
+	}
+
+	public boolean isHasConcept() {
+		return hasConcept;
+	}
+
+	public CategorySummary setId(final String id) {
+		this.id = id;
+		return this;
+	}
+
+	public CategorySummary setLabel(final String label) {
+		this.label = label;
+		return this;
+	}
+
+	public CategorySummary setHasConcept(final boolean hasConcept) {
+		this.hasConcept = hasConcept;
+		return this;
+	}
+
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/CategorySummaryList.java

@@ -0,0 +1,7 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+import java.util.ArrayList;
+
+public class CategorySummaryList extends ArrayList<CategorySummary> {
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/ConceptSummary.java

@@ -0,0 +1,52 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+import java.util.List;
+
+public class ConceptSummary {
+
+	private String id;
+
+	private String label;
+
+	public boolean hasSubConcept;
+
+	private List<ConceptSummary> concepts;
+
+	public String getId() {
+		return id;
+	}
+
+	public String getLabel() {
+		return label;
+	}
+
+	public List<ConceptSummary> getConcepts() {
+		return concepts;
+	}
+
+	public ConceptSummary setId(final String id) {
+		this.id = id;
+		return this;
+	}
+
+	public ConceptSummary setLabel(final String label) {
+		this.label = label;
+		return this;
+	}
+
+	public boolean isHasSubConcept() {
+		return hasSubConcept;
+	}
+
+	public ConceptSummary setHasSubConcept(final boolean hasSubConcept) {
+		this.hasSubConcept = hasSubConcept;
+		return this;
+	}
+
+	public ConceptSummary setConcept(final List<ConceptSummary> concepts) {
+		this.concepts = concepts;
+		return this;
+	}
+
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/ConceptSummaryList.java

@@ -0,0 +1,7 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+import java.util.ArrayList;
+
+public class ConceptSummaryList extends ArrayList<ConceptSummary> {
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/ContextSummary.java

@@ -0,0 +1,50 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+public class ContextSummary {
+
+	private String id;
+
+	private String label;
+
+	private String type;
+
+	private String status;
+
+	public String getId() {
+		return id;
+	}
+
+	public String getLabel() {
+		return label;
+	}
+
+	public String getType() {
+		return type;
+	}
+
+	public String getStatus() {
+		return status;
+	}
+
+	public ContextSummary setId(final String id) {
+		this.id = id;
+		return this;
+	}
+
+	public ContextSummary setLabel(final String label) {
+		this.label = label;
+		return this;
+	}
+
+	public ContextSummary setType(final String type) {
+		this.type = type;
+		return this;
+	}
+
+	public ContextSummary setStatus(final String status) {
+		this.status = status;
+		return this;
+	}
+
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/api/context/ContextSummaryList.java

@@ -0,0 +1,7 @@
+
+package eu.dnetlib.dhp.common.api.context;
+
+import java.util.ArrayList;
+
+public class ContextSummaryList extends ArrayList<ContextSummary> {
+}

dhp-common/src/main/java/eu/dnetlib/dhp/common/collection/HttpConnector2.java

@@ -8,10 +8,13 @@ import java.io.InputStream;
 import java.net.*;
 import java.util.List;
 import java.util.Map;
+import java.util.concurrent.TimeUnit;
 
 import org.apache.commons.io.IOUtils;
 import org.apache.commons.lang3.math.NumberUtils;
+import org.apache.commons.lang3.time.DateUtils;
 import org.apache.http.HttpHeaders;
+import org.joda.time.Instant;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -94,14 +97,16 @@ public class HttpConnector2 {
 			throw new CollectorException(msg);
 		}
 
-		log.info("Request attempt {} [{}]", retryNumber, requestUrl);
-
 		InputStream input = null;
 
+		long start = System.currentTimeMillis();
 		try {
 			if (getClientParams().getRequestDelay() > 0) {
 				backoffAndSleep(getClientParams().getRequestDelay());
 			}
+
+			log.info("Request attempt {} [{}]", retryNumber, requestUrl);
+
 			final HttpURLConnection urlConn = (HttpURLConnection) new URL(requestUrl).openConnection();
 			urlConn.setInstanceFollowRedirects(false);
 			urlConn.setReadTimeout(getClientParams().getReadTimeOut() * 1000);
@@ -115,9 +120,8 @@ public class HttpConnector2 {
 					urlConn.addRequestProperty(headerEntry.getKey(), headerEntry.getValue());
 				}
 			}
-			if (log.isDebugEnabled()) {
+
-				logHeaderFields(urlConn);
+			logHeaderFields(urlConn);
-			}
 
 			int retryAfter = obtainRetryAfter(urlConn.getHeaderFields());
 			String rateLimit = urlConn.getHeaderField(Constants.HTTPHEADER_IETF_DRAFT_RATELIMIT_LIMIT);
@@ -132,9 +136,7 @@ public class HttpConnector2 {
 			}
 
 			if (is2xx(urlConn.getResponseCode())) {
-				input = urlConn.getInputStream();
+				return getInputStream(urlConn, start);
-				responseType = urlConn.getContentType();
-				return input;
 			}
 			if (is3xx(urlConn.getResponseCode())) {
 				// REDIRECTS
@@ -144,6 +146,7 @@ public class HttpConnector2 {
 					.put(
 						REPORT_PREFIX + urlConn.getResponseCode(),
 						String.format("Moved to: %s", newUrl));
+				logRequestTime(start);
 				urlConn.disconnect();
 				if (retryAfter > 0) {
 					backoffAndSleep(retryAfter);
@@ -159,26 +162,50 @@ public class HttpConnector2 {
 						if (retryAfter > 0) {
 							log
 								.warn(
-									"{} - waiting and repeating request after suggested retry-after {} sec.",
+									"waiting and repeating request after suggested retry-after {} sec for URL {}",
-									requestUrl, retryAfter);
+									retryAfter, requestUrl);
 							backoffAndSleep(retryAfter * 1000);
 						} else {
 							log
 								.warn(
-									"{} - waiting and repeating request after default delay of {} sec.",
+									"waiting and repeating request after default delay of {} sec for URL {}",
-									requestUrl, getClientParams().getRetryDelay());
+									getClientParams().getRetryDelay(), requestUrl);
-							backoffAndSleep(retryNumber * getClientParams().getRetryDelay() * 1000);
+							backoffAndSleep(retryNumber * getClientParams().getRetryDelay());
 						}
 						report.put(REPORT_PREFIX + urlConn.getResponseCode(), requestUrl);
+
+						logRequestTime(start);
+
 						urlConn.disconnect();
+
 						return attemptDownload(requestUrl, retryNumber + 1, report);
+					case 422: // UNPROCESSABLE ENTITY
+						report.put(REPORT_PREFIX + urlConn.getResponseCode(), requestUrl);
+						log.warn("waiting and repeating request after 10 sec for URL {}", requestUrl);
+						backoffAndSleep(10000);
+						urlConn.disconnect();
+						logRequestTime(start);
+						try {
+							return getInputStream(urlConn, start);
+						} catch (IOException e) {
+							log
+								.error(
+									"server returned 422 and got IOException accessing the response body from URL {}",
+									requestUrl);
+							log.error("IOException:", e);
+							return attemptDownload(requestUrl, retryNumber + 1, report);
+						}
 					default:
+						log.error("gor error {} from URL: {}", urlConn.getResponseCode(), urlConn.getURL());
+						log.error("response message: {}", urlConn.getResponseMessage());
 						report
 							.put(
 								REPORT_PREFIX + urlConn.getResponseCode(),
 								String
 									.format(
 										"%s Error: %s", requestUrl, urlConn.getResponseMessage()));
+						logRequestTime(start);
+						urlConn.disconnect();
 						throw new CollectorException(urlConn.getResponseCode() + " error " + report);
 				}
 			}
@@ -199,13 +226,27 @@ public class HttpConnector2 {
 		}
 	}
 
+	private InputStream getInputStream(HttpURLConnection urlConn, long start) throws IOException {
+		InputStream input = urlConn.getInputStream();
+		responseType = urlConn.getContentType();
+		logRequestTime(start);
+		return input;
+	}
+
+	private static void logRequestTime(long start) {
+		log
+			.info(
+				"request time elapsed: {}sec",
+				TimeUnit.MILLISECONDS.toSeconds(System.currentTimeMillis() - start));
+	}
+
 	private void logHeaderFields(final HttpURLConnection urlConn) throws IOException {
-		log.debug("StatusCode: {}", urlConn.getResponseMessage());
+		log.info("Response: {} - {}", urlConn.getResponseCode(), urlConn.getResponseMessage());
 
 		for (Map.Entry<String, List<String>> e : urlConn.getHeaderFields().entrySet()) {
 			if (e.getKey() != null) {
 				for (String v : e.getValue()) {
-					log.debug("  key: {} - value: {}", e.getKey(), v);
+					log.info("  key: {} - value: {}", e.getKey(), v);
 				}
 			}
 		}
@@ -225,7 +266,7 @@ public class HttpConnector2 {
 		for (String key : headerMap.keySet()) {
 			if ((key != null) && key.equalsIgnoreCase(HttpHeaders.RETRY_AFTER) && (!headerMap.get(key).isEmpty())
 				&& NumberUtils.isCreatable(headerMap.get(key).get(0))) {
-				return Integer.parseInt(headerMap.get(key).get(0)) + 10;
+				return Integer.parseInt(headerMap.get(key).get(0));
 			}
 		}
 		return -1;

dhp-common/src/main/java/eu/dnetlib/dhp/oozie/RunSQLSparkJob.java

@@ -0,0 +1,77 @@
+
+package eu.dnetlib.dhp.oozie;
+
+import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkHiveSession;
+
+import java.net.URL;
+import java.nio.charset.StandardCharsets;
+import java.util.HashMap;
+import java.util.Map;
+import java.util.Optional;
+
+import org.apache.commons.lang3.time.DurationFormatUtils;
+import org.apache.commons.text.StringSubstitutor;
+import org.apache.spark.SparkConf;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import com.google.common.io.Resources;
+
+import eu.dnetlib.dhp.application.ArgumentApplicationParser;
+
+public class RunSQLSparkJob {
+	private static final Logger log = LoggerFactory.getLogger(RunSQLSparkJob.class);
+
+	private final ArgumentApplicationParser parser;
+
+	public RunSQLSparkJob(ArgumentApplicationParser parser) {
+		this.parser = parser;
+	}
+
+	public static void main(String[] args) throws Exception {
+
+		Map<String, String> params = new HashMap<>();
+		for (int i = 0; i < args.length - 1; i++) {
+			if (args[i].startsWith("--")) {
+				params.put(args[i].substring(2), args[++i]);
+			}
+		}
+
+		/*
+		 * String jsonConfiguration = IOUtils .toString( Objects .requireNonNull( RunSQLSparkJob.class
+		 * .getResourceAsStream( "/eu/dnetlib/dhp/oozie/run_sql_parameters.json"))); final ArgumentApplicationParser
+		 * parser = new ArgumentApplicationParser(jsonConfiguration); parser.parseArgument(args);
+		 */
+
+		Boolean isSparkSessionManaged = Optional
+			.ofNullable(params.get("isSparkSessionManaged"))
+			.map(Boolean::valueOf)
+			.orElse(Boolean.TRUE);
+		log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
+
+		URL url = com.google.common.io.Resources.getResource(params.get("sql"));
+		String raw_sql = Resources.toString(url, StandardCharsets.UTF_8);
+
+		String sql = StringSubstitutor.replace(raw_sql, params);
+		log.info("sql: {}", sql);
+
+		SparkConf conf = new SparkConf();
+		conf.set("hive.metastore.uris", params.get("hiveMetastoreUris"));
+
+		runWithSparkHiveSession(
+			conf,
+			isSparkSessionManaged,
+			spark -> {
+				for (String statement : sql.split(";\\s*/\\*\\s*EOS\\s*\\*/\\s*")) {
+					log.info("executing: {}", statement);
+					long startTime = System.currentTimeMillis();
+					spark.sql(statement).show();
+					log
+						.info(
+							"executed in {}",
+							DurationFormatUtils.formatDuration(System.currentTimeMillis() - startTime, "HH:mm:ss.S"));
+				}
+			});
+	}
+
+}

dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/GraphCleaningFunctions.java

@@ -312,7 +312,8 @@ public class GraphCleaningFunctions extends CleaningFunctions {
 		}
 
 		if (value instanceof Datasource) {
-			// nothing to evaluate here
+			final Datasource d = (Datasource) value;
+			return Objects.nonNull(d.getOfficialname()) && StringUtils.isNotBlank(d.getOfficialname().getValue());
 		} else if (value instanceof Project) {
 			final Project p = (Project) value;
 			return Objects.nonNull(p.getCode()) && StringUtils.isNotBlank(p.getCode().getValue());

dhp-common/src/main/resources/eu/dnetlib/dhp/oozie/run_sql_parameters.json

@@ -0,0 +1,20 @@
+[
+  {
+    "paramName": "issm",
+    "paramLongName": "isSparkSessionManaged",
+    "paramDescription": "when true will stop SparkSession after job execution",
+    "paramRequired": false
+  },
+  {
+    "paramName": "hmu",
+    "paramLongName": "hiveMetastoreUris",
+    "paramDescription": "the hive metastore uris",
+    "paramRequired": true
+  },
+  {
+    "paramName": "sql",
+    "paramLongName": "sql",
+    "paramDescription": "sql script to execute",
+    "paramRequired": true
+  }
+]

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/AbstractClusteringFunction.java

@@ -14,9 +14,9 @@ import eu.dnetlib.pace.config.Config;
 
 public abstract class AbstractClusteringFunction extends AbstractPaceFunctions implements ClusteringFunction {
 
-	protected Map<String, Integer> params;
+	protected Map<String, Object> params;
 
-	public AbstractClusteringFunction(final Map<String, Integer> params) {
+	public AbstractClusteringFunction(final Map<String, Object> params) {
 		this.params = params;
 	}
 
@@ -27,7 +27,7 @@ public abstract class AbstractClusteringFunction extends AbstractPaceFunctions i
 		return fields
 			.stream()
 			.filter(f -> !f.isEmpty())
-			.map(this::normalize)
+			.map(s -> normalize(s))
 			.map(s -> filterAllStopWords(s))
 			.map(s -> doApply(conf, s))
 			.map(c -> filterBlacklisted(c, ngramBlacklist))
@@ -36,11 +36,24 @@ public abstract class AbstractClusteringFunction extends AbstractPaceFunctions i
 			.collect(Collectors.toCollection(HashSet::new));
 	}
 
-	public Map<String, Integer> getParams() {
+	public Map<String, Object> getParams() {
 		return params;
 	}
 
 	protected Integer param(String name) {
-		return params.get(name);
+		Object val = params.get(name);
+		if (val == null)
+			return null;
+		if (val instanceof Number) {
+			return ((Number) val).intValue();
+		}
+		return Integer.parseInt(val.toString());
+	}
+
+	protected int paramOrDefault(String name, int i) {
+		Integer res = param(name);
+		if (res == null)
+			res = i;
+		return res;
 	}
 }

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/Acronyms.java

@@ -13,7 +13,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("acronyms")
 public class Acronyms extends AbstractClusteringFunction {
 
-	public Acronyms(Map<String, Integer> params) {
+	public Acronyms(Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/ClusteringFunction.java

@@ -11,6 +11,6 @@ public interface ClusteringFunction {
 
 	public Collection<String> apply(Config config, List<String> fields);
 
-	public Map<String, Integer> getParams();
+	public Map<String, Object> getParams();
 
 }

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/ImmutableFieldValue.java

@@ -12,7 +12,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("immutablefieldvalue")
 public class ImmutableFieldValue extends AbstractClusteringFunction {
 
-	public ImmutableFieldValue(final Map<String, Integer> params) {
+	public ImmutableFieldValue(final Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/JSONListClustering.java

@@ -0,0 +1,69 @@
+
+package eu.dnetlib.pace.clustering;
+
+import java.util.Collection;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Map;
+import java.util.stream.Collectors;
+
+import org.apache.commons.lang3.StringUtils;
+
+import com.jayway.jsonpath.Configuration;
+import com.jayway.jsonpath.DocumentContext;
+import com.jayway.jsonpath.JsonPath;
+import com.jayway.jsonpath.Option;
+
+import eu.dnetlib.pace.common.AbstractPaceFunctions;
+import eu.dnetlib.pace.config.Config;
+import eu.dnetlib.pace.util.MapDocumentUtil;
+
+@ClusteringClass("jsonlistclustering")
+public class JSONListClustering extends AbstractPaceFunctions implements ClusteringFunction {
+
+	private Map<String, Object> params;
+
+	public JSONListClustering(Map<String, Object> params) {
+		this.params = params;
+	}
+
+	@Override
+	public Map<String, Object> getParams() {
+		return params;
+	}
+
+	@Override
+	public Collection<String> apply(Config conf, List<String> fields) {
+		return fields
+			.stream()
+			.filter(f -> !f.isEmpty())
+			.map(s -> doApply(conf, s))
+			.filter(StringUtils::isNotBlank)
+			.collect(Collectors.toCollection(HashSet::new));
+	}
+
+	private String doApply(Config conf, String json) {
+		StringBuilder st = new StringBuilder(); // to build the string used for comparisons basing on the jpath into
+		// parameters
+		final DocumentContext documentContext = JsonPath
+			.using(Configuration.defaultConfiguration().addOptions(Option.SUPPRESS_EXCEPTIONS))
+			.parse(json);
+
+		// for each path in the param list
+		for (String key : params.keySet().stream().filter(k -> k.contains("jpath")).collect(Collectors.toList())) {
+			String path = params.get(key).toString();
+			String value = MapDocumentUtil.getJPathString(path, documentContext);
+			if (value == null || value.isEmpty())
+				value = "";
+			st.append(value);
+			st.append(" ");
+		}
+
+		st.setLength(st.length() - 1);
+
+		if (StringUtils.isBlank(st)) {
+			return "1";
+		}
+		return st.toString();
+	}
+}

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/KeywordsClustering.java

@@ -11,7 +11,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("keywordsclustering")
 public class KeywordsClustering extends AbstractClusteringFunction {
 
-	public KeywordsClustering(Map<String, Integer> params) {
+	public KeywordsClustering(Map<String, Object> params) {
 		super(params);
 	}
 
@@ -19,8 +19,8 @@ public class KeywordsClustering extends AbstractClusteringFunction {
 	protected Collection<String> doApply(final Config conf, String s) {
 
 		// takes city codes and keywords codes without duplicates
-		Set<String> keywords = getKeywords(s, conf.translationMap(), params.getOrDefault("windowSize", 4));
+		Set<String> keywords = getKeywords(s, conf.translationMap(), paramOrDefault("windowSize", 4));
-		Set<String> cities = getCities(s, params.getOrDefault("windowSize", 4));
+		Set<String> cities = getCities(s, paramOrDefault("windowSize", 4));
 
 		// list of combination to return as result
 		final Collection<String> combinations = new LinkedHashSet<String>();
@@ -28,7 +28,7 @@ public class KeywordsClustering extends AbstractClusteringFunction {
 		for (String keyword : keywordsToCodes(keywords, conf.translationMap())) {
 			for (String city : citiesToCodes(cities)) {
 				combinations.add(keyword + "-" + city);
-				if (combinations.size() >= params.getOrDefault("max", 2)) {
+				if (combinations.size() >= paramOrDefault("max", 2)) {
 					return combinations;
 				}
 			}
@@ -42,8 +42,8 @@ public class KeywordsClustering extends AbstractClusteringFunction {
 		return fields
 			.stream()
 			.filter(f -> !f.isEmpty())
-			.map(this::cleanup)
+			.map(KeywordsClustering::cleanup)
-			.map(this::normalize)
+			.map(KeywordsClustering::normalize)
 			.map(s -> filterAllStopWords(s))
 			.map(s -> doApply(conf, s))
 			.map(c -> filterBlacklisted(c, ngramBlacklist))

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/LastNameFirstInitial.java

@@ -16,7 +16,7 @@ public class LastNameFirstInitial extends AbstractClusteringFunction {
 
 	private boolean DEFAULT_AGGRESSIVE = true;
 
-	public LastNameFirstInitial(final Map<String, Integer> params) {
+	public LastNameFirstInitial(final Map<String, Object> params) {
 		super(params);
 	}
 
@@ -25,7 +25,7 @@ public class LastNameFirstInitial extends AbstractClusteringFunction {
 		return fields
 			.stream()
 			.filter(f -> !f.isEmpty())
-			.map(this::normalize)
+			.map(LastNameFirstInitial::normalize)
 			.map(s -> doApply(conf, s))
 			.map(c -> filterBlacklisted(c, ngramBlacklist))
 			.flatMap(c -> c.stream())
@@ -33,8 +33,7 @@ public class LastNameFirstInitial extends AbstractClusteringFunction {
 			.collect(Collectors.toCollection(HashSet::new));
 	}
 
-	@Override
+	public static String normalize(final String s) {
-	protected String normalize(final String s) {
 		return fixAliases(transliterate(nfd(unicodeNormalization(s))))
 			// do not compact the regexes in a single expression, would cause StackOverflowError in case of large input
 			// strings

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/LowercaseClustering.java

@@ -15,7 +15,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("lowercase")
 public class LowercaseClustering extends AbstractClusteringFunction {
 
-	public LowercaseClustering(final Map<String, Integer> params) {
+	public LowercaseClustering(final Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/NgramPairs.java

@@ -12,11 +12,11 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("ngrampairs")
 public class NgramPairs extends Ngrams {
 
-	public NgramPairs(Map<String, Integer> params) {
+	public NgramPairs(Map<String, Object> params) {
 		super(params, false);
 	}
 
-	public NgramPairs(Map<String, Integer> params, boolean sorted) {
+	public NgramPairs(Map<String, Object> params, boolean sorted) {
 		super(params, sorted);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/Ngrams.java

@@ -10,11 +10,11 @@ public class Ngrams extends AbstractClusteringFunction {
 
 	private final boolean sorted;
 
-	public Ngrams(Map<String, Integer> params) {
+	public Ngrams(Map<String, Object> params) {
 		this(params, false);
 	}
 
-	public Ngrams(Map<String, Integer> params, boolean sorted) {
+	public Ngrams(Map<String, Object> params, boolean sorted) {
 		super(params);
 		this.sorted = sorted;
 	}

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/NumAuthorsTitleSuffixPrefixChain.java

@@ -0,0 +1,113 @@
+
+package eu.dnetlib.pace.clustering;
+
+import java.util.*;
+import java.util.stream.Collectors;
+import java.util.stream.StreamSupport;
+
+import com.google.common.base.Splitter;
+import com.google.common.collect.Sets;
+
+import eu.dnetlib.pace.config.Config;
+
+@ClusteringClass("numAuthorsTitleSuffixPrefixChain")
+public class NumAuthorsTitleSuffixPrefixChain extends AbstractClusteringFunction {
+
+	public NumAuthorsTitleSuffixPrefixChain(Map<String, Object> params) {
+		super(params);
+	}
+
+	@Override
+	public Collection<String> apply(Config conf, List<String> fields) {
+
+		try {
+			int num_authors = Math.min(Integer.parseInt(fields.get(0)), 21); // SIZE threshold is 20, +1
+
+			if (num_authors > 0) {
+				return super.apply(conf, fields.subList(1, fields.size()))
+					.stream()
+					.map(s -> num_authors + "-" + s)
+					.collect(Collectors.toList());
+			}
+		} catch (NumberFormatException e) {
+			// missing or null authors array
+		}
+
+		return Collections.emptyList();
+	}
+
+	@Override
+	protected Collection<String> doApply(Config conf, String s) {
+		return suffixPrefixChain(cleanup(s), param("mod"));
+	}
+
+	private Collection<String> suffixPrefixChain(String s, int mod) {
+		// create the list of words from the string (remove short words)
+		List<String> wordsList = Arrays
+			.stream(s.split(" "))
+			.filter(si -> si.length() > 3)
+			.collect(Collectors.toList());
+
+		final int words = wordsList.size();
+		final int letters = s.length();
+
+		// create the prefix: number of words + number of letters/mod
+		String prefix = words / mod + "-";
+
+		return doSuffixPrefixChain(wordsList, prefix);
+
+	}
+
+	private Collection<String> doSuffixPrefixChain(List<String> wordsList, String prefix) {
+
+		Set<String> set = Sets.newLinkedHashSet();
+		switch (wordsList.size()) {
+			case 0:
+				break;
+			case 1:
+				set.add(wordsList.get(0));
+				break;
+			case 2:
+				set
+					.add(
+						prefix +
+							suffix(wordsList.get(0), 3) +
+							prefix(wordsList.get(1), 3));
+
+				set
+					.add(
+						prefix +
+							prefix(wordsList.get(0), 3) +
+							suffix(wordsList.get(1), 3));
+
+				break;
+			default:
+				set
+					.add(
+						prefix +
+							suffix(wordsList.get(0), 3) +
+							prefix(wordsList.get(1), 3) +
+							suffix(wordsList.get(2), 3));
+
+				set
+					.add(
+						prefix +
+							prefix(wordsList.get(0), 3) +
+							suffix(wordsList.get(1), 3) +
+							prefix(wordsList.get(2), 3));
+				break;
+		}
+
+		return set;
+
+	}
+
+	private String suffix(String s, int len) {
+		return s.substring(s.length() - len);
+	}
+
+	private String prefix(String s, int len) {
+		return s.substring(0, len);
+	}
+
+}

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/PersonClustering.java

@@ -17,11 +17,11 @@ import eu.dnetlib.pace.model.Person;
 @ClusteringClass("personClustering")
 public class PersonClustering extends AbstractPaceFunctions implements ClusteringFunction {
 
-	private Map<String, Integer> params;
+	private Map<String, Object> params;
 
 	private static final int MAX_TOKENS = 5;
 
-	public PersonClustering(final Map<String, Integer> params) {
+	public PersonClustering(final Map<String, Object> params) {
 		this.params = params;
 	}
 
@@ -77,7 +77,7 @@ public class PersonClustering extends AbstractPaceFunctions implements Clusterin
 //	}
 
 	@Override
-	public Map<String, Integer> getParams() {
+	public Map<String, Object> getParams() {
 		return params;
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/PersonHash.java

@@ -15,7 +15,7 @@ public class PersonHash extends AbstractClusteringFunction {
 
 	private boolean DEFAULT_AGGRESSIVE = false;
 
-	public PersonHash(final Map<String, Integer> params) {
+	public PersonHash(final Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/RandomClusteringFunction.java

@@ -8,7 +8,7 @@ import eu.dnetlib.pace.config.Config;
 
 public class RandomClusteringFunction extends AbstractClusteringFunction {
 
-	public RandomClusteringFunction(Map<String, Integer> params) {
+	public RandomClusteringFunction(Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/SortedNgramPairs.java

@@ -1,7 +1,10 @@
 
 package eu.dnetlib.pace.clustering;
 
-import java.util.*;
+import java.util.Collection;
+import java.util.Collections;
+import java.util.List;
+import java.util.Map;
 
 import com.google.common.base.Joiner;
 import com.google.common.base.Splitter;
@@ -12,7 +15,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("sortedngrampairs")
 public class SortedNgramPairs extends NgramPairs {
 
-	public SortedNgramPairs(Map<String, Integer> params) {
+	public SortedNgramPairs(Map<String, Object> params) {
 		super(params, false);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/SpaceTrimmingFieldValue.java

@@ -15,7 +15,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("spacetrimmingfieldvalue")
 public class SpaceTrimmingFieldValue extends AbstractClusteringFunction {
 
-	public SpaceTrimmingFieldValue(final Map<String, Integer> params) {
+	public SpaceTrimmingFieldValue(final Map<String, Object> params) {
 		super(params);
 	}
 
@@ -25,7 +25,7 @@ public class SpaceTrimmingFieldValue extends AbstractClusteringFunction {
 
 		res
 			.add(
-				StringUtils.isBlank(s) ? RandomStringUtils.random(getParams().get("randomLength"))
+				StringUtils.isBlank(s) ? RandomStringUtils.random(param("randomLength"))
 					: s.toLowerCase().replaceAll("\\s+", ""));
 
 		return res;

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/SuffixPrefix.java

@@ -12,7 +12,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("suffixprefix")
 public class SuffixPrefix extends AbstractClusteringFunction {
 
-	public SuffixPrefix(Map<String, Integer> params) {
+	public SuffixPrefix(Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/UrlClustering.java

@@ -15,12 +15,17 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("urlclustering")
 public class UrlClustering extends AbstractPaceFunctions implements ClusteringFunction {
 
-	protected Map<String, Integer> params;
+	protected Map<String, Object> params;
 
-	public UrlClustering(final Map<String, Integer> params) {
+	public UrlClustering(final Map<String, Object> params) {
 		this.params = params;
 	}
 
+	@Override
+	public Map<String, Object> getParams() {
+		return params;
+	}
+
 	@Override
 	public Collection<String> apply(final Config conf, List<String> fields) {
 		try {
@@ -35,11 +40,6 @@ public class UrlClustering extends AbstractPaceFunctions implements ClusteringFu
 		}
 	}
 
-	@Override
-	public Map<String, Integer> getParams() {
-		return null;
-	}
-
 	private URL asUrl(String value) {
 		try {
 			return new URL(value);

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/WordsStatsSuffixPrefixChain.java

@@ -11,7 +11,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("wordsStatsSuffixPrefixChain")
 public class WordsStatsSuffixPrefixChain extends AbstractClusteringFunction {
 
-	public WordsStatsSuffixPrefixChain(Map<String, Integer> params) {
+	public WordsStatsSuffixPrefixChain(Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/clustering/WordsSuffixPrefix.java

@@ -12,7 +12,7 @@ import eu.dnetlib.pace.config.Config;
 @ClusteringClass("wordssuffixprefix")
 public class WordsSuffixPrefix extends AbstractClusteringFunction {
 
-	public WordsSuffixPrefix(Map<String, Integer> params) {
+	public WordsSuffixPrefix(Map<String, Object> params) {
 		super(params);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/common/AbstractPaceFunctions.java

@@ -16,7 +16,6 @@ import org.apache.commons.lang3.StringUtils;
 import com.google.common.base.Joiner;
 import com.google.common.base.Splitter;
 import com.google.common.collect.Iterables;
-import com.google.common.collect.Lists;
 import com.google.common.collect.Sets;
 import com.ibm.icu.text.Transliterator;
 
@@ -27,7 +26,7 @@ import eu.dnetlib.pace.clustering.NGramUtils;
  *
  * @author claudio
  */
-public abstract class AbstractPaceFunctions {
+public class AbstractPaceFunctions {
 
 	// city map to be used when translating the city names into codes
 	private static Map<String, String> cityMap = AbstractPaceFunctions
@@ -62,11 +61,14 @@ public abstract class AbstractPaceFunctions {
 
 	private static Pattern hexUnicodePattern = Pattern.compile("\\\\u(\\p{XDigit}{4})");
 
-	protected String concat(final List<String> l) {
+	private static Pattern romanNumberPattern = Pattern
+		.compile("^M{0,4}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})$");
+
+	protected static String concat(final List<String> l) {
 		return Joiner.on(" ").skipNulls().join(l);
 	}
 
-	protected String cleanup(final String s) {
+	public static String cleanup(final String s) {
 		final String s1 = HTML_REGEX.matcher(s).replaceAll("");
 		final String s2 = unicodeNormalization(s1.toLowerCase());
 		final String s3 = nfd(s2);
@@ -82,7 +84,7 @@ public abstract class AbstractPaceFunctions {
 		return s12;
 	}
 
-	protected String fixXML(final String a) {
+	protected static String fixXML(final String a) {
 
 		return a
 			.replaceAll("&ndash;", " ")
@@ -91,7 +93,7 @@ public abstract class AbstractPaceFunctions {
 			.replaceAll("&minus;", " ");
 	}
 
-	protected boolean checkNumbers(final String a, final String b) {
+	protected static boolean checkNumbers(final String a, final String b) {
 		final String numbersA = getNumbers(a);
 		final String numbersB = getNumbers(b);
 		final String romansA = getRomans(a);
@@ -99,7 +101,7 @@ public abstract class AbstractPaceFunctions {
 		return !numbersA.equals(numbersB) || !romansA.equals(romansB);
 	}
 
-	protected String getRomans(final String s) {
+	protected static String getRomans(final String s) {
 		final StringBuilder sb = new StringBuilder();
 		for (final String t : s.split(" ")) {
 			sb.append(isRoman(t) ? t : "");
@@ -107,13 +109,12 @@ public abstract class AbstractPaceFunctions {
 		return sb.toString();
 	}
 
-	protected boolean isRoman(final String s) {
+	protected static boolean isRoman(final String s) {
-		return s
+		Matcher m = romanNumberPattern.matcher(s);
-			.replaceAll("^M{0,4}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})$", "qwertyuiop")
+		return m.matches() && m.hitEnd();
-			.equals("qwertyuiop");
 	}
 
-	protected String getNumbers(final String s) {
+	protected static String getNumbers(final String s) {
 		final StringBuilder sb = new StringBuilder();
 		for (final String t : s.split(" ")) {
 			sb.append(isNumber(t) ? t : "");
@@ -121,7 +122,7 @@ public abstract class AbstractPaceFunctions {
 		return sb.toString();
 	}
 
-	public boolean isNumber(String strNum) {
+	public static boolean isNumber(String strNum) {
 		if (strNum == null) {
 			return false;
 		}
@@ -147,7 +148,7 @@ public abstract class AbstractPaceFunctions {
 		}
 	}
 
-	protected String removeSymbols(final String s) {
+	protected static String removeSymbols(final String s) {
 		final StringBuilder sb = new StringBuilder();
 
 		s.chars().forEach(ch -> {
@@ -157,11 +158,11 @@ public abstract class AbstractPaceFunctions {
 		return sb.toString().replaceAll("\\s+", " ");
 	}
 
-	protected boolean notNull(final String s) {
+	protected static boolean notNull(final String s) {
 		return s != null;
 	}
 
-	protected String normalize(final String s) {
+	public static String normalize(final String s) {
 		return fixAliases(transliterate(nfd(unicodeNormalization(s))))
 			.toLowerCase()
 			// do not compact the regexes in a single expression, would cause StackOverflowError in case of large input
@@ -174,16 +175,16 @@ public abstract class AbstractPaceFunctions {
 			.trim();
 	}
 
-	public String nfd(final String s) {
+	public static String nfd(final String s) {
 		return Normalizer.normalize(s, Normalizer.Form.NFD);
 	}
 
-	public String utf8(final String s) {
+	public static String utf8(final String s) {
 		byte[] bytes = s.getBytes(StandardCharsets.UTF_8);
 		return new String(bytes, StandardCharsets.UTF_8);
 	}
 
-	public String unicodeNormalization(final String s) {
+	public static String unicodeNormalization(final String s) {
 
 		Matcher m = hexUnicodePattern.matcher(s);
 		StringBuffer buf = new StringBuffer(s.length());
@@ -195,7 +196,7 @@ public abstract class AbstractPaceFunctions {
 		return buf.toString();
 	}
 
-	protected String filterStopWords(final String s, final Set<String> stopwords) {
+	protected static String filterStopWords(final String s, final Set<String> stopwords) {
 		final StringTokenizer st = new StringTokenizer(s);
 		final StringBuilder sb = new StringBuilder();
 		while (st.hasMoreTokens()) {
@@ -208,7 +209,7 @@ public abstract class AbstractPaceFunctions {
 		return sb.toString().trim();
 	}
 
-	public String filterAllStopWords(String s) {
+	public static String filterAllStopWords(String s) {
 
 		s = filterStopWords(s, stopwords_en);
 		s = filterStopWords(s, stopwords_de);
@@ -221,7 +222,8 @@ public abstract class AbstractPaceFunctions {
 		return s;
 	}
 
-	protected Collection<String> filterBlacklisted(final Collection<String> set, final Set<String> ngramBlacklist) {
+	protected static Collection<String> filterBlacklisted(final Collection<String> set,
+		final Set<String> ngramBlacklist) {
 		final Set<String> newset = Sets.newLinkedHashSet();
 		for (final String s : set) {
 			if (!ngramBlacklist.contains(s)) {
@@ -268,7 +270,7 @@ public abstract class AbstractPaceFunctions {
 		return m;
 	}
 
-	public String removeKeywords(String s, Set<String> keywords) {
+	public static String removeKeywords(String s, Set<String> keywords) {
 
 		s = " " + s + " ";
 		for (String k : keywords) {
@@ -278,39 +280,39 @@ public abstract class AbstractPaceFunctions {
 		return s.trim();
 	}
 
-	public double commonElementsPercentage(Set<String> s1, Set<String> s2) {
+	public static double commonElementsPercentage(Set<String> s1, Set<String> s2) {
 
 		double longer = Math.max(s1.size(), s2.size());
 		return (double) s1.stream().filter(s2::contains).count() / longer;
 	}
 
 	// convert the set of keywords to codes
-	public Set<String> toCodes(Set<String> keywords, Map<String, String> translationMap) {
+	public static Set<String> toCodes(Set<String> keywords, Map<String, String> translationMap) {
 		return keywords.stream().map(s -> translationMap.get(s)).collect(Collectors.toSet());
 	}
 
-	public Set<String> keywordsToCodes(Set<String> keywords, Map<String, String> translationMap) {
+	public static Set<String> keywordsToCodes(Set<String> keywords, Map<String, String> translationMap) {
 		return toCodes(keywords, translationMap);
 	}
 
-	public Set<String> citiesToCodes(Set<String> keywords) {
+	public static Set<String> citiesToCodes(Set<String> keywords) {
 		return toCodes(keywords, cityMap);
 	}
 
-	protected String firstLC(final String s) {
+	protected static String firstLC(final String s) {
 		return StringUtils.substring(s, 0, 1).toLowerCase();
 	}
 
-	protected Iterable<String> tokens(final String s, final int maxTokens) {
+	protected static Iterable<String> tokens(final String s, final int maxTokens) {
 		return Iterables.limit(Splitter.on(" ").omitEmptyStrings().trimResults().split(s), maxTokens);
 	}
 
-	public String normalizePid(String pid) {
+	public static String normalizePid(String pid) {
 		return DOI_PREFIX.matcher(pid.toLowerCase()).replaceAll("");
 	}
 
 	// get the list of keywords into the input string
-	public Set<String> getKeywords(String s1, Map<String, String> translationMap, int windowSize) {
+	public static Set<String> getKeywords(String s1, Map<String, String> translationMap, int windowSize) {
 
 		String s = s1;
 
@@ -340,7 +342,7 @@ public abstract class AbstractPaceFunctions {
 		return codes;
 	}
 
-	public Set<String> getCities(String s1, int windowSize) {
+	public static Set<String> getCities(String s1, int windowSize) {
 		return getKeywords(s1, cityMap, windowSize);
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/model/ClusteringDef.java

@@ -18,7 +18,7 @@ public class ClusteringDef implements Serializable {
 
 	private List<String> fields;
 
-	private Map<String, Integer> params;
+	private Map<String, Object> params;
 
 	public ClusteringDef() {
 	}
@@ -43,11 +43,11 @@ public class ClusteringDef implements Serializable {
 		this.fields = fields;
 	}
 
-	public Map<String, Integer> getParams() {
+	public Map<String, Object> getParams() {
 		return params;
 	}
 
-	public void setParams(final Map<String, Integer> params) {
+	public void setParams(final Map<String, Object> params) {
 		this.params = params;
 	}
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/model/FieldDef.java

@@ -2,6 +2,7 @@
 package eu.dnetlib.pace.model;
 
 import java.io.Serializable;
+import java.util.HashSet;
 import java.util.List;
 
 import com.fasterxml.jackson.core.JsonProcessingException;
@@ -36,6 +37,16 @@ public class FieldDef implements Serializable {
 	 */
 	private int length = -1;
 
+	private HashSet<String> filter;
+
+	private boolean sorted;
+
+	public boolean isSorted() {
+		return sorted;
+	}
+
+	private String clean;
+
 	public FieldDef() {
 	}
 
@@ -91,6 +102,30 @@ public class FieldDef implements Serializable {
 		this.path = path;
 	}
 
+	public HashSet<String> getFilter() {
+		return filter;
+	}
+
+	public void setFilter(HashSet<String> filter) {
+		this.filter = filter;
+	}
+
+	public boolean getSorted() {
+		return sorted;
+	}
+
+	public void setSorted(boolean sorted) {
+		this.sorted = sorted;
+	}
+
+	public String getClean() {
+		return clean;
+	}
+
+	public void setClean(String clean) {
+		this.clean = clean;
+	}
+
 	@Override
 	public String toString() {
 		try {

dhp-pace-core/src/main/java/eu/dnetlib/pace/model/SparkDeduper.scala

@@ -5,9 +5,9 @@ import eu.dnetlib.pace.util.{BlockProcessor, SparkReporter}
 import org.apache.spark.SparkContext
 import org.apache.spark.sql.catalyst.expressions.Literal
 import org.apache.spark.sql.expressions._
-import org.apache.spark.sql.functions.{col, lit, udf}
+import org.apache.spark.sql.functions.{col, desc, expr, lit, udf}
 import org.apache.spark.sql.types._
-import org.apache.spark.sql.{Column, Dataset, Row, functions}
+import org.apache.spark.sql.{Column, Dataset, Row, SaveMode, functions}
 
 import java.util.function.Predicate
 import java.util.stream.Collectors
@@ -80,6 +80,8 @@ case class SparkDeduper(conf: DedupConfig) extends Serializable {
         .withColumn("key", functions.explode(clusterValuesUDF(cd).apply(functions.array(inputColumns: _*))))
         // Add position column having the position of the row within the set of rows having the same key value ordered by the sorting value
         .withColumn("position", functions.row_number().over(Window.partitionBy("key").orderBy(col(model.orderingFieldName), col(model.identifierFieldName))))
+       // .withColumn("count", functions.max("position").over(Window.partitionBy("key").orderBy(col(model.orderingFieldName), col(model.identifierFieldName)).rowsBetween(Window.unboundedPreceding,Window.unboundedFollowing) ))
+       // .filter("count > 1")
 
       if (df_with_clustering_keys == null)
         df_with_clustering_keys = ds
@@ -88,20 +90,44 @@ case class SparkDeduper(conf: DedupConfig) extends Serializable {
     }
 
     //TODO: analytics
+    /*df_with_clustering_keys.groupBy(col("clustering"), col("key"))
+      .agg(expr("max(count) AS size"))
+      .orderBy(desc("size"))
+      .show*/
 
     val df_with_blocks = df_with_clustering_keys
-      // filter out rows with position exceeding the maxqueuesize parameter
+      // split the clustering block into smaller blocks of queuemaxsize
-      .filter(col("position").leq(conf.getWf.getQueueMaxSize))
+      .groupBy(col("clustering"), col("key"), functions.floor(col("position").divide(lit(conf.getWf.getQueueMaxSize))))
-      .groupBy("clustering", "key")
       .agg(functions.collect_set(functions.struct(model.schema.fieldNames.map(col): _*)).as("block"))
       .filter(functions.size(new Column("block")).gt(1))
+       .union(
+        //adjacency blocks
+        df_with_clustering_keys
+          // filter out leading and trailing elements
+          .filter(col("position").gt(conf.getWf.getSlidingWindowSize/2))
+          //.filter(col("position").lt(col("count").minus(conf.getWf.getSlidingWindowSize/2)))
+          // create small blocks of records on "the border" of maxqueuesize: getSlidingWindowSize/2 elements before and after
+          .filter(
+            col("position").mod(conf.getWf.getQueueMaxSize).lt(conf.getWf.getSlidingWindowSize/2) // slice of the start of block
+            || col("position").mod(conf.getWf.getQueueMaxSize).gt(conf.getWf.getQueueMaxSize - (conf.getWf.getSlidingWindowSize/2)) //slice of the end of the block
+          )
+          .groupBy(col("clustering"), col("key"), functions.floor((col("position") + lit(conf.getWf.getSlidingWindowSize/2)).divide(lit(conf.getWf.getQueueMaxSize))))
+          .agg(functions.collect_set(functions.struct(model.schema.fieldNames.map(col): _*)).as("block"))
+          .filter(functions.size(new Column("block")).gt(1))
+      )
 
     df_with_blocks
   }
 
   def clusterValuesUDF(cd: ClusteringDef) = {
     udf[mutable.WrappedArray[String], mutable.WrappedArray[Any]](values => {
-      values.flatMap(f => cd.clusteringFunction().apply(conf, Seq(f.toString).asJava).asScala)
+      val valueList = values.flatMap {
+        case a: mutable.WrappedArray[Any] => a.map(_.toString)
+        case s: Any => Seq(s.toString)
+      }.asJava;
+
+      mutable.WrappedArray.make(cd.clusteringFunction().apply(conf, valueList).toArray())
+
     })
   }
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/model/SparkModel.scala

@@ -1,13 +1,16 @@
 package eu.dnetlib.pace.model
 
 import com.jayway.jsonpath.{Configuration, JsonPath}
+import eu.dnetlib.pace.common.AbstractPaceFunctions
 import eu.dnetlib.pace.config.{DedupConfig, Type}
 import eu.dnetlib.pace.util.MapDocumentUtil
+import org.apache.commons.lang3.StringUtils
 import org.apache.spark.sql.catalyst.encoders.RowEncoder
 import org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema
 import org.apache.spark.sql.types.{DataTypes, Metadata, StructField, StructType}
 import org.apache.spark.sql.{Dataset, Row}
 
+import java.util.Locale
 import java.util.regex.Pattern
 import scala.collection.JavaConverters._
 
@@ -60,7 +63,7 @@ case class SparkModel(conf: DedupConfig) {
     values(identityFieldPosition) = MapDocumentUtil.getJPathString(conf.getWf.getIdPath, documentContext)
 
     schema.fieldNames.zipWithIndex.foldLeft(values) {
-      case ((res, (fname, index))) => {
+      case ((res, (fname, index))) =>
         val fdef = conf.getPace.getModelMap.get(fname)
 
         if (fdef != null) {
@@ -96,13 +99,52 @@ case class SparkModel(conf: DedupConfig) {
             case Type.DoubleArray =>
               MapDocumentUtil.getJPathArray(fdef.getPath, json)
           }
+
+          val filter = fdef.getFilter
+
+          if (StringUtils.isNotBlank(fdef.getClean)) {
+            res(index) = res(index) match {
+              case x: Seq[String] => x.map(clean(_, fdef.getClean)).toSeq
+              case _ => clean(res(index).toString, fdef.getClean)
+            }
+          }
+
+          if (filter != null && !filter.isEmpty) {
+            res(index) = res(index) match {
+              case x: String if filter.contains(x.toLowerCase(Locale.ROOT)) => null
+              case x: Seq[String] => x.filter(s => !filter.contains(s.toLowerCase(Locale.ROOT))).toSeq
+              case _ => res(index)
+            }
+          }
+
+          if (fdef.getSorted) {
+            res(index) = res(index) match {
+              case x: Seq[String] => x.sorted.toSeq
+              case _ => res(index)
+            }
+          }
         }
 
         res
-      }
     }
 
     new GenericRowWithSchema(values, schema)
   }
+
+  def clean(value: String, cleantype: String) : String = {
+    val res = cleantype match {
+      case "title" => AbstractPaceFunctions.cleanup(value)
+      case _ => value
+    }
+
+//    if (!res.equals(AbstractPaceFunctions.normalize(value))) {
+//      println(res)
+//      println(AbstractPaceFunctions.normalize(value))
+//      println()
+//    }
+
+    res
+  }
+
 }
 

dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/AuthorsMatch.java

@@ -23,7 +23,6 @@ public class AuthorsMatch extends AbstractListComparator {
 	private String MODE; // full or surname
 	private int SIZE_THRESHOLD;
 	private String TYPE; // count or percentage
-	private int common;
 
 	public AuthorsMatch(Map<String, String> params) {
 		super(params, new com.wcohen.ss.JaroWinkler());
@@ -35,7 +34,6 @@ public class AuthorsMatch extends AbstractListComparator {
 		FULLNAME_THRESHOLD = Double.parseDouble(params.getOrDefault("fullname_th", "0.9"));
 		SIZE_THRESHOLD = Integer.parseInt(params.getOrDefault("size_th", "20"));
 		TYPE = params.getOrDefault("type", "percentage");
-		common = 0;
 	}
 
 	protected AuthorsMatch(double w, AbstractStringDistance ssalgo) {
@@ -44,22 +42,27 @@ public class AuthorsMatch extends AbstractListComparator {
 
 	@Override
 	public double compare(final List<String> a, final List<String> b, final Config conf) {
-
 		if (a.isEmpty() || b.isEmpty())
 			return -1;
 
 		if (a.size() > SIZE_THRESHOLD || b.size() > SIZE_THRESHOLD)
 			return 1.0;
 
-		List<Person> aList = a.stream().map(author -> new Person(author, false)).collect(Collectors.toList());
+		int maxMiss = Integer.MAX_VALUE;
 		List<Person> bList = b.stream().map(author -> new Person(author, false)).collect(Collectors.toList());
 
-		common = 0;
+		Double threshold = getDoubleParam("threshold");
+
+		if (threshold != null && threshold >= 0.0 && threshold <= 1.0 && a.size() == b.size()) {
+			maxMiss = (int) Math.floor((1 - threshold) * Math.max(a.size(), b.size()));
+		}
+
+		int common = 0;
 		// compare each element of List1 with each element of List2
-		for (Person p1 : aList)
+		for (int i = 0; i < a.size(); i++) {
+			Person p1 = new Person(a.get(i), false);
 
 			for (Person p2 : bList) {
-
 				// both persons are inaccurate
 				if (!p1.isAccurate() && !p2.isAccurate()) {
 					// compare just normalized fullnames
@@ -118,11 +121,15 @@ public class AuthorsMatch extends AbstractListComparator {
 					}
 
 				}
-
 			}
 
+			if (i - common > maxMiss) {
+				return 0.0;
+			}
+		}
+
 		// normalization factor to compute the score
-		int normFactor = aList.size() == bList.size() ? aList.size() : (aList.size() + bList.size() - common);
+		int normFactor = a.size() == b.size() ? a.size() : (a.size() + b.size() - common);
 
 		if (TYPE.equals("percentage")) {
 			return (double) common / normFactor;

dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/InstanceTypeMatch.java

@@ -25,6 +25,7 @@ public class InstanceTypeMatch extends AbstractListComparator {
 		translationMap.put("Conference object", "*");
 		translationMap.put("Other literature type", "*");
 		translationMap.put("Unknown", "*");
+		translationMap.put("UNKNOWN", "*");
 
 		// article types
 		translationMap.put("Article", "Article");
@@ -76,5 +77,4 @@ public class InstanceTypeMatch extends AbstractListComparator {
 	protected double normalize(final double d) {
 		return d;
 	}
-
 }

dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/LevensteinTitle.java

@@ -3,6 +3,7 @@ package eu.dnetlib.pace.tree;
 
 import java.util.Map;
 
+import org.apache.commons.lang3.StringUtils;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 
@@ -30,16 +31,25 @@ public class LevensteinTitle extends AbstractStringComparator {
 	}
 
 	@Override
-	public double distance(final String a, final String b, final Config conf) {
+	public double distance(final String ca, final String cb, final Config conf) {
-		final String ca = cleanup(a);
-		final String cb = cleanup(b);
-
 		final boolean check = checkNumbers(ca, cb);
 
 		if (check)
 			return 0.5;
 
-		return normalize(ssalgo.score(ca, cb), ca.length(), cb.length());
+		Double threshold = getDoubleParam("threshold");
+
+		// reduce Levenshtein algo complexity when target threshold is known
+		if (threshold != null && threshold >= 0.0 && threshold <= 1.0) {
+			int maxdistance = (int) Math.floor((1 - threshold) * Math.max(ca.length(), cb.length()));
+			int score = StringUtils.getLevenshteinDistance(ca, cb, maxdistance);
+			if (score == -1) {
+				return 0;
+			}
+			return normalize(score, ca.length(), cb.length());
+		} else {
+			return normalize(StringUtils.getLevenshteinDistance(ca, cb), ca.length(), cb.length());
+		}
 	}
 
 	private double normalize(final double score, final int la, final int lb) {

dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/MaxLengthMatch.java

@@ -0,0 +1,29 @@
+
+package eu.dnetlib.pace.tree;
+
+import java.util.Map;
+
+import eu.dnetlib.pace.config.Config;
+import eu.dnetlib.pace.tree.support.AbstractStringComparator;
+import eu.dnetlib.pace.tree.support.ComparatorClass;
+
+@ComparatorClass("maxLengthMatch")
+public class MaxLengthMatch extends AbstractStringComparator {
+
+	private final int limit;
+
+	public MaxLengthMatch(Map<String, String> params) {
+		super(params);
+
+		limit = Integer.parseInt(params.getOrDefault("limit", "200"));
+	}
+
+	@Override
+	public double compare(String a, String b, final Config conf) {
+		return a.length() < limit && b.length() < limit ? 1.0 : -1.0;
+	}
+
+	protected String toString(final Object object) {
+		return toFirstString(object);
+	}
+}

dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/support/AbstractComparator.java

@@ -127,4 +127,14 @@ public abstract class AbstractComparator<T> extends AbstractPaceFunctions implem
 		return this.weight;
 	}
 
+	public Double getDoubleParam(String name) {
+		String svalue = params.get(name);
+
+		try {
+			return Double.parseDouble(svalue);
+		} catch (Throwable t) {
+		}
+
+		return null;
+	}
 }

dhp-pace-core/src/main/java/eu/dnetlib/pace/util/BlockProcessor.java

@@ -67,8 +67,10 @@ public class BlockProcessor {
 
 	private void processRows(final List<Row> queue, final Reporter context) {
 
-		for (int pivotPos = 0; pivotPos < queue.size(); pivotPos++) {
+		IncrementalConnectedComponents icc = new IncrementalConnectedComponents(queue.size());
-			final Row pivot = queue.get(pivotPos);
+
+		for (int i = 0; i < queue.size(); i++) {
+			final Row pivot = queue.get(i);
 
 			final String idPivot = pivot.getString(identifierFieldPos); // identifier
 			final Object fieldsPivot = getJavaValue(pivot, orderFieldPos);
@@ -76,9 +78,9 @@ public class BlockProcessor {
 			final WfConfig wf = dedupConf.getWf();
 
 			if (fieldPivot != null) {
-				int i = 0;
+				for (int j = icc.nextUnconnected(i, i + 1); j >= 0
-				for (int windowPos = pivotPos + 1; windowPos < queue.size(); windowPos++) {
+					&& j < queue.size(); j = icc.nextUnconnected(i, j + 1)) {
-					final Row curr = queue.get(windowPos);
+					final Row curr = queue.get(j);
 					final String idCurr = curr.getString(identifierFieldPos); // identifier
 
 					if (mustSkip(idCurr)) {
@@ -86,7 +88,7 @@ public class BlockProcessor {
 						break;
 					}
 
-					if (++i > wf.getSlidingWindowSize()) {
+					if (wf.getSlidingWindowSize() > 0 && (j - i) > wf.getSlidingWindowSize()) {
 						break;
 					}
 
@@ -97,7 +99,9 @@ public class BlockProcessor {
 
 						final TreeProcessor treeProcessor = new TreeProcessor(dedupConf);
 
-						emitOutput(treeProcessor.compare(pivot, curr), idPivot, idCurr, context);
+						if (emitOutput(treeProcessor.compare(pivot, curr), idPivot, idCurr, context)) {
+							icc.connect(i, j);
+						}
 					}
 				}
 			}
@@ -115,7 +119,8 @@ public class BlockProcessor {
 		return null;
 	}
 
-	private void emitOutput(final boolean result, final String idPivot, final String idCurr, final Reporter context) {
+	private boolean emitOutput(final boolean result, final String idPivot, final String idCurr,
+		final Reporter context) {
 
 		if (result) {
 			if (idPivot.compareTo(idCurr) <= 0) {
@@ -127,6 +132,8 @@ public class BlockProcessor {
 		} else {
 			context.incrementCounter(dedupConf.getWf().getEntityType(), "d < " + dedupConf.getWf().getThreshold(), 1);
 		}
+
+		return result;
 	}
 
 	private boolean mustSkip(final String idPivot) {
@@ -142,5 +149,4 @@ public class BlockProcessor {
 
 		context.emit(type, from, to);
 	}
-
 }

dhp-pace-core/src/main/java/eu/dnetlib/pace/util/IncrementalConnectedComponents.java

@@ -0,0 +1,50 @@
+
+package eu.dnetlib.pace.util;
+
+import java.util.BitSet;
+
+public class IncrementalConnectedComponents {
+	final private int size;
+
+	final private BitSet[] indexes;
+
+	IncrementalConnectedComponents(int size) {
+		this.size = size;
+		this.indexes = new BitSet[size];
+	}
+
+	public void connect(int i, int j) {
+		if (indexes[i] == null) {
+			if (indexes[j] == null) {
+				indexes[i] = new BitSet(size);
+			} else {
+				indexes[i] = indexes[j];
+			}
+		} else {
+			if (indexes[j] != null && indexes[i] != indexes[j]) {
+				// merge adjacency lists for i and j
+				indexes[i].or(indexes[j]);
+			}
+		}
+
+		indexes[i].set(i);
+		indexes[i].set(j);
+		indexes[j] = indexes[i];
+	}
+
+	public int nextUnconnected(int i, int j) {
+		if (indexes[i] == null) {
+			return j;
+		}
+		int result = indexes[i].nextClearBit(j);
+
+		return (result >= size) ? -1 : result;
+	}
+
+	public BitSet getConnections(int i) {
+		if (indexes[i] == null) {
+			return null;
+		}
+		return indexes[i];
+	}
+}

dhp-pace-core/src/main/java/eu/dnetlib/pace/util/MapDocumentUtil.java

@@ -97,6 +97,8 @@ public class MapDocumentUtil {
 			Object o = json.read(jsonPath);
 			if (o instanceof String)
 				return (String) o;
+			if (o instanceof Number)
+				return (String) o.toString();
 			if (o instanceof JSONArray && ((JSONArray) o).size() > 0)
 				return (String) ((JSONArray) o).get(0);
 			return "";

dhp-pace-core/src/main/java/eu/dnetlib/pace/util/PaceResolver.java

@@ -40,7 +40,7 @@ public class PaceResolver implements Serializable {
 				Collectors.toMap(cl -> cl.getAnnotation(ComparatorClass.class).value(), cl -> (Class<Comparator>) cl));
 	}
 
-	public ClusteringFunction getClusteringFunction(String name, Map<String, Integer> params) throws PaceException {
+	public ClusteringFunction getClusteringFunction(String name, Map<String, Object> params) throws PaceException {
 		try {
 			return clusteringFunctions.get(name).getDeclaredConstructor(Map.class).newInstance(params);
 		} catch (InstantiationException | IllegalAccessException | InvocationTargetException

dhp-pace-core/src/test/java/eu/dnetlib/pace/clustering/ClusteringFunctionTest.java

@@ -15,7 +15,7 @@ import eu.dnetlib.pace.config.DedupConfig;
 
 public class ClusteringFunctionTest extends AbstractPaceTest {
 
-	private static Map<String, Integer> params;
+	private static Map<String, Object> params;
 	private static DedupConfig conf;
 
 	@BeforeAll
@@ -40,10 +40,10 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testNgram() {
-		params.put("ngramLen", 3);
+		params.put("ngramLen", "3");
-		params.put("max", 8);
+		params.put("max", "8");
-		params.put("maxPerToken", 2);
+		params.put("maxPerToken", "2");
-		params.put("minNgramLen", 1);
+		params.put("minNgramLen", "1");
 
 		final ClusteringFunction ngram = new Ngrams(params);
 
@@ -54,8 +54,8 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testNgramPairs() {
-		params.put("ngramLen", 3);
+		params.put("ngramLen", "3");
-		params.put("max", 2);
+		params.put("max", "2");
 
 		final ClusteringFunction np = new NgramPairs(params);
 
@@ -66,8 +66,8 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testSortedNgramPairs() {
-		params.put("ngramLen", 3);
+		params.put("ngramLen", "3");
-		params.put("max", 2);
+		params.put("max", "2");
 
 		final ClusteringFunction np = new SortedNgramPairs(params);
 
@@ -87,9 +87,9 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testAcronym() {
-		params.put("max", 4);
+		params.put("max", "4");
-		params.put("minLen", 1);
+		params.put("minLen", "1");
-		params.put("maxLen", 3);
+		params.put("maxLen", "3");
 
 		final ClusteringFunction acro = new Acronyms(params);
 
@@ -100,8 +100,8 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testSuffixPrefix() {
-		params.put("len", 3);
+		params.put("len", "3");
-		params.put("max", 4);
+		params.put("max", "4");
 
 		final ClusteringFunction sp = new SuffixPrefix(params);
 
@@ -109,8 +109,8 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 		System.out.println(s);
 		System.out.println(sp.apply(conf, Lists.newArrayList(s)));
 
-		params.put("len", 3);
+		params.put("len", "3");
-		params.put("max", 1);
+		params.put("max", "1");
 
 		System.out.println(sp.apply(conf, Lists.newArrayList("Framework for general-purpose deduplication")));
 	}
@@ -118,8 +118,8 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 	@Test
 	public void testWordsSuffixPrefix() {
 
-		params.put("len", 3);
+		params.put("len", "3");
-		params.put("max", 4);
+		params.put("max", "4");
 
 		final ClusteringFunction sp = new WordsSuffixPrefix(params);
 
@@ -130,7 +130,7 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 
 	@Test
 	public void testWordsStatsSuffixPrefix() {
-		params.put("mod", 10);
+		params.put("mod", "10");
 
 		final ClusteringFunction sp = new WordsStatsSuffixPrefixChain(params);
 
@@ -167,7 +167,7 @@ public class ClusteringFunctionTest extends AbstractPaceTest {
 	@Test
 	public void testFieldValue() {
 
-		params.put("randomLength", 5);
+		params.put("randomLength", "5");
 
 		final ClusteringFunction sp = new SpaceTrimmingFieldValue(params);
 

dhp-pace-core/src/test/java/eu/dnetlib/pace/util/IncrementalConnectedComponentsTest.java

@@ -0,0 +1,40 @@
+
+package eu.dnetlib.pace.util;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertNull;
+
+import org.junit.jupiter.api.Test;
+
+public class IncrementalConnectedComponentsTest {
+
+	@Test
+	public void transitiveClosureTest() {
+		IncrementalConnectedComponents icc = new IncrementalConnectedComponents(10);
+
+		icc.connect(0, 1);
+		icc.connect(0, 2);
+		icc.connect(0, 3);
+
+		icc.connect(1, 2);
+		icc.connect(1, 4);
+		icc.connect(1, 5);
+
+		icc.connect(6, 7);
+		icc.connect(6, 9);
+
+		assertEquals(icc.getConnections(0).toString(), "{0, 1, 2, 3, 4, 5}");
+		assertEquals(icc.getConnections(1).toString(), "{0, 1, 2, 3, 4, 5}");
+		assertEquals(icc.getConnections(2).toString(), "{0, 1, 2, 3, 4, 5}");
+		assertEquals(icc.getConnections(3).toString(), "{0, 1, 2, 3, 4, 5}");
+		assertEquals(icc.getConnections(4).toString(), "{0, 1, 2, 3, 4, 5}");
+		assertEquals(icc.getConnections(5).toString(), "{0, 1, 2, 3, 4, 5}");
+
+		assertEquals(icc.getConnections(6).toString(), "{6, 7, 9}");
+		assertEquals(icc.getConnections(7).toString(), "{6, 7, 9}");
+		assertEquals(icc.getConnections(9).toString(), "{6, 7, 9}");
+
+		assertNull(icc.getConnections(8));
+	}
+
+}

dhp-workflows/dhp-actionmanager/src/main/java/eu/dnetlib/dhp/actionmanager/promote/PromoteAction.java

@@ -0,0 +1,39 @@
+/*
+ * Copyright (c) 2024.
+ * SPDX-FileCopyrightText: © 2023 Consiglio Nazionale delle Ricerche
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+package eu.dnetlib.dhp.actionmanager.promote;
+
+/** Encodes the Actionset promotion strategies */
+public class PromoteAction {
+
+	/** The supported actionset promotion strategies
+	 *
+	 * ENRICH: promotes only records in the actionset matching another record in the
+	 *  graph and enriches them applying the given MergeAndGet strategy
+	 * UPSERT: promotes all the records in an actionset, matching records are updated
+	 *  using the given MergeAndGet strategy, the non-matching record as inserted as they are.
+	 */
+	public enum Strategy {
+		ENRICH, UPSERT
+	}
+
+	/**
+	 * Returns the string representation of the join type implementing the given PromoteAction.
+	 *
+	 * @param strategy the strategy to be used to promote the Actionset contents
+	 * @return the join type used to implement the promotion strategy
+	 */
+	public static String joinTypeForStrategy(PromoteAction.Strategy strategy) {
+		switch (strategy) {
+			case ENRICH:
+				return "left_outer";
+			case UPSERT:
+				return "full_outer";
+			default:
+				throw new IllegalStateException("unsupported PromoteAction: " + strategy.toString());
+		}
+	}
+}

dhp-workflows/dhp-actionmanager/src/main/java/eu/dnetlib/dhp/actionmanager/promote/PromoteActionPayloadForGraphTableJob.java

@@ -67,8 +67,9 @@ public class PromoteActionPayloadForGraphTableJob {
 		String outputGraphTablePath = parser.get("outputGraphTablePath");
 		logger.info("outputGraphTablePath: {}", outputGraphTablePath);
 
-		MergeAndGet.Strategy strategy = MergeAndGet.Strategy.valueOf(parser.get("mergeAndGetStrategy").toUpperCase());
+		MergeAndGet.Strategy mergeAndGetStrategy = MergeAndGet.Strategy
-		logger.info("strategy: {}", strategy);
+			.valueOf(parser.get("mergeAndGetStrategy").toUpperCase());
+		logger.info("mergeAndGetStrategy: {}", mergeAndGetStrategy);
 
 		Boolean shouldGroupById = Optional
 			.ofNullable(parser.get("shouldGroupById"))
@@ -76,6 +77,12 @@ public class PromoteActionPayloadForGraphTableJob {
 			.orElse(true);
 		logger.info("shouldGroupById: {}", shouldGroupById);
 
+		PromoteAction.Strategy promoteActionStrategy = Optional
+			.ofNullable(parser.get("promoteActionStrategy"))
+			.map(PromoteAction.Strategy::valueOf)
+			.orElse(PromoteAction.Strategy.UPSERT);
+		logger.info("promoteActionStrategy: {}", promoteActionStrategy);
+
 		@SuppressWarnings("unchecked")
 		Class<? extends Oaf> rowClazz = (Class<? extends Oaf>) Class.forName(graphTableClassName);
 		@SuppressWarnings("unchecked")
@@ -97,7 +104,8 @@ public class PromoteActionPayloadForGraphTableJob {
 					inputGraphTablePath,
 					inputActionPayloadPath,
 					outputGraphTablePath,
-					strategy,
+					mergeAndGetStrategy,
+					promoteActionStrategy,
 					rowClazz,
 					actionPayloadClazz,
 					shouldGroupById);
@@ -124,14 +132,16 @@ public class PromoteActionPayloadForGraphTableJob {
 		String inputGraphTablePath,
 		String inputActionPayloadPath,
 		String outputGraphTablePath,
-		MergeAndGet.Strategy strategy,
+		MergeAndGet.Strategy mergeAndGetStrategy,
+		PromoteAction.Strategy promoteActionStrategy,
 		Class<G> rowClazz,
 		Class<A> actionPayloadClazz, Boolean shouldGroupById) {
 		Dataset<G> rowDS = readGraphTable(spark, inputGraphTablePath, rowClazz);
 		Dataset<A> actionPayloadDS = readActionPayload(spark, inputActionPayloadPath, actionPayloadClazz);
 
 		Dataset<G> result = promoteActionPayloadForGraphTable(
-			rowDS, actionPayloadDS, strategy, rowClazz, actionPayloadClazz, shouldGroupById)
+			rowDS, actionPayloadDS, mergeAndGetStrategy, promoteActionStrategy, rowClazz, actionPayloadClazz,
+			shouldGroupById)
 				.map((MapFunction<G, G>) value -> value, Encoders.bean(rowClazz));
 
 		saveGraphTable(result, outputGraphTablePath);
@@ -183,7 +193,8 @@ public class PromoteActionPayloadForGraphTableJob {
 	private static <G extends Oaf, A extends Oaf> Dataset<G> promoteActionPayloadForGraphTable(
 		Dataset<G> rowDS,
 		Dataset<A> actionPayloadDS,
-		MergeAndGet.Strategy strategy,
+		MergeAndGet.Strategy mergeAndGetStrategy,
+		PromoteAction.Strategy promoteActionStrategy,
 		Class<G> rowClazz,
 		Class<A> actionPayloadClazz,
 		Boolean shouldGroupById) {
@@ -195,8 +206,9 @@ public class PromoteActionPayloadForGraphTableJob {
 
 		SerializableSupplier<Function<G, String>> rowIdFn = ModelSupport::idFn;
 		SerializableSupplier<Function<A, String>> actionPayloadIdFn = ModelSupport::idFn;
-		SerializableSupplier<BiFunction<G, A, G>> mergeRowWithActionPayloadAndGetFn = MergeAndGet.functionFor(strategy);
+		SerializableSupplier<BiFunction<G, A, G>> mergeRowWithActionPayloadAndGetFn = MergeAndGet
-		SerializableSupplier<BiFunction<G, G, G>> mergeRowsAndGetFn = MergeAndGet.functionFor(strategy);
+			.functionFor(mergeAndGetStrategy);
+		SerializableSupplier<BiFunction<G, G, G>> mergeRowsAndGetFn = MergeAndGet.functionFor(mergeAndGetStrategy);
 		SerializableSupplier<G> zeroFn = zeroFn(rowClazz);
 		SerializableSupplier<Function<G, Boolean>> isNotZeroFn = PromoteActionPayloadForGraphTableJob::isNotZeroFnUsingIdOrSourceAndTarget;
 
@@ -207,6 +219,7 @@ public class PromoteActionPayloadForGraphTableJob {
 				rowIdFn,
 				actionPayloadIdFn,
 				mergeRowWithActionPayloadAndGetFn,
+				promoteActionStrategy,
 				rowClazz,
 				actionPayloadClazz);
 

dhp-workflows/dhp-actionmanager/src/main/java/eu/dnetlib/dhp/actionmanager/promote/PromoteActionPayloadFunctions.java

@@ -34,6 +34,7 @@ public class PromoteActionPayloadFunctions {
 	 * @param rowIdFn Function used to get the id of graph table row
 	 * @param actionPayloadIdFn Function used to get id of action payload instance
 	 * @param mergeAndGetFn Function used to merge graph table row and action payload instance
+	 * @param promoteActionStrategy the Actionset promotion strategy
 	 * @param rowClazz Class of graph table
 	 * @param actionPayloadClazz Class of action payload
 	 * @param <G> Type of graph table row
@@ -46,6 +47,7 @@ public class PromoteActionPayloadFunctions {
 		SerializableSupplier<Function<G, String>> rowIdFn,
 		SerializableSupplier<Function<A, String>> actionPayloadIdFn,
 		SerializableSupplier<BiFunction<G, A, G>> mergeAndGetFn,
+		PromoteAction.Strategy promoteActionStrategy,
 		Class<G> rowClazz,
 		Class<A> actionPayloadClazz) {
 		if (!isSubClass(rowClazz, actionPayloadClazz)) {
@@ -61,7 +63,7 @@ public class PromoteActionPayloadFunctions {
 			.joinWith(
 				actionPayloadWithIdDS,
 				rowWithIdDS.col("_1").equalTo(actionPayloadWithIdDS.col("_1")),
-				"full_outer")
+				PromoteAction.joinTypeForStrategy(promoteActionStrategy))
 			.map(
 				(MapFunction<Tuple2<Tuple2<String, G>, Tuple2<String, A>>, G>) value -> {
 					Optional<G> rowOpt = Optional.ofNullable(value._1()).map(Tuple2::_2);

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/promote/promote_action_payload_for_graph_table_input_parameters.json

@@ -41,6 +41,12 @@
     "paramDescription": "strategy for merging graph table objects with action payload instances, MERGE_FROM_AND_GET or SELECT_NEWER_AND_GET",
     "paramRequired": true
   },
+  {
+    "paramName": "pas",
+    "paramLongName": "promoteActionStrategy",
+    "paramDescription": "strategy for promoting the actionset contents into the graph tables, ENRICH or UPSERT (default)",
+    "paramRequired": false
+  },
   {
     "paramName": "sgid",
     "paramLongName": "shouldGroupById",

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/dataset/oozie_app/workflow.xml

@@ -115,6 +115,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
             <arg>--outputGraphTablePath</arg><arg>${workingDir}/dataset</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="DecisionPromoteResultActionPayloadForDatasetTable"/>
@@ -167,6 +168,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Result</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/dataset</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="End"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/datasource/oozie_app/workflow.xml

@@ -106,6 +106,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Datasource</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/datasource</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
         </spark>
         <ok to="End"/>
         <error to="Kill"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/organization/oozie_app/workflow.xml

@@ -106,6 +106,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Organization</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/organization</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
         </spark>
         <ok to="End"/>
         <error to="Kill"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/otherresearchproduct/oozie_app/workflow.xml

@@ -114,6 +114,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
             <arg>--outputGraphTablePath</arg><arg>${workingDir}/otherresearchproduct</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="DecisionPromoteResultActionPayloadForOtherResearchProductTable"/>
@@ -166,6 +167,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Result</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/otherresearchproduct</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="End"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/project/oozie_app/workflow.xml

@@ -106,6 +106,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Project</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/project</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
         </spark>
         <ok to="End"/>
         <error to="Kill"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/publication/oozie_app/workflow.xml

@@ -115,6 +115,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
             <arg>--outputGraphTablePath</arg><arg>${workingDir}/publication</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="DecisionPromoteResultActionPayloadForPublicationTable"/>
@@ -167,6 +168,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Result</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/publication</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="End"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/relation/oozie_app/workflow.xml

@@ -107,6 +107,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Relation</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/relation</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
         </spark>
         <ok to="End"/>
         <error to="Kill"/>

dhp-workflows/dhp-actionmanager/src/main/resources/eu/dnetlib/dhp/actionmanager/wf/software/oozie_app/workflow.xml

@@ -114,6 +114,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
             <arg>--outputGraphTablePath</arg><arg>${workingDir}/software</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="DecisionPromoteResultActionPayloadForSoftwareTable"/>
@@ -166,6 +167,7 @@
             <arg>--actionPayloadClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Result</arg>
             <arg>--outputGraphTablePath</arg><arg>${outputGraphRootPath}/software</arg>
             <arg>--mergeAndGetStrategy</arg><arg>${mergeAndGetStrategy}</arg>
+            <arg>--promoteActionStrategy</arg><arg>${promoteActionStrategy}</arg>
             <arg>--shouldGroupById</arg><arg>${shouldGroupById}</arg>
         </spark>
         <ok to="End"/>

dhp-workflows/dhp-actionmanager/src/test/java/eu/dnetlib/dhp/actionmanager/promote/PromoteActionPayloadFunctionsTest.java

@@ -54,7 +54,7 @@ public class PromoteActionPayloadFunctionsTest {
 				RuntimeException.class,
 				() -> PromoteActionPayloadFunctions
 					.joinGraphTableWithActionPayloadAndMerge(
-						null, null, null, null, null, OafImplSubSub.class, OafImpl.class));
+						null, null, null, null, null, null, OafImplSubSub.class, OafImpl.class));
 		}
 
 		@Test
@@ -104,6 +104,7 @@ public class PromoteActionPayloadFunctionsTest {
 					rowIdFn,
 					actionPayloadIdFn,
 					mergeAndGetFn,
+					PromoteAction.Strategy.UPSERT,
 					OafImplSubSub.class,
 					OafImplSubSub.class)
 				.collectAsList();
@@ -183,6 +184,7 @@ public class PromoteActionPayloadFunctionsTest {
 					rowIdFn,
 					actionPayloadIdFn,
 					mergeAndGetFn,
+					PromoteAction.Strategy.UPSERT,
 					OafImplSubSub.class,
 					OafImplSub.class)
 				.collectAsList();

dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/createunresolvedentities/PrepareFOSSparkJob.java

@@ -124,8 +124,19 @@ public class PrepareFOSSparkJob implements Serializable {
 		FOSDataModel first) {
 		level1.add(first.getLevel1());
 		level2.add(first.getLevel2());
-		level3.add(first.getLevel3() + "@@" + first.getScoreL3());
+		if (Optional.ofNullable(first.getLevel3()).isPresent() &&
-		level4.add(first.getLevel4() + "@@" + first.getScoreL4());
+			!first.getLevel3().equalsIgnoreCase(NA) && !first.getLevel3().equalsIgnoreCase(NULL)
+			&& first.getLevel3() != null)
+			level3.add(first.getLevel3() + "@@" + first.getScoreL3());
+		else
+			level3.add(NULL);
+		if (Optional.ofNullable(first.getLevel4()).isPresent() &&
+			!first.getLevel4().equalsIgnoreCase(NA) &&
+			!first.getLevel4().equalsIgnoreCase(NULL) &&
+			first.getLevel4() != null)
+			level4.add(first.getLevel4() + "@@" + first.getScoreL4());
+		else
+			level4.add(NULL);
 	}
 
 }

dhp-workflows/dhp-aggregation/src/main/resources/eu/dnetlib/dhp/sx/bio/ebi/baseline_to_oaf_params.json

@@ -2,7 +2,7 @@
   {"paramName":"mt",  "paramLongName":"master",         "paramDescription": "should be local or yarn",                  "paramRequired": true},
   {"paramName":"i",   "paramLongName":"isLookupUrl",    "paramDescription": "isLookupUrl",                              "paramRequired": true},
   {"paramName":"w",   "paramLongName":"workingPath",    "paramDescription": "the path of the sequencial file to read",  "paramRequired": true},
-  {"paramName":"t",   "paramLongName":"targetPath",     "paramDescription": "the oaf path ",                            "paramRequired": true},
+  {"paramName":"mo",   "paramLongName":"mdstoreOutputVersion",     "paramDescription": "the oaf path ",                            "paramRequired": true},
   {"paramName":"s",   "paramLongName":"skipUpdate",     "paramDescription": "skip update ",                             "paramRequired": false},
   {"paramName":"h",   "paramLongName":"hdfsServerUri",  "paramDescription": "the  working path ",                       "paramRequired": true}
 ]

dhp-workflows/dhp-aggregation/src/main/resources/eu/dnetlib/dhp/sx/bio/pubmed/oozie_app/workflow.xml

@@ -1,4 +1,4 @@
-    <workflow-app name="Download_Transform_Pubmed_Workflow" xmlns="uri:oozie:workflow:0.5">
+<workflow-app name="Download_Transform_Pubmed_Workflow" xmlns="uri:oozie:workflow:0.5">
     <parameters>
         <property>
             <name>baselineWorkingPath</name>
@@ -9,8 +9,12 @@
             <description>The IS lookUp service endopoint</description>
         </property>
         <property>
-            <name>targetPath</name>
+            <name>mdStoreOutputId</name>
-            <description>The target path</description>
+            <description>the identifier of the cleaned MDStore</description>
+        </property>
+        <property>
+            <name>mdStoreManagerURI</name>
+            <description>the path of the cleaned mdstore</description>
         </property>
         <property>
             <name>skipUpdate</name>
@@ -19,12 +23,31 @@
         </property>
     </parameters>
 
-    <start to="ConvertDataset"/>
+    <start to="StartTransaction"/>
+
 
     <kill name="Kill">
         <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
     </kill>
 
+    <action name="StartTransaction">
+        <java>
+            <configuration>
+                <property>
+                    <name>oozie.launcher.mapreduce.user.classpath.first</name>
+                    <value>true</value>
+                </property>
+            </configuration>
+            <main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
+            <arg>--action</arg><arg>NEW_VERSION</arg>
+            <arg>--mdStoreID</arg><arg>${mdStoreOutputId}</arg>
+            <arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
+            <capture-output/>
+        </java>
+        <ok to="ConvertDataset"/>
+        <error to="RollBack"/>
+    </action>
+
     <action name="ConvertDataset">
         <spark xmlns="uri:oozie:spark-action:0.2">
             <master>yarn</master>
@@ -43,16 +66,52 @@
                 --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
             </spark-opts>
             <arg>--workingPath</arg><arg>${baselineWorkingPath}</arg>
-            <arg>--targetPath</arg><arg>${targetPath}</arg>
+            <arg>--mdstoreOutputVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
             <arg>--master</arg><arg>yarn</arg>
             <arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
             <arg>--hdfsServerUri</arg><arg>${nameNode}</arg>
             <arg>--skipUpdate</arg><arg>${skipUpdate}</arg>
         </spark>
+        <ok to="CommitVersion"/>
+        <error to="RollBack"/>
+    </action>
+
+    <action name="CommitVersion">
+        <java>
+            <configuration>
+                <property>
+                    <name>oozie.launcher.mapreduce.user.classpath.first</name>
+                    <value>true</value>
+                </property>
+            </configuration>
+            <main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
+            <arg>--action</arg><arg>COMMIT</arg>
+            <arg>--namenode</arg><arg>${nameNode}</arg>
+            <arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
+            <arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
+        </java>
         <ok to="End"/>
         <error to="Kill"/>
     </action>
 
+    <action name="RollBack">
+        <java>
+            <configuration>
+                <property>
+                    <name>oozie.launcher.mapreduce.user.classpath.first</name>
+                    <value>true</value>
+                </property>
+            </configuration>
+            <main-class>eu.dnetlib.dhp.aggregation.mdstore.MDStoreActionNode</main-class>
+            <arg>--action</arg><arg>ROLLBACK</arg>
+            <arg>--mdStoreVersion</arg><arg>${wf:actionData('StartTransaction')['mdStoreVersion']}</arg>
+            <arg>--mdStoreManagerURI</arg><arg>${mdStoreManagerURI}</arg>
+        </java>
+        <ok to="Kill"/>
+        <error to="Kill"/>
+    </action>
+
+
     <end name="End"/>
 
 </workflow-app>

dhp-workflows/dhp-aggregation/src/main/scala/eu/dnetlib/dhp/sx/bio/ebi/SparkCreateBaselineDataFrame.scala

@@ -2,9 +2,12 @@ package eu.dnetlib.dhp.sx.bio.ebi
 
 import eu.dnetlib.dhp.application.ArgumentApplicationParser
 import eu.dnetlib.dhp.collection.CollectionUtils
+import eu.dnetlib.dhp.common.Constants.{MDSTORE_DATA_PATH, MDSTORE_SIZE_PATH}
 import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup
+import eu.dnetlib.dhp.schema.mdstore.MDStoreVersion
 import eu.dnetlib.dhp.schema.oaf.{Oaf, Result}
 import eu.dnetlib.dhp.sx.bio.pubmed._
+import eu.dnetlib.dhp.utils.DHPUtils.{MAPPER, writeHdfsFile}
 import eu.dnetlib.dhp.utils.ISLookupClientFactory
 import org.apache.commons.io.IOUtils
 import org.apache.hadoop.conf.Configuration
@@ -164,11 +167,15 @@ object SparkCreateBaselineDataFrame {
     val workingPath = parser.get("workingPath")
     log.info("workingPath: {}", workingPath)
 
-    val targetPath = parser.get("targetPath")
+    val mdstoreOutputVersion = parser.get("mdstoreOutputVersion")
-    log.info("targetPath: {}", targetPath)
+    log.info("mdstoreOutputVersion: {}", mdstoreOutputVersion)
+
+    val cleanedMdStoreVersion = MAPPER.readValue(mdstoreOutputVersion, classOf[MDStoreVersion])
+    val outputBasePath = cleanedMdStoreVersion.getHdfsPath
+    log.info("outputBasePath: {}", outputBasePath)
 
     val hdfsServerUri = parser.get("hdfsServerUri")
-    log.info("hdfsServerUri: {}", targetPath)
+    log.info("hdfsServerUri: {}", hdfsServerUri)
 
     val skipUpdate = parser.get("skipUpdate")
     log.info("skipUpdate: {}", skipUpdate)
@@ -216,8 +223,11 @@ object SparkCreateBaselineDataFrame {
         .map(a => PubMedToOaf.convert(a, vocabularies))
         .as[Oaf]
         .filter(p => p != null),
-      targetPath
+      s"$outputBasePath/$MDSTORE_DATA_PATH"
     )
 
+    val df = spark.read.text(s"$outputBasePath/$MDSTORE_DATA_PATH")
+    val mdStoreSize = df.count
+    writeHdfsFile(spark.sparkContext.hadoopConfiguration, s"$mdStoreSize", s"$outputBasePath/$MDSTORE_SIZE_PATH")
   }
 }

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/AbstractSparkAction.java

@@ -101,6 +101,10 @@ abstract class AbstractSparkAction implements Serializable {
 		return SparkSession.builder().config(conf).getOrCreate();
 	}
 
+	protected static SparkSession getSparkWithHiveSession(SparkConf conf) {
+		return SparkSession.builder().enableHiveSupport().config(conf).getOrCreate();
+	}
+
 	protected static <T> void save(Dataset<T> dataset, String outPath, SaveMode mode) {
 		dataset.write().option("compression", "gzip").mode(mode).json(outPath);
 	}

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/DedupRecordFactory.java

@@ -1,128 +1,212 @@
 
 package eu.dnetlib.dhp.oa.dedup;
 
-import java.lang.reflect.InvocationTargetException;
 import java.util.*;
-import java.util.stream.Collectors;
+import java.util.stream.Stream;
 
 import org.apache.commons.beanutils.BeanUtils;
 import org.apache.commons.lang3.StringUtils;
+import org.apache.spark.api.java.function.FlatMapFunction;
 import org.apache.spark.api.java.function.MapFunction;
-import org.apache.spark.api.java.function.MapGroupsFunction;
+import org.apache.spark.api.java.function.ReduceFunction;
-import org.apache.spark.sql.Dataset;
+import org.apache.spark.sql.*;
-import org.apache.spark.sql.Encoders;
-import org.apache.spark.sql.SparkSession;
-
-import com.fasterxml.jackson.databind.DeserializationFeature;
-import com.fasterxml.jackson.databind.ObjectMapper;
-import com.google.common.collect.Lists;
 
 import eu.dnetlib.dhp.oa.dedup.model.Identifier;
 import eu.dnetlib.dhp.oa.merge.AuthorMerger;
 import eu.dnetlib.dhp.schema.common.ModelSupport;
-import eu.dnetlib.dhp.schema.oaf.*;
+import eu.dnetlib.dhp.schema.oaf.Author;
+import eu.dnetlib.dhp.schema.oaf.DataInfo;
+import eu.dnetlib.dhp.schema.oaf.OafEntity;
+import eu.dnetlib.dhp.schema.oaf.Result;
 import scala.Tuple2;
+import scala.Tuple3;
+import scala.collection.JavaConversions;
 
 public class DedupRecordFactory {
+	public static final class DedupRecordReduceState {
+		public final String dedupId;
 
-	protected static final ObjectMapper OBJECT_MAPPER = new ObjectMapper()
+		public final ArrayList<String> aliases = new ArrayList<>();
-		.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
+
+		public final HashSet<String> acceptanceDate = new HashSet<>();
+
+		public OafEntity entity;
+
+		public DedupRecordReduceState(String dedupId, String id, OafEntity entity) {
+			this.dedupId = dedupId;
+			this.entity = entity;
+			if (entity == null) {
+				aliases.add(id);
+			} else {
+				if (Result.class.isAssignableFrom(entity.getClass())) {
+					Result result = (Result) entity;
+					if (result.getDateofacceptance() != null
+						&& StringUtils.isNotBlank(result.getDateofacceptance().getValue())) {
+						acceptanceDate.add(result.getDateofacceptance().getValue());
+					}
+				}
+			}
+		}
+
+		public String getDedupId() {
+			return dedupId;
+		}
+	}
+
+	private static final int MAX_ACCEPTANCE_DATE = 20;
 
 	private DedupRecordFactory() {
 	}
 
-	public static <T extends OafEntity> Dataset<T> createDedupRecord(
+	public static Dataset<OafEntity> createDedupRecord(
 		final SparkSession spark,
 		final DataInfo dataInfo,
 		final String mergeRelsInputPath,
 		final String entitiesInputPath,
-		final Class<T> clazz) {
+		final Class<OafEntity> clazz) {
 
-		long ts = System.currentTimeMillis();
+		final long ts = System.currentTimeMillis();
+		final Encoder<OafEntity> beanEncoder = Encoders.bean(clazz);
+		final Encoder<OafEntity> kryoEncoder = Encoders.kryo(clazz);
 
 		// <id, json_entity>
-		Dataset<Tuple2<String, T>> entities = spark
+		Dataset<Row> entities = spark
 			.read()
-			.textFile(entitiesInputPath)
+			.schema(Encoders.bean(clazz).schema())
+			.json(entitiesInputPath)
+			.as(beanEncoder)
 			.map(
-				(MapFunction<String, Tuple2<String, T>>) it -> {
+				(MapFunction<OafEntity, Tuple2<String, OafEntity>>) entity -> {
-					T entity = OBJECT_MAPPER.readValue(it, clazz);
 					return new Tuple2<>(entity.getId(), entity);
 				},
-				Encoders.tuple(Encoders.STRING(), Encoders.kryo(clazz)));
+				Encoders.tuple(Encoders.STRING(), kryoEncoder))
+			.selectExpr("_1 AS id", "_2 AS kryoObject");
 
 		// <source, target>: source is the dedup_id, target is the id of the mergedIn
-		Dataset<Tuple2<String, String>> mergeRels = spark
+		Dataset<Row> mergeRels = spark
 			.read()
 			.load(mergeRelsInputPath)
-			.as(Encoders.bean(Relation.class))
 			.where("relClass == 'merges'")
-			.map(
+			.selectExpr("source as dedupId", "target as id");
-				(MapFunction<Relation, Tuple2<String, String>>) r -> new Tuple2<>(r.getSource(), r.getTarget()),
-				Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
 
 		return mergeRels
-			.joinWith(entities, mergeRels.col("_2").equalTo(entities.col("_1")), "inner")
+			.join(entities, JavaConversions.asScalaBuffer(Collections.singletonList("id")), "left")
+			.select("dedupId", "id", "kryoObject")
+			.as(Encoders.tuple(Encoders.STRING(), Encoders.STRING(), kryoEncoder))
 			.map(
-				(MapFunction<Tuple2<Tuple2<String, String>, Tuple2<String, T>>, Tuple2<String, T>>) value -> new Tuple2<>(
+				(MapFunction<Tuple3<String, String, OafEntity>, DedupRecordReduceState>) t -> new DedupRecordReduceState(
-					value._1()._1(), value._2()._2()),
+					t._1(), t._2(), t._3()),
-				Encoders.tuple(Encoders.STRING(), Encoders.kryo(clazz)))
+				Encoders.kryo(DedupRecordReduceState.class))
 			.groupByKey(
-				(MapFunction<Tuple2<String, T>, String>) Tuple2::_1, Encoders.STRING())
+				(MapFunction<DedupRecordReduceState, String>) DedupRecordReduceState::getDedupId, Encoders.STRING())
-			.mapGroups(
+			.reduceGroups(
-				(MapGroupsFunction<String, Tuple2<String, T>, T>) (key,
+				(ReduceFunction<DedupRecordReduceState>) (t1, t2) -> {
-					values) -> entityMerger(key, values, ts, dataInfo, clazz),
+					if (t1.entity == null) {
-				Encoders.bean(clazz));
+						t2.aliases.addAll(t1.aliases);
+						return t2;
+					}
+					if (t1.acceptanceDate.size() < MAX_ACCEPTANCE_DATE) {
+						t1.acceptanceDate.addAll(t2.acceptanceDate);
+					}
+					t1.aliases.addAll(t2.aliases);
+					t1.entity = reduceEntity(t1.entity, t2.entity);
+
+					return t1;
+				})
+			.flatMap((FlatMapFunction<Tuple2<String, DedupRecordReduceState>, OafEntity>) t -> {
+				String dedupId = t._1();
+				DedupRecordReduceState agg = t._2();
+
+				if (agg.acceptanceDate.size() >= MAX_ACCEPTANCE_DATE) {
+					return Collections.emptyIterator();
+				}
+
+				return Stream
+					.concat(
+						Stream
+							.of(agg.getDedupId())
+							.map(id -> createDedupOafEntity(id, agg.entity, dataInfo, ts)),
+						agg.aliases
+							.stream()
+							.map(id -> createMergedDedupAliasOafEntity(id, agg.entity, dataInfo, ts)))
+					.iterator();
+			}, beanEncoder);
 	}
 
-	public static <T extends OafEntity> T entityMerger(
+	private static OafEntity createDedupOafEntity(String id, OafEntity base, DataInfo dataInfo, long ts) {
-		String id, Iterator<Tuple2<String, T>> entities, long ts, DataInfo dataInfo, Class<T> clazz)
+		try {
-		throws IllegalAccessException, InstantiationException, InvocationTargetException {
+			OafEntity res = (OafEntity) BeanUtils.cloneBean(base);
+			res.setId(id);
+			res.setDataInfo(dataInfo);
+			res.setLastupdatetimestamp(ts);
+			return res;
+		} catch (Exception e) {
+			throw new RuntimeException(e);
+		}
+	}
 
-		final Comparator<Identifier<T>> idComparator = new IdentifierComparator<>();
+	private static OafEntity createMergedDedupAliasOafEntity(String id, OafEntity base, DataInfo dataInfo, long ts) {
+		try {
+			OafEntity res = createDedupOafEntity(id, base, dataInfo, ts);
+			DataInfo ds = (DataInfo) BeanUtils.cloneBean(dataInfo);
+			ds.setDeletedbyinference(true);
+			res.setDataInfo(ds);
+			return res;
+		} catch (Exception e) {
+			throw new RuntimeException(e);
+		}
+	}
 
-		final LinkedList<T> entityList = Lists
+	private static OafEntity reduceEntity(OafEntity entity, OafEntity duplicate) {
-			.newArrayList(entities)
-			.stream()
-			.map(t -> Identifier.newInstance(t._2()))
-			.sorted(idComparator)
-			.map(Identifier::getEntity)
-			.collect(Collectors.toCollection(LinkedList::new));
 
-		final T entity = clazz.newInstance();
+		if (duplicate == null) {
-		final T first = entityList.removeFirst();
+			return entity;
-
-		BeanUtils.copyProperties(entity, first);
-
-		final List<List<Author>> authors = Lists.newArrayList();
-
-		entityList
-			.forEach(
-				duplicate -> {
-					entity.mergeFrom(duplicate);
-					if (ModelSupport.isSubClass(duplicate, Result.class)) {
-						Result r1 = (Result) duplicate;
-						Optional
-							.ofNullable(r1.getAuthor())
-							.ifPresent(a -> authors.add(a));
-					}
-				});
-
-		// set authors and date
-		if (ModelSupport.isSubClass(entity, Result.class)) {
-			Optional
-				.ofNullable(((Result) entity).getAuthor())
-				.ifPresent(a -> authors.add(a));
-
-			((Result) entity).setAuthor(AuthorMerger.merge(authors));
 		}
 
-		entity.setId(id);
+		int compare = new IdentifierComparator<>()
+			.compare(Identifier.newInstance(entity), Identifier.newInstance(duplicate));
 
-		entity.setLastupdatetimestamp(ts);
+		if (compare > 0) {
-		entity.setDataInfo(dataInfo);
+			OafEntity swap = duplicate;
+			duplicate = entity;
+			entity = swap;
+		}
+
+		entity.mergeFrom(duplicate);
+
+		if (ModelSupport.isSubClass(duplicate, Result.class)) {
+			Result re = (Result) entity;
+			Result rd = (Result) duplicate;
+
+			List<List<Author>> authors = new ArrayList<>();
+			if (re.getAuthor() != null) {
+				authors.add(re.getAuthor());
+			}
+			if (rd.getAuthor() != null) {
+				authors.add(rd.getAuthor());
+			}
+
+			re.setAuthor(AuthorMerger.merge(authors));
+		}
 
 		return entity;
 	}
 
+	public static <T extends OafEntity> T entityMerger(
+		String id, Iterator<Tuple2<String, T>> entities, long ts, DataInfo dataInfo, Class<T> clazz) {
+		T base = entities.next()._2();
+
+		while (entities.hasNext()) {
+			T duplicate = entities.next()._2();
+			if (duplicate != null)
+				base = (T) reduceEntity(base, duplicate);
+		}
+
+		base.setId(id);
+		base.setDataInfo(dataInfo);
+		base.setLastupdatetimestamp(ts);
+
+		return base;
+	}
+
 }

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/IdGenerator.java

@@ -1,6 +1,7 @@
 
 package eu.dnetlib.dhp.oa.dedup;
 
+import static eu.dnetlib.dhp.utils.DHPUtils.md5;
 import static org.apache.commons.lang3.StringUtils.substringAfter;
 import static org.apache.commons.lang3.StringUtils.substringBefore;
 
@@ -14,33 +15,36 @@ import eu.dnetlib.dhp.schema.oaf.utils.PidType;
 public class IdGenerator implements Serializable {
 
 	// pick the best pid from the list (consider date and pidtype)
-	public static <T extends OafEntity> String generate(List<Identifier<T>> pids, String defaultID) {
+	public static <T extends OafEntity> String generate(List<? extends Identifier> pids, String defaultID) {
 		if (pids == null || pids.isEmpty())
 			return defaultID;
 
 		return generateId(pids);
 	}
 
-	private static <T extends OafEntity> String generateId(List<Identifier<T>> pids) {
+	private static String generateId(List<? extends Identifier> pids) {
-		Identifier<T> bp = pids
+		Identifier bp = pids
 			.stream()
 			.min(Identifier::compareTo)
 			.orElseThrow(() -> new IllegalStateException("unable to generate id"));
 
-		String prefix = substringBefore(bp.getOriginalID(), "|");
+		return generate(bp.getOriginalID());
-		String ns = substringBefore(substringAfter(bp.getOriginalID(), "|"), "::");
+	}
-		String suffix = substringAfter(bp.getOriginalID(), "::");
+
+	public static String generate(String originalId) {
+		String prefix = substringBefore(originalId, "|");
+		String ns = substringBefore(substringAfter(originalId, "|"), "::");
+		String suffix = substringAfter(originalId, "::");
 
 		final String pidType = substringBefore(ns, "_");
 		if (PidType.isValid(pidType)) {
 			return prefix + "|" + dedupify(ns) + "::" + suffix;
 		} else {
-			return prefix + "|dedup_wf_001::" + suffix;
+			return prefix + "|dedup_wf_002::" + md5(originalId); // hash the whole originalId to avoid collisions
 		}
 	}
 
 	private static String dedupify(String ns) {
-
 		StringBuilder prefix;
 		if (PidType.valueOf(substringBefore(ns, "_")) == PidType.openorgs) {
 			prefix = new StringBuilder(substringBefore(ns, "_"));
@@ -53,5 +57,4 @@ public class IdGenerator implements Serializable {
 		}
 		return prefix.substring(0, 12);
 	}
-
 }

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkCreateMergeRels.java

@@ -3,49 +3,47 @@ package eu.dnetlib.dhp.oa.dedup;
 
 import static eu.dnetlib.dhp.schema.common.ModelConstants.DNET_PROVENANCE_ACTIONS;
 import static eu.dnetlib.dhp.schema.common.ModelConstants.PROVENANCE_DEDUP;
+import static org.apache.spark.sql.functions.*;
 
 import java.io.IOException;
-import java.util.*;
+import java.time.LocalDate;
-import java.util.stream.Collectors;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.Optional;
 
 import org.apache.commons.io.IOUtils;
+import org.apache.commons.lang3.StringUtils;
 import org.apache.spark.SparkConf;
-import org.apache.spark.api.java.JavaPairRDD;
-import org.apache.spark.api.java.JavaSparkContext;
 import org.apache.spark.api.java.function.FlatMapFunction;
-import org.apache.spark.api.java.function.MapFunction;
+import org.apache.spark.sql.*;
-import org.apache.spark.api.java.function.MapGroupsFunction;
-import org.apache.spark.graphx.Edge;
-import org.apache.spark.rdd.RDD;
 import org.apache.spark.sql.Dataset;
-import org.apache.spark.sql.Encoders;
+import org.apache.spark.sql.catalyst.encoders.RowEncoder;
-import org.apache.spark.sql.SaveMode;
+import org.apache.spark.sql.expressions.UserDefinedFunction;
-import org.apache.spark.sql.SparkSession;
+import org.apache.spark.sql.expressions.Window;
+import org.apache.spark.sql.expressions.WindowSpec;
+import org.apache.spark.sql.types.DataTypes;
+import org.apache.spark.sql.types.StructType;
 import org.dom4j.DocumentException;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 import org.xml.sax.SAXException;
 
-import com.google.common.collect.Lists;
 import com.google.common.hash.Hashing;
+import com.kwartile.lib.cc.ConnectedComponent;
 
 import eu.dnetlib.dhp.application.ArgumentApplicationParser;
-import eu.dnetlib.dhp.oa.dedup.graph.ConnectedComponent;
-import eu.dnetlib.dhp.oa.dedup.graph.GraphProcessor;
-import eu.dnetlib.dhp.oa.dedup.model.Identifier;
 import eu.dnetlib.dhp.schema.common.EntityType;
 import eu.dnetlib.dhp.schema.common.ModelConstants;
 import eu.dnetlib.dhp.schema.common.ModelSupport;
-import eu.dnetlib.dhp.schema.oaf.DataInfo;
+import eu.dnetlib.dhp.schema.oaf.*;
-import eu.dnetlib.dhp.schema.oaf.OafEntity;
+import eu.dnetlib.dhp.schema.oaf.utils.PidType;
-import eu.dnetlib.dhp.schema.oaf.Qualifier;
-import eu.dnetlib.dhp.schema.oaf.Relation;
 import eu.dnetlib.dhp.utils.ISLookupClientFactory;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
 import eu.dnetlib.pace.config.DedupConfig;
-import eu.dnetlib.pace.util.MapDocumentUtil;
+import scala.Tuple3;
-import scala.Tuple2;
+import scala.collection.JavaConversions;
 
 public class SparkCreateMergeRels extends AbstractSparkAction {
 
@@ -68,10 +66,12 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
 		log.info("isLookupUrl {}", isLookUpUrl);
 
 		SparkConf conf = new SparkConf();
+		conf.set("hive.metastore.uris", parser.get("hiveMetastoreUris"));
 		conf.set("spark.serializer", "org.apache.spark.serializer.KryoSerializer");
+
 		conf.registerKryoClasses(ModelSupport.getOafModelClasses());
 
-		new SparkCreateMergeRels(parser, getSparkSession(conf))
+		new SparkCreateMergeRels(parser, getSparkWithHiveSession(conf))
 			.run(ISLookupClientFactory.getLookUpService(isLookUpUrl));
 	}
 
@@ -87,14 +87,15 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
 			.ofNullable(parser.get("cutConnectedComponent"))
 			.map(Integer::valueOf)
 			.orElse(0);
+
+		final String pivotHistoryDatabase = parser.get("pivotHistoryDatabase");
+
 		log.info("connected component cut: '{}'", cut);
 		log.info("graphBasePath: '{}'", graphBasePath);
 		log.info("isLookUpUrl:   '{}'", isLookUpUrl);
 		log.info("actionSetId:   '{}'", actionSetId);
 		log.info("workingPath:   '{}'", workingPath);
 
-		final JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
-
 		for (DedupConfig dedupConf : getConfigurations(isLookUpService, actionSetId)) {
 			final String subEntity = dedupConf.getWf().getSubEntityValue();
 			final Class<OafEntity> clazz = ModelSupport.entityTypes.get(EntityType.valueOf(subEntity));
@@ -106,113 +107,173 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
 
 			final String mergeRelPath = DedupUtility.createMergeRelPath(workingPath, actionSetId, subEntity);
 
-			// <hash(id), id>
+			final Dataset<Row> simRels = spark
-			JavaPairRDD<Object, String> vertexes = createVertexes(sc, graphBasePath, subEntity, dedupConf);
-
-			final RDD<Edge<String>> edgeRdd = spark
 				.read()
 				.load(DedupUtility.createSimRelPath(workingPath, actionSetId, subEntity))
-				.as(Encoders.bean(Relation.class))
+				.select("source", "target");
-				.javaRDD()
-				.map(it -> new Edge<>(hash(it.getSource()), hash(it.getTarget()), it.getRelClass()))
-				.rdd();
 
-			Dataset<Tuple2<String, String>> rawMergeRels = spark
+			UserDefinedFunction hashUDF = functions
-				.createDataset(
+				.udf(
-					GraphProcessor
+					(String s) -> hash(s), DataTypes.LongType);
-						.findCCs(vertexes.rdd(), edgeRdd, maxIterations, cut)
-						.toJavaRDD()
-						.filter(k -> k.getIds().size() > 1)
-						.flatMap(this::ccToRels)
-						.rdd(),
-					Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
 
-			Dataset<Tuple2<String, OafEntity>> entities = spark
+			// <hash(id), id>
+			Dataset<Row> vertexIdMap = simRels
+				.selectExpr("source as id")
+				.union(simRels.selectExpr("target as id"))
+				.distinct()
+				.withColumn("vertexId", hashUDF.apply(functions.col("id")));
+
+			// transform simrels into pairs of numeric ids
+			final Dataset<Row> edges = spark
 				.read()
-				.textFile(DedupUtility.createEntityPath(graphBasePath, subEntity))
+				.load(DedupUtility.createSimRelPath(workingPath, actionSetId, subEntity))
-				.map(
+				.select("source", "target")
-					(MapFunction<String, Tuple2<String, OafEntity>>) it -> {
+				.withColumn("source", hashUDF.apply(functions.col("source")))
-						OafEntity entity = OBJECT_MAPPER.readValue(it, clazz);
+				.withColumn("target", hashUDF.apply(functions.col("target")));
-						return new Tuple2<>(entity.getId(), entity);
-					},
-					Encoders.tuple(Encoders.STRING(), Encoders.kryo(clazz)));
 
-			Dataset<Relation> mergeRels = rawMergeRels
+			// resolve connected components
-				.joinWith(entities, rawMergeRels.col("_2").equalTo(entities.col("_1")), "inner")
+			// ("vertexId", "groupId")
-				// <tmp_source,target>,<target,entity>
+			Dataset<Row> cliques = ConnectedComponent
-				.map(
+				.runOnPairs(edges, 50, spark);
-					(MapFunction<Tuple2<Tuple2<String, String>, Tuple2<String, OafEntity>>, Tuple2<String, OafEntity>>) value -> new Tuple2<>(
+
-						value._1()._1(), value._2()._2()),
+			// transform "vertexId" back to its original string value
-					Encoders.tuple(Encoders.STRING(), Encoders.kryo(clazz)))
+			// groupId is kept numeric as its string value is not used
-				// <tmp_source,entity>
+			// ("id", "groupId")
-				.groupByKey(
+			Dataset<Row> rawMergeRels = cliques
-					(MapFunction<Tuple2<String, OafEntity>, String>) Tuple2::_1, Encoders.STRING())
+				.join(vertexIdMap, JavaConversions.asScalaBuffer(Collections.singletonList("vertexId")), "inner")
-				.mapGroups(
+				.drop("vertexId")
-					(MapGroupsFunction<String, Tuple2<String, OafEntity>, ConnectedComponent>) this::generateID,
+				.distinct();
-					Encoders.bean(ConnectedComponent.class))
+
-				// <root_id, list(target)>
+			// empty dataframe if historydatabase is not used
+			Dataset<Row> pivotHistory = spark
+				.createDataset(
+					Collections.emptyList(),
+					RowEncoder
+						.apply(StructType.fromDDL("id STRING, lastUsage STRING")));
+
+			if (StringUtils.isNotBlank(pivotHistoryDatabase)) {
+				pivotHistory = spark
+					.read()
+					.table(pivotHistoryDatabase + "." + subEntity)
+					.selectExpr("id", "lastUsage");
+			}
+
+			// depending on resulttype collectefrom and dateofacceptance are evaluated differently
+			String collectedfromExpr = "false AS collectedfrom";
+			String dateExpr = "'' AS date";
+
+			if (Result.class.isAssignableFrom(clazz)) {
+				if (Publication.class.isAssignableFrom(clazz)) {
+					collectedfromExpr = "array_contains(collectedfrom.key, '" + ModelConstants.CROSSREF_ID
+						+ "') AS collectedfrom";
+				} else if (eu.dnetlib.dhp.schema.oaf.Dataset.class.isAssignableFrom(clazz)) {
+					collectedfromExpr = "array_contains(collectedfrom.key, '" + ModelConstants.DATACITE_ID
+						+ "') AS collectedfrom";
+				}
+
+				dateExpr = "dateofacceptance.value AS date";
+			}
+
+			// cap pidType at w3id as from there on they are considered equal
+			UserDefinedFunction mapPid = udf(
+				(String s) -> Math.min(PidType.tryValueOf(s).ordinal(), PidType.w3id.ordinal()), DataTypes.IntegerType);
+
+			UserDefinedFunction validDate = udf((String date) -> {
+				if (StringUtils.isNotBlank(date)
+					&& date.matches(DatePicker.DATE_PATTERN) && DatePicker.inRange(date)) {
+					return date;
+				}
+				return LocalDate.now().plusWeeks(1).toString();
+			}, DataTypes.StringType);
+
+			Dataset<Row> pivotingData = spark
+				.read()
+				.schema(Encoders.bean(clazz).schema())
+				.json(DedupUtility.createEntityPath(graphBasePath, subEntity))
+				.selectExpr(
+					"id",
+					"regexp_extract(id, '^\\\\d+\\\\|([^_]+).*::', 1) AS pidType",
+					collectedfromExpr,
+					dateExpr)
+				.withColumn("pidType", mapPid.apply(col("pidType"))) // ordinal of pid type
+				.withColumn("date", validDate.apply(col("date")));
+
+			// ordering to selected pivot id
+			WindowSpec w = Window
+				.partitionBy("groupId")
+				.orderBy(
+					col("lastUsage").desc_nulls_last(),
+					col("pidType").asc_nulls_last(),
+					col("collectedfrom").desc_nulls_last(),
+					col("date").asc_nulls_last(),
+					col("id").asc_nulls_last());
+
+			Dataset<Relation> output = rawMergeRels
+				.join(pivotHistory, JavaConversions.asScalaBuffer(Collections.singletonList("id")), "full")
+				.join(pivotingData, JavaConversions.asScalaBuffer(Collections.singletonList("id")), "left")
+				.withColumn("pivot", functions.first("id").over(w))
+				.withColumn("position", functions.row_number().over(w))
 				.flatMap(
-					(FlatMapFunction<ConnectedComponent, Relation>) cc -> ccToMergeRel(cc, dedupConf),
+					(FlatMapFunction<Row, Tuple3<String, String, String>>) (Row r) -> {
-					Encoders.bean(Relation.class));
+						String id = r.getAs("id");
+						String dedupId = IdGenerator.generate(id);
 
-			saveParquet(mergeRels, mergeRelPath, SaveMode.Overwrite);
+						String pivot = r.getAs("pivot");
+						String pivotDedupId = IdGenerator.generate(pivot);
 
+						// filter out id == pivotDedupId
+						// those are caused by claim expressed on pivotDedupId
+						// information will be merged after creating deduprecord
+						if (id.equals(pivotDedupId)) {
+							return Collections.emptyIterator();
+						}
+
+						ArrayList<Tuple3<String, String, String>> res = new ArrayList<>();
+
+						// singleton pivots have null groupId as they do not match rawMergeRels
+						if (r.isNullAt(r.fieldIndex("groupId"))) {
+							// the record is existing if it matches pivotingData
+							if (!r.isNullAt(r.fieldIndex("collectedfrom"))) {
+								// create relation with old dedup id
+								res.add(new Tuple3<>(id, dedupId, null));
+							}
+							return res.iterator();
+						}
+
+						// this was a pivot in a previous graph but it has been merged into a new group with different
+						// pivot
+						if (!r.isNullAt(r.fieldIndex("lastUsage")) && !pivot.equals(id)
+							&& !dedupId.equals(pivotDedupId)) {
+							// materialize the previous dedup record as a merge relation with the new one
+							res.add(new Tuple3<>(dedupId, pivotDedupId, null));
+						}
+
+						// add merge relations
+						if (cut <= 0 || r.<Integer> getAs("position") <= cut) {
+							res.add(new Tuple3<>(id, pivotDedupId, pivot));
+						}
+
+						return res.iterator();
+					}, Encoders.tuple(Encoders.STRING(), Encoders.STRING(), Encoders.STRING()))
+				.distinct()
+				.flatMap(
+					(FlatMapFunction<Tuple3<String, String, String>, Relation>) (Tuple3<String, String, String> r) -> {
+						String id = r._1();
+						String dedupId = r._2();
+						String pivot = r._3();
+
+						ArrayList<Relation> res = new ArrayList<>();
+						res.add(rel(pivot, dedupId, id, ModelConstants.MERGES, dedupConf));
+						res.add(rel(pivot, id, dedupId, ModelConstants.IS_MERGED_IN, dedupConf));
+
+						return res.iterator();
+					}, Encoders.bean(Relation.class));
+
+			saveParquet(output, mergeRelPath, SaveMode.Overwrite);
 		}
 	}
 
-	private <T extends OafEntity> ConnectedComponent generateID(String key, Iterator<Tuple2<String, T>> values) {
+	private static Relation rel(String pivot, String source, String target, String relClass, DedupConfig dedupConf) {
-
-		List<Identifier<T>> identifiers = Lists
-			.newArrayList(values)
-			.stream()
-			.map(v -> Identifier.newInstance(v._2()))
-			.collect(Collectors.toList());
-
-		String rootID = IdGenerator.generate(identifiers, key);
-
-		if (Objects.equals(rootID, key))
-			throw new IllegalStateException("generated default ID: " + rootID);
-
-		return new ConnectedComponent(rootID,
-			identifiers.stream().map(i -> i.getEntity().getId()).collect(Collectors.toSet()));
-	}
-
-	private JavaPairRDD<Object, String> createVertexes(JavaSparkContext sc, String graphBasePath, String subEntity,
-		DedupConfig dedupConf) {
-
-		return sc
-			.textFile(DedupUtility.createEntityPath(graphBasePath, subEntity))
-			.mapToPair(json -> {
-				String id = MapDocumentUtil.getJPathString(dedupConf.getWf().getIdPath(), json);
-				return new Tuple2<>(hash(id), id);
-			});
-	}
-
-	private Iterator<Tuple2<String, String>> ccToRels(ConnectedComponent cc) {
-		return cc
-			.getIds()
-			.stream()
-			.map(id -> new Tuple2<>(cc.getCcId(), id))
-			.iterator();
-	}
-
-	private Iterator<Relation> ccToMergeRel(ConnectedComponent cc, DedupConfig dedupConf) {
-		return cc
-			.getIds()
-			.stream()
-			.flatMap(
-				id -> {
-					List<Relation> tmp = new ArrayList<>();
-
-					tmp.add(rel(cc.getCcId(), id, ModelConstants.MERGES, dedupConf));
-					tmp.add(rel(id, cc.getCcId(), ModelConstants.IS_MERGED_IN, dedupConf));
-
-					return tmp.stream();
-				})
-			.iterator();
-	}
-
-	private Relation rel(String source, String target, String relClass, DedupConfig dedupConf) {
 
 		String entityType = dedupConf.getWf().getEntityType();
 
@@ -238,6 +299,14 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
 		// TODO calculate the trust value based on the similarity score of the elements in the CC
 
 		r.setDataInfo(info);
+
+		if (pivot != null) {
+			KeyValue pivotKV = new KeyValue();
+			pivotKV.setKey("pivot");
+			pivotKV.setValue(pivot);
+
+			r.setProperties(Arrays.asList(pivotKV));
+		}
 		return r;
 	}
 

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkWhitelistSimRels.java

@@ -91,18 +91,12 @@ public class SparkWhitelistSimRels extends AbstractSparkAction {
 			Dataset<Row> entities = spark
 				.read()
 				.textFile(DedupUtility.createEntityPath(graphBasePath, subEntity))
-				.repartition(numPartitions)
+				.select(functions.get_json_object(new Column("value"), dedupConf.getWf().getIdPath()).as("id"))
-				.withColumn("id", functions.get_json_object(new Column("value"), dedupConf.getWf().getIdPath()));
+				.distinct();
 
-			Dataset<Row> whiteListRels1 = whiteListRels
+			Dataset<Relation> whiteListSimRels = whiteListRels
-				.join(entities, entities.col("id").equalTo(whiteListRels.col("from")), "inner")
+				.join(entities, entities.col("id").equalTo(whiteListRels.col("from")), "leftsemi")
-				.select("from", "to");
+				.join(entities, functions.col("to").equalTo(entities.col("id")), "leftsemi")
-
-			Dataset<Row> whiteListRels2 = whiteListRels1
-				.join(entities, whiteListRels1.col("to").equalTo(entities.col("id")), "inner")
-				.select("from", "to");
-
-			Dataset<Relation> whiteListSimRels = whiteListRels2
 				.map(
 					(MapFunction<Row, Relation>) r -> DedupUtility
 						.createSimRel(r.getString(0), r.getString(1), entity),

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/graph/ConnectedComponent.java

@@ -1,100 +0,0 @@
-
-package eu.dnetlib.dhp.oa.dedup.graph;
-
-import java.io.IOException;
-import java.io.Serializable;
-import java.util.Set;
-import java.util.stream.Collectors;
-
-import org.apache.commons.lang3.StringUtils;
-import org.codehaus.jackson.annotate.JsonIgnore;
-
-import com.fasterxml.jackson.databind.ObjectMapper;
-
-import eu.dnetlib.dhp.utils.DHPUtils;
-import eu.dnetlib.pace.util.PaceException;
-
-public class ConnectedComponent implements Serializable {
-
-	private String ccId;
-	private Set<String> ids;
-
-	private static final String CONNECTED_COMPONENT_ID_PREFIX = "connect_comp";
-
-	public ConnectedComponent(Set<String> ids, final int cut) {
-		this.ids = ids;
-
-		this.ccId = createDefaultID();
-
-		if (cut > 0 && ids.size() > cut) {
-			this.ids = ids
-				.stream()
-				.filter(id -> !ccId.equalsIgnoreCase(id))
-				.limit(cut - 1)
-				.collect(Collectors.toSet());
-//			this.ids.add(ccId); ??
-		}
-	}
-
-	public ConnectedComponent(String ccId, Set<String> ids) {
-		this.ccId = ccId;
-		this.ids = ids;
-	}
-
-	public String createDefaultID() {
-		if (ids.size() > 1) {
-			final String s = getMin();
-			String prefix = s.split("\\|")[0];
-			ccId = prefix + "|" + CONNECTED_COMPONENT_ID_PREFIX + "::" + DHPUtils.md5(s);
-			return ccId;
-		} else {
-			return ids.iterator().next();
-		}
-	}
-
-	@JsonIgnore
-	public String getMin() {
-
-		final StringBuilder min = new StringBuilder();
-
-		ids
-			.forEach(
-				id -> {
-					if (StringUtils.isBlank(min.toString())) {
-						min.append(id);
-					} else {
-						if (min.toString().compareTo(id) > 0) {
-							min.setLength(0);
-							min.append(id);
-						}
-					}
-				});
-		return min.toString();
-	}
-
-	@Override
-	public String toString() {
-		ObjectMapper mapper = new ObjectMapper();
-		try {
-			return mapper.writeValueAsString(this);
-		} catch (IOException e) {
-			throw new PaceException("Failed to create Json: ", e);
-		}
-	}
-
-	public Set<String> getIds() {
-		return ids;
-	}
-
-	public void setIds(Set<String> ids) {
-		this.ids = ids;
-	}
-
-	public String getCcId() {
-		return ccId;
-	}
-
-	public void setCcId(String ccId) {
-		this.ccId = ccId;
-	}
-}

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/graph/GraphProcessor.scala

@@ -1,37 +0,0 @@
-package eu.dnetlib.dhp.oa.dedup.graph
-
-import org.apache.spark.graphx._
-import org.apache.spark.rdd.RDD
-
-import scala.collection.JavaConversions;
-
-object GraphProcessor {
-
-  def findCCs(vertexes: RDD[(VertexId, String)], edges: RDD[Edge[String]], maxIterations: Int, cut:Int): RDD[ConnectedComponent] = {
-    val graph: Graph[String, String] = Graph(vertexes, edges).partitionBy(PartitionStrategy.RandomVertexCut) //TODO remember to remove partitionby
-    val cc = graph.connectedComponents(maxIterations).vertices
-
-    val joinResult = vertexes.leftOuterJoin(cc).map {
-      case (id, (openaireId, cc)) => {
-        if (cc.isEmpty) {
-          (id, openaireId)
-        }
-        else {
-          (cc.get, openaireId)
-        }
-      }
-    }
-    val connectedComponents = joinResult.groupByKey()
-      .map[ConnectedComponent](cc => asConnectedComponent(cc, cut))
-    connectedComponents
-  }
-
-
-
-  def asConnectedComponent(group: (VertexId, Iterable[String]), cut:Int): ConnectedComponent = {
-    val docs = group._2.toSet[String]
-    val connectedComponent = new ConnectedComponent(JavaConversions.setAsJavaSet[String](docs), cut);
-    connectedComponent
-  }
-
-}

dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/model/Identifier.java

@@ -3,21 +3,21 @@ package eu.dnetlib.dhp.oa.dedup.model;
 
 import java.io.Serializable;
 import java.text.SimpleDateFormat;
-import java.util.*;
+import java.time.LocalDate;
-import java.util.stream.Collectors;
+import java.util.Date;
+import java.util.List;
+import java.util.Objects;
 
 import org.apache.commons.lang3.StringUtils;
 
-import com.google.common.collect.Sets;
-
 import eu.dnetlib.dhp.oa.dedup.DatePicker;
 import eu.dnetlib.dhp.oa.dedup.IdentifierComparator;
 import eu.dnetlib.dhp.schema.common.EntityType;
-import eu.dnetlib.dhp.schema.common.ModelConstants;
 import eu.dnetlib.dhp.schema.common.ModelSupport;
-import eu.dnetlib.dhp.schema.oaf.*;
+import eu.dnetlib.dhp.schema.oaf.Field;
-import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
+import eu.dnetlib.dhp.schema.oaf.KeyValue;
-import eu.dnetlib.dhp.schema.oaf.utils.PidComparator;
+import eu.dnetlib.dhp.schema.oaf.OafEntity;
+import eu.dnetlib.dhp.schema.oaf.Result;
 import eu.dnetlib.dhp.schema.oaf.utils.PidType;
 
 public class Identifier<T extends OafEntity> implements Serializable, Comparable<Identifier<T>> {
@@ -50,7 +50,7 @@ public class Identifier<T extends OafEntity> implements Serializable, Comparable
 		if (Objects.nonNull(date)) {
 			return date;
 		} else {
-			String sDate = BASE_DATE;
+			String sDate = LocalDate.now().plusDays(1).toString();
 			if (ModelSupport.isSubClass(getEntity(), Result.class)) {
 				Result result = (Result) getEntity();
 				if (isWellformed(result.getDateofacceptance())) {

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/createCC_parameters.json

@@ -28,5 +28,17 @@
     "paramLongName": "workingPath",
     "paramDescription": "path for the working directory",
     "paramRequired": true
+  },
+  {
+    "paramName":"h",
+    "paramLongName":"hiveMetastoreUris",
+    "paramDescription": "the hive metastore uris",
+    "paramRequired": true
+  },
+  {
+    "paramName": "p",
+    "paramLongName": "pivotHistoryDatabase",
+    "paramDescription": "Pivot history database",
+    "paramRequired": false
   }
 ]

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/pivothistory/oozie_app/config-default.xml

@@ -0,0 +1,26 @@
+<configuration>
+    <property>
+        <name>jobTracker</name>
+        <value>yarnRM</value>
+    </property>
+    <property>
+        <name>nameNode</name>
+        <value>hdfs://nameservice1</value>
+    </property>
+    <property>
+        <name>oozie.use.system.libpath</name>
+        <value>true</value>
+    </property>
+    <property>
+        <name>oozie.action.sharelib.for.spark</name>
+        <value>spark2</value>
+    </property>
+    <property>
+        <name>hiveMetastoreUris</name>
+        <value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
+    </property>
+    <property>
+        <name>sparkSqlWarehouseDir</name>
+        <value>/user/hive/warehouse</value>
+    </property>
+</configuration>

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/pivothistory/oozie_app/sql.sql

@@ -0,0 +1,62 @@
+
+CREATE TABLE `${pivot_history_db}`.`dataset_new` STORED AS PARQUET AS
+WITH pivots (
+    SELECT property.value AS id, '${new_graph_date}' AS usedIn  FROM `${new_graph_db}`.`relation`
+    LEFT SEMI JOIN `${new_graph_db}`.`dataset` ON relation.source = dataset.id
+    LATERAL VIEW EXPLODE(properties) AS property WHERE relClass = 'isMergedIn' AND property.key = 'pivot'
+UNION
+    SELECT id, usedIn FROM `${pivot_history_db}`.`dataset` LATERAL VIEW EXPLODE(usages) AS usedIn
+)
+SELECT id, min(usedIn) as firstUsage, max(usedIn) as lastUsage, collect_set(usedIn) as usages
+    FROM pivots
+    GROUP BY id; /*EOS*/
+CREATE TABLE `${pivot_history_db}`.`publication_new` STORED AS PARQUET AS
+WITH pivots (
+    SELECT property.value AS id, '${new_graph_date}' AS usedIn  FROM `${new_graph_db}`.`relation`
+    LEFT SEMI JOIN `${new_graph_db}`.`publication` ON relation.source = publication.id
+    LATERAL VIEW EXPLODE(properties) AS property WHERE relClass = 'isMergedIn' AND property.key = 'pivot'
+UNION
+    SELECT id, usedIn FROM `${pivot_history_db}`.`publication` LATERAL VIEW EXPLODE(usages) AS usedIn
+)
+SELECT id, min(usedIn) as firstUsage, max(usedIn) as lastUsage, collect_set(usedIn) as usages
+    FROM pivots
+    GROUP BY id; /*EOS*/
+CREATE TABLE `${pivot_history_db}`.`software_new` STORED AS PARQUET AS
+WITH pivots (
+    SELECT property.value AS id, '${new_graph_date}' AS usedIn  FROM `${new_graph_db}`.`relation`
+    LEFT SEMI JOIN `${new_graph_db}`.`software` ON relation.source = software.id
+    LATERAL VIEW EXPLODE(properties) AS property WHERE relClass = 'isMergedIn' AND property.key = 'pivot'
+UNION
+    SELECT id, usedIn FROM `${pivot_history_db}`.`software` LATERAL VIEW EXPLODE(usages) AS usedIn
+)
+SELECT id, min(usedIn) as firstUsage, max(usedIn) as lastUsage, collect_set(usedIn) as usages
+    FROM pivots
+    GROUP BY id; /*EOS*/
+CREATE TABLE `${pivot_history_db}`.`otherresearchproduct_new` STORED AS PARQUET AS
+WITH pivots (
+    SELECT property.value AS id, '${new_graph_date}' AS usedIn  FROM `${new_graph_db}`.`relation`
+     LEFT SEMI JOIN `${new_graph_db}`.`otherresearchproduct` ON relation.source = otherresearchproduct.id
+     LATERAL VIEW EXPLODE(properties) AS property WHERE relClass = 'isMergedIn' AND property.key = 'pivot'
+UNION
+    SELECT id, usedIn FROM `${pivot_history_db}`.`otherresearchproduct` LATERAL VIEW EXPLODE(usages) AS usedIn
+)
+SELECT id, min(usedIn) as firstUsage, max(usedIn) as lastUsage, collect_set(usedIn) as usages
+    FROM pivots
+    GROUP BY id; /*EOS*/
+
+
+DROP TABLE IF EXISTS `${pivot_history_db}`.`dataset_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`dataset` RENAME TO `${pivot_history_db}`.`dataset_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`dataset_new` RENAME TO `${pivot_history_db}`.`dataset`; /*EOS*/
+
+DROP TABLE IF EXISTS `${pivot_history_db}`.`publication_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`publication` RENAME TO `${pivot_history_db}`.`publication_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`publication_new` RENAME TO `${pivot_history_db}`.`publication`; /*EOS*/
+
+DROP TABLE IF EXISTS `${pivot_history_db}`.`software_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`software` RENAME TO `${pivot_history_db}`.`software_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`software_new` RENAME TO `${pivot_history_db}`.`software`; /*EOS*/
+
+DROP TABLE IF EXISTS `${pivot_history_db}`.`otherresearchproduct_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`otherresearchproduct` RENAME TO `${pivot_history_db}`.`otherresearchproduct_old`; /*EOS*/
+ALTER TABLE `${pivot_history_db}`.`otherresearchproduct_new` RENAME TO `${pivot_history_db}`.`otherresearchproduct`; /*EOS*/

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/pivothistory/oozie_app/workflow.xml

@@ -0,0 +1,95 @@
+<workflow-app name="Update pivot history" xmlns="uri:oozie:workflow:0.5">
+    <parameters>
+        <!-- properties used in SQL -->
+        <property>
+            <name>pivot_history_db</name>
+            <!-- <value>openaire_beta_pivots_test</value> -->
+            <description>Pivot history DB on hive</description>
+        </property>
+        <property>
+            <name>new_graph_db</name>
+            <!--<value>openaire_beta_20231208</value> -->
+            <description>New graph DB on hive</description>
+    </property>
+    <property>
+        <name>new_graph_date</name>
+        <!-- <value>20231208</value> -->
+        <description>Creation date of new graph db</description>
+    </property>
+
+    <!-- RunSQLSparkJob properties -->
+        <property>
+            <name>hiveMetastoreUris</name>
+            <description>hive server metastore URIs</description>
+        </property>
+        <property>
+            <name>sparkSqlWarehouseDir</name>
+        </property>
+        <!-- General oozie workflow properties -->
+        <property>
+            <name>sparkClusterOpts</name>
+            <value>--conf spark.network.timeout=600 --conf spark.extraListeners= --conf spark.sql.queryExecutionListeners= --conf spark.yarn.historyServer.address=http://iis-cdh5-test-m3.ocean.icm.edu.pl:18088 --conf spark.eventLog.dir=hdfs://nameservice1/user/spark/applicationHistory</value>
+            <description>spark cluster-wide options</description>
+        </property>
+        <property>
+            <name>sparkResourceOpts</name>
+            <value>--executor-memory=3G --conf spark.executor.memoryOverhead=3G --executor-cores=6 --driver-memory=8G --driver-cores=4</value>
+            <description>spark resource options</description>
+        </property>
+        <property>
+            <name>sparkApplicationOpts</name>
+            <value>--conf spark.sql.shuffle.partitions=3840</value>
+            <description>spark resource options</description>
+        </property>
+    </parameters>
+
+    <global>
+        <job-tracker>${jobTracker}</job-tracker>
+        <name-node>${nameNode}</name-node>
+        <configuration>
+            <property>
+                <name>mapreduce.job.queuename</name>
+                <value>${queueName}</value>
+            </property>
+            <property>
+                <name>oozie.launcher.mapred.job.queue.name</name>
+                <value>${oozieLauncherQueueName}</value>
+            </property>
+            <property>
+                <name>oozie.action.sharelib.for.spark</name>
+                <value>${oozieActionShareLibForSpark2}</value>
+            </property>
+        </configuration>
+    </global>
+
+    <start to="UpgradePivotHistory"/>
+
+    <kill name="Kill">
+        <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
+    </kill>
+
+    <action name="UpgradePivotHistory">
+        <spark xmlns="uri:oozie:spark-action:0.2">
+            <master>yarn</master>
+            <mode>cluster</mode>
+            <name>Upgrade Pivot History</name>
+            <class>eu.dnetlib.dhp.oozie.RunSQLSparkJob</class>
+            <jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
+            <spark-opts>
+                --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
+                ${sparkClusterOpts}
+                ${sparkResourceOpts}
+                ${sparkApplicationOpts}
+            </spark-opts>
+            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
+            <arg>--sql</arg><arg>eu/dnetlib/dhp/oa/dedup/pivothistory/oozie_app/sql.sql</arg>
+            <arg>--pivot_history_db</arg><arg>${pivot_history_db}</arg>
+            <arg>--new_graph_db</arg><arg>${new_graph_db}</arg>
+            <arg>--new_graph_date</arg><arg>${new_graph_date}</arg>
+        </spark>
+        <ok to="End"/>
+        <error to="Kill"/>
+    </action>
+
+    <end name="End"/>
+</workflow-app>

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/config-default.xml

@@ -15,4 +15,8 @@
         <name>oozie.action.sharelib.for.spark</name>
         <value>spark2</value>
     </property>
+    <property>
+        <name>hiveMetastoreUris</name>
+        <value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
+    </property>
 </configuration>

dhp-workflows/dhp-dedup-openaire/src/main/resources/eu/dnetlib/dhp/oa/dedup/scan/oozie_app/workflow.xml

@@ -188,6 +188,8 @@
             <arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
             <arg>--actionSetId</arg><arg>${actionSetId}</arg>
             <arg>--cutConnectedComponent</arg><arg>${cutConnectedComponent}</arg>
+            <arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
+            <arg>--pivotHistoryDatabase</arg><arg>${pivotHistoryDatabase}</arg>
         </spark>
         <ok to="CreateDedupRecord"/>
         <error to="Kill"/>

dhp-workflows/dhp-dedup-openaire/src/main/scala/com/kwartile/lib/cc/ConnectedComponent.scala

@@ -0,0 +1,335 @@
+/** Copyright (c) 2017 Kwartile, Inc., http://www.kwartile.com
+  * Permission is hereby granted, free of charge, to any person obtaining a copy
+  * of this software and associated documentation files (the "Software"), to deal
+  * in the Software without restriction, including without limitation the rights
+  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+  * copies of the Software, and to permit persons to whom the Software is
+  * furnished to do so, subject to the following conditions:
+  *
+  * The above copyright notice and this permission notice shall be included in all
+  * copies or substantial portions of the Software.
+  *
+  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+  * SOFTWARE.
+  */
+
+/** Map-reduce implementation of Connected Component
+  * Given lists of subgraphs, returns all the nodes that are connected.
+  */
+
+package com.kwartile.lib.cc
+
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.{Dataset, Row, SparkSession}
+import org.apache.spark.storage.StorageLevel
+
+import scala.annotation.tailrec
+import scala.collection.mutable
+
+object ConnectedComponent extends Serializable {
+
+  /** Applies Small Star operation on RDD of nodePairs
+    *
+    * @param nodePairs on which to apply Small Star operations
+    * @return new nodePairs after the operation and conncectivy change count
+    */
+  private def smallStar(nodePairs: RDD[(Long, Long)]): (RDD[(Long, Long)], Long) = {
+
+    /** generate RDD of (self, List(neighbors)) where self > neighbors
+      * E.g.: nodePairs (1, 4), (6, 1), (3, 2), (6, 5)
+      * will result into (4, List(1)), (6, List(1)), (3, List(2)), (6, List(5))
+      */
+    val neighbors = nodePairs.map(x => {
+      val (self, neighbor) = (x._1, x._2)
+      if (self > neighbor)
+        (self, neighbor)
+      else
+        (neighbor, self)
+    })
+
+    /** reduce on self to get list of all its neighbors.
+      * E.g: (4, List(1)), (6, List(1)), (3, List(2)), (6, List(5))
+      * will result into (4, List(1)), (6, List(1, 5)), (3, List(2))
+      * Note:
+      * (1) you may need to tweak number of partitions.
+      * (2) also, watch out for data skew. In that case, consider using rangePartitioner
+      */
+    val empty = mutable.HashSet[Long]()
+    val allNeighbors = neighbors.aggregateByKey(empty)(
+      (lb, v) => lb += v,
+      (lb1, lb2) => lb1 ++ lb2
+    )
+
+    /** Apply Small Star operation on (self, List(neighbor)) to get newNodePairs and count the change in connectivity
+      */
+
+    val newNodePairsWithChangeCount = allNeighbors
+      .map(x => {
+        val self = x._1
+        val neighbors = x._2.toList
+        val minNode = argMin(self :: neighbors)
+        val newNodePairs = (self :: neighbors)
+          .map(neighbor => {
+            (neighbor, minNode)
+          })
+          .filter(x => {
+            val neighbor = x._1
+            val minNode = x._2
+            (neighbor <= self && neighbor != minNode) || (self == neighbor)
+          })
+        val uniqueNewNodePairs = newNodePairs.toSet.toList
+
+        /** We count the change by taking a diff of the new node pairs with the old node pairs
+          */
+        val connectivityChangeCount = (uniqueNewNodePairs diff neighbors.map((self, _))).length
+        (uniqueNewNodePairs, connectivityChangeCount)
+      })
+      .persist(StorageLevel.MEMORY_AND_DISK_SER)
+
+    /** Sum all the changeCounts
+      */
+    val totalConnectivityCountChange = newNodePairsWithChangeCount
+      .mapPartitions(iter => {
+        val (v, l) = iter.toSeq.unzip
+        val sum = l.sum
+        Iterator(sum)
+      })
+      .sum
+      .toLong
+
+    val newNodePairs = newNodePairsWithChangeCount.map(x => x._1).flatMap(x => x)
+    newNodePairsWithChangeCount.unpersist(false)
+    (newNodePairs, totalConnectivityCountChange)
+  }
+
+  /** Apply Large Star operation on a RDD of nodePairs
+    *
+    * @param nodePairs on which to apply Large Star operations
+    * @return new nodePairs after the operation and conncectivy change count
+    */
+  private def largeStar(nodePairs: RDD[(Long, Long)]): (RDD[(Long, Long)], Long) = {
+
+    /** generate RDD of (self, List(neighbors))
+      * E.g.: nodePairs (1, 4), (6, 1), (3, 2), (6, 5)
+      * will result into (4, List(1)), (1, List(4)), (6, List(1)), (1, List(6)), (3, List(2)), (2, List(3)), (6, List(5)), (5, List(6))
+      */
+
+    val neighbors = nodePairs.flatMap(x => {
+      val (self, neighbor) = (x._1, x._2)
+      if (self == neighbor)
+        List((self, neighbor))
+      else
+        List((self, neighbor), (neighbor, self))
+    })
+
+    /** reduce on self to get list of all its neighbors.
+      * E.g: (4, List(1)), (1, List(4)), (6, List(1)), (1, List(6)), (3, List(2)), (2, List(3)), (6, List(5)), (5, List(6))
+      * will result into (4, List(1)), (1, List(4, 6)), (6, List(1, 5)), (3, List(2)), (2, List(3)), (5, List(6))
+      * Note:
+      * (1) you may need to tweak number of partitions.
+      * (2) also, watch out for data skew. In that case, consider using rangePartitioner
+      */
+
+    val localAdd = (s: mutable.HashSet[Long], v: Long) => s += v
+    val partitionAdd = (s1: mutable.HashSet[Long], s2: mutable.HashSet[Long]) => s1 ++= s2
+    val allNeighbors =
+      neighbors.aggregateByKey(mutable.HashSet.empty[Long] /*, rangePartitioner*/ )(localAdd, partitionAdd)
+
+    /** Apply Large Star operation on (self, List(neighbor)) to get newNodePairs and count the change in connectivity
+      */
+
+    val newNodePairsWithChangeCount = allNeighbors
+      .map(x => {
+        val self = x._1
+        val neighbors = x._2.toList
+        val minNode = argMin(self :: neighbors)
+        val newNodePairs = (self :: neighbors)
+          .map(neighbor => {
+            (neighbor, minNode)
+          })
+          .filter(x => {
+            val neighbor = x._1
+            val minNode = x._2
+            neighbor > self || neighbor == minNode
+          })
+
+        val uniqueNewNodePairs = newNodePairs.toSet.toList
+        val connectivityChangeCount = (uniqueNewNodePairs diff neighbors.map((self, _))).length
+        (uniqueNewNodePairs, connectivityChangeCount)
+      })
+      .persist(StorageLevel.MEMORY_AND_DISK_SER)
+
+    val totalConnectivityCountChange = newNodePairsWithChangeCount
+      .mapPartitions(iter => {
+        val (v, l) = iter.toSeq.unzip
+        val sum = l.sum
+        Iterator(sum)
+      })
+      .sum
+      .toLong
+
+    /** Sum all the changeCounts
+      */
+    val newNodePairs = newNodePairsWithChangeCount.map(x => x._1).flatMap(x => x)
+    newNodePairsWithChangeCount.unpersist(false)
+    (newNodePairs, totalConnectivityCountChange)
+  }
+
+  private def argMin(nodes: List[Long]): Long = {
+    nodes.min(Ordering.by((node: Long) => node))
+  }
+
+  /** Build nodePairs given a list of nodes.  A list of nodes represents a subgraph.
+    *
+    * @param nodes that are part of a subgraph
+    * @return nodePairs for a subgraph
+    */
+  private def buildPairs(nodes: List[Long]): List[(Long, Long)] = {
+    buildPairs(nodes.head, nodes.tail, null.asInstanceOf[List[(Long, Long)]])
+  }
+
+  @tailrec
+  private def buildPairs(node: Long, neighbors: List[Long], partialPairs: List[(Long, Long)]): List[(Long, Long)] = {
+    if (neighbors.isEmpty) {
+      if (partialPairs != null)
+        List((node, node)) ::: partialPairs
+      else
+        List((node, node))
+    } else if (neighbors.length == 1) {
+      val neighbor = neighbors(0)
+      if (node > neighbor)
+        if (partialPairs != null) List((node, neighbor)) ::: partialPairs else List((node, neighbor))
+      else if (partialPairs != null) List((neighbor, node)) ::: partialPairs
+      else List((neighbor, node))
+    } else {
+      val newPartialPairs = neighbors
+        .map(neighbor => {
+          if (node > neighbor)
+            List((node, neighbor))
+          else
+            List((neighbor, node))
+        })
+        .flatMap(x => x)
+
+      if (partialPairs != null)
+        buildPairs(neighbors.head, neighbors.tail, newPartialPairs ::: partialPairs)
+      else
+        buildPairs(neighbors.head, neighbors.tail, newPartialPairs)
+    }
+  }
+
+  /** Implements alternatingAlgo.  Converges when the changeCount is either 0 or does not change from the previous iteration
+    *
+    * @param nodePairs                        for a graph
+    * @param largeStarConnectivityChangeCount change count that resulted from the previous iteration
+    * @param smallStarConnectivityChangeCount change count that resulted from the previous iteration
+    * @param didConverge                      flag to indicate the alorigth converged
+    * @param currIterationCount               counter to capture number of iterations
+    * @param maxIterationCount                maximum number iterations to try before giving up
+    * @return RDD of nodePairs
+    */
+
+  @tailrec
+  private def alternatingAlgo(
+    nodePairs: RDD[(Long, Long)],
+    largeStarConnectivityChangeCount: Long,
+    smallStarConnectivityChangeCount: Long,
+    didConverge: Boolean,
+    currIterationCount: Int,
+    maxIterationCount: Int
+  ): (RDD[(Long, Long)], Boolean, Long) = {
+
+    val iterationCount = currIterationCount + 1
+    if (didConverge)
+      (nodePairs, true, currIterationCount)
+    else if (currIterationCount >= maxIterationCount) {
+      (nodePairs, false, currIterationCount)
+    } else {
+
+      val (nodePairsLargeStar, currLargeStarConnectivityChangeCount) = largeStar(nodePairs)
+      val (nodePairsSmallStar, currSmallStarConnectivityChangeCount) = smallStar(nodePairsLargeStar)
+
+      if (
+        (currLargeStarConnectivityChangeCount == largeStarConnectivityChangeCount &&
+        currSmallStarConnectivityChangeCount == smallStarConnectivityChangeCount) ||
+        (currSmallStarConnectivityChangeCount == 0 && currLargeStarConnectivityChangeCount == 0)
+      ) {
+        alternatingAlgo(
+          nodePairsSmallStar,
+          currLargeStarConnectivityChangeCount,
+          currSmallStarConnectivityChangeCount,
+          true,
+          iterationCount,
+          maxIterationCount
+        )
+      } else {
+        alternatingAlgo(
+          nodePairsSmallStar,
+          currLargeStarConnectivityChangeCount,
+          currSmallStarConnectivityChangeCount,
+          false,
+          iterationCount,
+          maxIterationCount
+        )
+      }
+    }
+  }
+
+  /** Driver function
+    *
+    * @param cliques           list of nodes representing subgraphs (or cliques)
+    * @param maxIterationCount maximum number iterations to try before giving up
+    * @return Connected Components as nodePairs where second member of the nodePair is the minimum node in the component
+    */
+  def run(cliques: RDD[List[Long]], maxIterationCount: Int): (RDD[(Long, Long)], Boolean, Long) = {
+
+    val nodePairs = cliques
+      .map(aClique => {
+        buildPairs(aClique)
+      })
+      .flatMap(x => x)
+
+    val (cc, didConverge, iterCount) = alternatingAlgo(nodePairs, 9999999L, 9999999L, false, 0, maxIterationCount)
+
+    if (didConverge) {
+      (cc, didConverge, iterCount)
+    } else {
+      (null.asInstanceOf[RDD[(Long, Long)]], didConverge, iterCount)
+    }
+  }
+
+  def runOnPairs(nodePairs: RDD[(Long, Long)], maxIterationCount: Int): (RDD[(Long, Long)], Boolean, Long) = {
+    val (cc, didConverge, iterCount) = alternatingAlgo(nodePairs, 9999999L, 9999999L, false, 0, maxIterationCount)
+
+    if (didConverge) {
+      (cc, didConverge, iterCount)
+    } else {
+      (null.asInstanceOf[RDD[(Long, Long)]], didConverge, iterCount)
+    }
+  }
+
+  def runOnPairs(nodePairs: Dataset[Row], maxIterationCount: Int)(implicit spark: SparkSession): Dataset[Row] = {
+    import spark.implicits._
+
+    val (cc, didConverge, iterCount) = alternatingAlgo(
+      nodePairs.map(e => (e.getLong(0), e.getLong(1))).rdd,
+      9999999L,
+      9999999L,
+      false,
+      0,
+      maxIterationCount
+    )
+
+    if (didConverge) {
+      cc.toDF("vertexId", "groupId")
+    } else {
+      null.asInstanceOf[Dataset[Row]]
+    }
+  }
+
+}

dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java

@@ -41,9 +41,13 @@ import com.google.common.collect.Sets;
 
 import eu.dnetlib.dhp.application.ArgumentApplicationParser;
 import eu.dnetlib.dhp.schema.common.ModelConstants;
+import eu.dnetlib.dhp.schema.common.ModelSupport;
 import eu.dnetlib.dhp.schema.oaf.*;
+import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
+import eu.dnetlib.dhp.schema.sx.OafUtils;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
+import scala.Tuple2;
 
 @ExtendWith(MockitoExtension.class)
 @TestMethodOrder(MethodOrderer.OrderAnnotation.class)
@@ -97,6 +101,7 @@ public class SparkDedupTest implements Serializable {
 
 		final SparkConf conf = new SparkConf();
 		conf.set("spark.sql.shuffle.partitions", "200");
+		conf.set("spark.sql.warehouse.dir", testOutputBasePath + "/spark-warehouse");
 		spark = SparkSession
 			.builder()
 			.appName(SparkDedupTest.class.getSimpleName())
@@ -186,11 +191,11 @@ public class SparkDedupTest implements Serializable {
 		System.out.println("ds_simrel = " + ds_simrel);
 		System.out.println("orp_simrel = " + orp_simrel);
 
-		assertEquals(1538, orgs_simrel);
+		assertEquals(751, orgs_simrel);
-		assertEquals(3523, pubs_simrel);
+		assertEquals(546, pubs_simrel);
-		assertEquals(168, sw_simrel);
+		assertEquals(113, sw_simrel);
-		assertEquals(221, ds_simrel);
+		assertEquals(148, ds_simrel);
-		assertEquals(3392, orp_simrel);
+		assertEquals(280, orp_simrel);
 
 	}
 
@@ -235,10 +240,10 @@ public class SparkDedupTest implements Serializable {
 			.count();
 
 		// entities simrels supposed to be equal to the number of previous step (no rels in whitelist)
-		assertEquals(1538, orgs_simrel);
+		assertEquals(751, orgs_simrel);
-		assertEquals(3523, pubs_simrel);
+		assertEquals(546, pubs_simrel);
-		assertEquals(221, ds_simrel);
+		assertEquals(148, ds_simrel);
-		assertEquals(3392, orp_simrel);
+		assertEquals(280, orp_simrel);
 //		System.out.println("orgs_simrel = " + orgs_simrel);
 //		System.out.println("pubs_simrel = " + pubs_simrel);
 //		System.out.println("ds_simrel = " + ds_simrel);
@@ -268,7 +273,7 @@ public class SparkDedupTest implements Serializable {
 						&& rel.getTarget().equalsIgnoreCase(whiteList.get(1).split(WHITELIST_SEPARATOR)[1]))
 				.count() > 0);
 
-		assertEquals(170, sw_simrel.count());
+		assertEquals(115, sw_simrel.count());
 //		System.out.println("sw_simrel = " + sw_simrel.count());
 
 	}
@@ -292,7 +297,9 @@ public class SparkDedupTest implements Serializable {
 					"-w",
 					testOutputBasePath,
 					"-cc",
-					"3"
+					"3",
+					"-h",
+					""
 				});
 
 		new SparkCreateMergeRels(parser, spark).run(isLookUpService);
@@ -365,6 +372,113 @@ public class SparkDedupTest implements Serializable {
 			.deleteDirectory(new File(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel"));
 	}
 
+	@Test
+	@Order(3)
+	void createMergeRelsWithPivotHistoryTest() throws Exception {
+
+		ArgumentApplicationParser parser = new ArgumentApplicationParser(
+			classPathResourceAsString("/eu/dnetlib/dhp/oa/dedup/createCC_parameters.json"));
+
+		spark.sql("CREATE DATABASE IF NOT EXISTS pivot_history_test");
+		ModelSupport.oafTypes.keySet().forEach(entityType -> {
+			try {
+				spark
+					.read()
+					.json(
+						Paths
+							.get(SparkDedupTest.class.getResource("/eu/dnetlib/dhp/dedup/pivot_history").toURI())
+							.toFile()
+							.getAbsolutePath())
+					.write()
+					.mode("overwrite")
+					.saveAsTable("pivot_history_test." + entityType);
+			} catch (URISyntaxException e) {
+				throw new RuntimeException(e);
+			}
+		});
+
+		parser
+			.parseArgument(
+				new String[] {
+					"-i",
+					testGraphBasePath,
+					"-asi",
+					testActionSetId,
+					"-la",
+					"lookupurl",
+					"-w",
+					testOutputBasePath,
+					"-h",
+					"",
+					"-pivotHistoryDatabase",
+					"pivot_history_test"
+
+				});
+
+		new SparkCreateMergeRels(parser, spark).run(isLookUpService);
+
+		long orgs_mergerel = spark
+			.read()
+			.load(testOutputBasePath + "/" + testActionSetId + "/organization_mergerel")
+			.count();
+		final Dataset<Relation> pubs = spark
+			.read()
+			.load(testOutputBasePath + "/" + testActionSetId + "/publication_mergerel")
+			.as(Encoders.bean(Relation.class));
+		long sw_mergerel = spark
+			.read()
+			.load(testOutputBasePath + "/" + testActionSetId + "/software_mergerel")
+			.count();
+		long ds_mergerel = spark
+			.read()
+			.load(testOutputBasePath + "/" + testActionSetId + "/dataset_mergerel")
+			.count();
+
+		long orp_mergerel = spark
+			.read()
+			.load(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel")
+			.count();
+
+		final List<Relation> merges = pubs
+			.filter("source == '50|arXiv_dedup_::c93aeb433eb90ed7a86e29be00791b7c'")
+			.collectAsList();
+		assertEquals(3, merges.size());
+		Set<String> dups = Sets
+			.newHashSet(
+				"50|doi_________::3b1d0d8e8f930826665df9d6b82fbb73",
+				"50|doi_________::d5021b53204e4fdeab6ff5d5bc468032",
+				"50|arXiv_______::c93aeb433eb90ed7a86e29be00791b7c");
+		merges.forEach(r -> {
+			assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
+			assertEquals(ModelConstants.DEDUP, r.getSubRelType());
+			assertEquals(ModelConstants.MERGES, r.getRelClass());
+			assertTrue(dups.contains(r.getTarget()));
+		});
+
+		final List<Relation> mergedIn = pubs
+			.filter("target == '50|arXiv_dedup_::c93aeb433eb90ed7a86e29be00791b7c'")
+			.collectAsList();
+		assertEquals(3, mergedIn.size());
+		mergedIn.forEach(r -> {
+			assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
+			assertEquals(ModelConstants.DEDUP, r.getSubRelType());
+			assertEquals(ModelConstants.IS_MERGED_IN, r.getRelClass());
+			assertTrue(dups.contains(r.getSource()));
+		});
+
+		assertEquals(1268, orgs_mergerel);
+		assertEquals(1112, pubs.count());
+		assertEquals(292, sw_mergerel);
+		assertEquals(476, ds_mergerel);
+		assertEquals(742, orp_mergerel);
+//		System.out.println("orgs_mergerel = " + orgs_mergerel);
+//		System.out.println("pubs_mergerel = " + pubs_mergerel);
+//		System.out.println("sw_mergerel = " + sw_mergerel);
+//		System.out.println("ds_mergerel = " + ds_mergerel);
+//		System.out.println("orp_mergerel = " + orp_mergerel);
+
+	}
+
 	@Test
 	@Order(4)
 	void createMergeRelsTest() throws Exception {
@@ -382,7 +496,9 @@ public class SparkDedupTest implements Serializable {
 					"-la",
 					"lookupurl",
 					"-w",
-					testOutputBasePath
+					testOutputBasePath,
+					"-h",
+					""
 				});
 
 		new SparkCreateMergeRels(parser, spark).run(isLookUpService);
@@ -437,10 +553,10 @@ public class SparkDedupTest implements Serializable {
 		});
 
 		assertEquals(1268, orgs_mergerel);
-		assertEquals(1450, pubs.count());
+		assertEquals(1112, pubs.count());
-		assertEquals(286, sw_mergerel);
+		assertEquals(292, sw_mergerel);
-		assertEquals(472, ds_mergerel);
+		assertEquals(476, ds_mergerel);
-		assertEquals(738, orp_mergerel);
+		assertEquals(742, orp_mergerel);
 //		System.out.println("orgs_mergerel = " + orgs_mergerel);
 //		System.out.println("pubs_mergerel = " + pubs_mergerel);
 //		System.out.println("sw_mergerel = " + sw_mergerel);
@@ -492,8 +608,8 @@ public class SparkDedupTest implements Serializable {
 			.count();
 
 		assertEquals(86, orgs_deduprecord);
-		assertEquals(68, pubs.count());
+		assertEquals(91, pubs.count());
-		assertEquals(49, sw_deduprecord);
+		assertEquals(47, sw_deduprecord);
 		assertEquals(97, ds_deduprecord);
 		assertEquals(92, orp_deduprecord);
 
@@ -629,11 +745,11 @@ public class SparkDedupTest implements Serializable {
 			.distinct()
 			.count();
 
-		assertEquals(902, publications);
+		assertEquals(925, publications);
 		assertEquals(839, organizations);
 		assertEquals(100, projects);
 		assertEquals(100, datasource);
-		assertEquals(198, softwares);
+		assertEquals(196, softwares);
 		assertEquals(389, dataset);
 		assertEquals(520, otherresearchproduct);
 

dhp-workflows/dhp-dedup-openaire/src/test/resources/eu/dnetlib/dhp/dedup/conf/ds.curr.conf.json

@@ -101,7 +101,8 @@
         "type" : "String",
         "path" : "$.title[?(@.qualifier.classid == 'main title')].value",
         "length" : 250,
-        "size" : 5
+        "size" : 5,
+        "clean": "title"
       },
       {
         "name" : "authors",

dhp-workflows/dhp-dedup-openaire/src/test/resources/eu/dnetlib/dhp/dedup/conf/orp.curr.conf.json

@@ -101,7 +101,8 @@
         "type" : "String",
         "path" : "$.title[?(@.qualifier.classid == 'main title')].value",
         "length" : 250,
-        "size" : 5
+        "size" : 5,
+        "clean": "title"
       },
       {
         "name" : "authors",

dhp-workflows/dhp-dedup-openaire/src/test/resources/eu/dnetlib/dhp/dedup/conf/pub.curr.conf.json

@@ -29,9 +29,8 @@
   },
   "pace": {
     "clustering" : [
-      { "name" : "ngrampairs", "fields" : [ "title" ], "params" : { "max" : "1", "ngramLen" : "3"} },
+      { "name" : "numAuthorsTitleSuffixPrefixChain", "fields" : [ "num_authors", "title" ], "params" : { "mod" : "10" } },
-      { "name" : "suffixprefix", "fields" : [ "title" ], "params" : { "max" : "1", "len" : "3" } },
+      { "name" : "jsonlistclustering", "fields" : [ "pid" ], "params" : { "jpath_value": "$.value", "jpath_classid": "$.qualifier.classid"} }
-      { "name" : "lowercase", "fields" : [ "doi" ], "params" : { } }
     ],
     "decisionTree": {
       "start": {
@@ -79,13 +78,37 @@
         "ignoreUndefined": "false"
       },
       "layer3": {
+        "fields": [
+          {
+            "field": "authors",
+            "comparator": "authorsMatch",
+            "weight": 1.0,
+            "countIfUndefined": "false",
+            "params": {
+              "surname_th": 0.75,
+              "fullname_th": 0.75,
+              "threshold": 0.6,
+              "mode": "full"
+            }
+          }
+        ],
+        "threshold": 0.6,
+        "aggregation": "MAX",
+        "positive": "layer4",
+        "negative": "NO_MATCH",
+        "undefined": "MATCH",
+        "ignoreUndefined": "true"
+      },
+      "layer4": {
         "fields": [
           {
             "field": "title",
             "comparator": "levensteinTitle",
             "weight": 1.0,
             "countIfUndefined": "true",
-            "params": {}
+            "params": {
+              "threshold": "0.99"
+            }
           }
         ],
         "threshold": 0.99,
@@ -97,23 +120,25 @@
       }
     },
     "model": [
-      {
-        "name": "doi",
-        "type": "String",
-        "path": "$.pid[?(@.qualifier.classid == 'doi')].value"
-      },
       {
         "name": "pid",
         "type": "JSON",
         "path": "$.pid",
         "overrideMatch": "true"
       },
+      {
+        "name": "alternateid",
+        "type": "JSON",
+        "path": "$.instance[*].alternateIdentifier[*]",
+        "overrideMatch": "true"
+      },
       {
         "name": "title",
         "type": "String",
         "path": "$.title[?(@.qualifier.classid == 'main title')].value",
         "length": 250,
-        "size": 5
+        "size": 5,
+        "clean": "title"
       },
       {
         "name": "authors",
@@ -122,9 +147,9 @@
         "size": 200
       },
       {
-        "name": "resulttype",
+        "name": "num_authors",
         "type": "String",
-        "path": "$.resulttype.classid"
+        "path": "$.author.length()"
       }
     ],
     "blacklists": {

dhp-workflows/dhp-dedup-openaire/src/test/resources/eu/dnetlib/dhp/dedup/conf/sw.curr.conf.json

@@ -75,7 +75,8 @@
             "type" : "String",
             "path" : "$.title[?(@.qualifier.classid == 'main title')].value",
             "length" : 250,
-            "size" : 5
+            "size" : 5,
+            "clean": "title"
           },
           {
             "name" : "url",

dhp-workflows/dhp-dedup-openaire/src/test/resources/eu/dnetlib/dhp/dedup/pivot_history/pivot_history.json

@@ -0,0 +1 @@
+{"id": "50|arXiv_______::c93aeb433eb90ed7a86e29be00791b7c", "firstUsage": "2022-01-01", "lastUsage": "2022-01-01", "dedupId": "50|arXiv_dedup_::c93aeb433eb90ed7a86e29be00791b7c" }

dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/crossref/Crossref2Oaf.scala

@@ -25,7 +25,7 @@ case class mappingAffiliation(name: String) {}
 
 case class mappingAuthor(
   given: Option[String],
-  family: String,
+  family: Option[String],
   sequence: Option[String],
   ORCID: Option[String],
   affiliation: Option[mappingAffiliation]
@@ -226,14 +226,14 @@ case object Crossref2Oaf {
 
     //Mapping Author
     val authorList: List[mappingAuthor] =
-      (json \ "author").extractOrElse[List[mappingAuthor]](List())
+      (json \ "author").extract[List[mappingAuthor]].filter(a => a.family.isDefined)
 
     val sorted_list = authorList.sortWith((a: mappingAuthor, b: mappingAuthor) =>
       a.sequence.isDefined && a.sequence.get.equalsIgnoreCase("first")
     )
 
     result.setAuthor(sorted_list.zipWithIndex.map { case (a, index) =>
-      generateAuhtor(a.given.orNull, a.family, a.ORCID.orNull, index)
+      generateAuhtor(a.given.orNull, a.family.get, a.ORCID.orNull, index)
     }.asJava)
 
     // Mapping instance

dhp-workflows/dhp-doiboost/src/test/scala/eu/dnetlib/dhp/doiboost/crossref/CrossrefMappingTest.scala

@@ -22,6 +22,13 @@ class CrossrefMappingTest {
   val logger: Logger = LoggerFactory.getLogger(Crossref2Oaf.getClass)
   val mapper = new ObjectMapper()
 
+  @Test
+  def testMissingAuthorParser():Unit = {
+    val json: String = Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/doiboost/crossref/s41567-022-01757-y.json")).mkString
+    val result = Crossref2Oaf.convert(json)
+    result.filter(o => o.isInstanceOf[Publication]).map(p=> p.asInstanceOf[Publication]).foreach(p =>assertTrue(p.getAuthor.size()>0))
+  }
+
   @Test
   def testFunderRelationshipsMapping(): Unit = {
     val template = Source

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/MoveResult.java

@@ -0,0 +1,84 @@
+
+package eu.dnetlib.dhp;
+
+import static eu.dnetlib.dhp.PropagationConstant.isSparkSessionManaged;
+import static eu.dnetlib.dhp.PropagationConstant.readPath;
+import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
+
+import java.io.Serializable;
+
+import org.apache.commons.io.IOUtils;
+import org.apache.spark.SparkConf;
+import org.apache.spark.sql.Dataset;
+import org.apache.spark.sql.SaveMode;
+import org.apache.spark.sql.SparkSession;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import eu.dnetlib.dhp.application.ArgumentApplicationParser;
+import eu.dnetlib.dhp.resulttocommunityfromorganization.SparkResultToCommunityFromOrganizationJob;
+import eu.dnetlib.dhp.schema.common.ModelSupport;
+import eu.dnetlib.dhp.schema.oaf.Result;
+
+/**
+ * @author miriam.baglioni
+ * @Date 15/01/24
+ */
+public class MoveResult implements Serializable {
+	private static final Logger log = LoggerFactory.getLogger(MoveResult.class);
+
+	public static void main(String[] args) throws Exception {
+		String jsonConfiguration = IOUtils
+			.toString(
+					MoveResult.class
+					.getResourceAsStream(
+						"/eu/dnetlib/dhp/wf/subworkflows/input_moveresult_parameters.json"));
+
+		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
+
+		parser.parseArgument(args);
+
+		Boolean isSparkSessionManaged = isSparkSessionManaged(parser);
+		log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
+
+		String inputPath = parser.get("sourcePath");
+		log.info("inputPath: {}", inputPath);
+
+		final String outputPath = parser.get("outputPath");
+		log.info("outputPath: {}", outputPath);
+
+		SparkConf conf = new SparkConf();
+
+		runWithSparkSession(
+			conf,
+			isSparkSessionManaged,
+			spark -> {
+				moveResults(spark, inputPath, outputPath);
+
+			});
+	}
+
+	public static <R extends Result> void moveResults(SparkSession spark, String inputPath, String outputPath) {
+
+		ModelSupport.entityTypes
+			.keySet()
+			.parallelStream()
+			.filter(e -> ModelSupport.isResult(e))
+			// .parallelStream()
+			.forEach(e -> {
+				Class<R> resultClazz = ModelSupport.entityTypes.get(e);
+				Dataset<R> resultDataset = readPath(spark, inputPath + e.name(), resultClazz);
+				if (resultDataset.count() > 0) {
+
+					resultDataset
+						.write()
+						.mode(SaveMode.Overwrite)
+						.option("compression", "gzip")
+						.json(outputPath + e.name());
+				}
+
+			});
+
+	}
+
+}

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/Utils.java

@@ -167,4 +167,11 @@ public class Utils implements Serializable {
 			});
 		return projectMap;
 	}
+
+	public static List<String> getCommunityIdList(String baseURL) throws IOException {
+		return getValidCommunities(baseURL)
+			.stream()
+			.map(community -> community.getId())
+			.collect(Collectors.toList());
+	}
 }

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

@@ -45,7 +45,7 @@ public class SparkBulkTagJob {
 			.toString(
 				SparkBulkTagJob.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/bulktag/input_bulkTag_parameters.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/bulktag/input_bulkTag_parameters.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 		parser.parseArgument(args);
@@ -105,7 +105,6 @@ public class SparkBulkTagJob {
 		Map<String, List<Pair<String, SelectionConstraints>>> dsm = cc.getEoscDatasourceMap();
 
 		for (String ds : datasources.collectAsList()) {
-			// final String dsId = ds.substring(3);
 			if (!dsm.containsKey(ds)) {
 				ArrayList<Pair<String, SelectionConstraints>> eoscList = new ArrayList<>();
 				dsm.put(ds, eoscList);
@@ -116,13 +115,11 @@ public class SparkBulkTagJob {
 
 	private static boolean isOKDatasource(Datasource ds) {
 		final String compatibility = ds.getOpenairecompatibility().getClassid();
-		boolean isOk = (compatibility.equalsIgnoreCase(OPENAIRE_3) ||
+		return (compatibility.equalsIgnoreCase(OPENAIRE_3) ||
 			compatibility.equalsIgnoreCase(OPENAIRE_4) ||
 			compatibility.equalsIgnoreCase(OPENAIRE_CRIS) ||
 			compatibility.equalsIgnoreCase(OPENAIRE_DATA)) &&
 			ds.getCollectedfrom().stream().anyMatch(cf -> cf.getKey().equals(EOSC));
-
-		return isOk;
 	}
 
 	private static <R extends Result> void execBulkTag(
@@ -151,7 +148,13 @@ public class SparkBulkTagJob {
 					.write()
 					.mode(SaveMode.Overwrite)
 					.option("compression", "gzip")
-					.json(outputPath + e.name());
+					.json(outputPath + e.name());// writing the tagging in the working dir for entity
+
+				readPath(spark, outputPath + e.name(), resultClazz) // copy the tagging in the actual result output path
+					.write()
+					.mode(SaveMode.Overwrite)
+					.option("compression", "gzip")
+					.json(inputPath + e.name());
 			});
 
 	}

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/PrepareDatasourceCountryAssociation.java

@@ -45,7 +45,7 @@ public class PrepareDatasourceCountryAssociation {
 			.toString(
 				PrepareDatasourceCountryAssociation.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/countrypropagation/input_prepareassoc_parameters.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/countrypropagation/input_prepareassoc_parameters.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 
@@ -66,7 +66,7 @@ public class PrepareDatasourceCountryAssociation {
 			conf,
 			isSparkSessionManaged,
 			spark -> {
-				removeOutputDir(spark, outputPath);
+				// removeOutputDir(spark, outputPath);
 				prepareDatasourceCountryAssociation(
 					spark,
 					Arrays.asList(parser.get("whitelist").split(";")),
@@ -90,7 +90,8 @@ public class PrepareDatasourceCountryAssociation {
 				(FilterFunction<Datasource>) ds -> !ds.getDataInfo().getDeletedbyinference() &&
 					Optional.ofNullable(ds.getDatasourcetype()).isPresent() &&
 					Optional.ofNullable(ds.getDatasourcetype().getClassid()).isPresent() &&
-					(allowedtypes.contains(ds.getDatasourcetype().getClassid()) ||
+					((Optional.ofNullable(ds.getJurisdiction()).isPresent() &&
+						allowedtypes.contains(ds.getJurisdiction().getClassid())) ||
 						whitelist.contains(ds.getId())));
 
 		// filtering of the relations taking the non deleted by inference and those with IsProvidedBy as relclass

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/PrepareResultCountrySet.java

@@ -32,7 +32,7 @@ public class PrepareResultCountrySet {
 			.toString(
 				PrepareResultCountrySet.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/countrypropagation/input_prepareresultcountry_parameters.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/countrypropagation/input_prepareresultcountry_parameters.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/SparkCountryPropagationJob.java

@@ -35,7 +35,7 @@ public class SparkCountryPropagationJob {
 			.toString(
 				SparkCountryPropagationJob.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/countrypropagation/input_countrypropagation_parameters.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/countrypropagation/input_countrypropagation_parameters.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/entitytoorganizationfromsemrel/PrepareInfo.java

@@ -60,7 +60,7 @@ public class PrepareInfo implements Serializable {
 			.toString(
 				SparkResultToOrganizationFromIstRepoJob.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/entitytoorganizationfromsemrel/input_preparation_parameter.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/entitytoorganizationfromsemrel/input_preparation_parameter.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/entitytoorganizationfromsemrel/SparkEntityToOrganizationFromSemRel.java

@@ -27,8 +27,8 @@ import eu.dnetlib.dhp.resulttoorganizationfrominstrepo.SparkResultToOrganization
 import eu.dnetlib.dhp.schema.common.ModelConstants;
 import eu.dnetlib.dhp.schema.oaf.Relation;
 
-public class SparkResultToOrganizationFromSemRel implements Serializable {
+public class SparkEntityToOrganizationFromSemRel implements Serializable {
-	private static final Logger log = LoggerFactory.getLogger(SparkResultToOrganizationFromSemRel.class);
+	private static final Logger log = LoggerFactory.getLogger(SparkEntityToOrganizationFromSemRel.class);
 	private static final int MAX_ITERATION = 5;
 	public static final String NEW_RESULT_RELATION_PATH = "/newResultRelation";
 	public static final String NEW_PROJECT_RELATION_PATH = "/newProjectRelation";
@@ -39,7 +39,7 @@ public class SparkResultToOrganizationFromSemRel implements Serializable {
 			.toString(
 				SparkResultToOrganizationFromIstRepoJob.class
 					.getResourceAsStream(
-						"/eu/dnetlib/dhp/entitytoorganizationfromsemrel/input_propagation_parameter.json"));
+						"/eu/dnetlib/dhp/wf/subworkflows/entitytoorganizationfromsemrel/input_propagation_parameter.json"));
 
 		final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
 

dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/entitytoorganizationfromsemrel/StepActions.java

@@ -3,8 +3,8 @@ package eu.dnetlib.dhp.entitytoorganizationfromsemrel;
 
 import static eu.dnetlib.dhp.PropagationConstant.*;
 import static eu.dnetlib.dhp.PropagationConstant.readPath;
-import static eu.dnetlib.dhp.entitytoorganizationfromsemrel.SparkResultToOrganizationFromSemRel.NEW_PROJECT_RELATION_PATH;
+import static eu.dnetlib.dhp.entitytoorganizationfromsemrel.SparkEntityToOrganizationFromSemRel.NEW_PROJECT_RELATION_PATH;
-import static eu.dnetlib.dhp.entitytoorganizationfromsemrel.SparkResultToOrganizationFromSemRel.NEW_RESULT_RELATION_PATH;
+import static eu.dnetlib.dhp.entitytoorganizationfromsemrel.SparkEntityToOrganizationFromSemRel.NEW_RESULT_RELATION_PATH;
 
 import java.io.Serializable;
 import java.util.*;
@@ -20,7 +20,6 @@ import org.jetbrains.annotations.NotNull;
 import com.fasterxml.jackson.databind.ObjectMapper;
 
 import eu.dnetlib.dhp.KeyValueSet;
-import eu.dnetlib.dhp.schema.common.ModelConstants;
 import eu.dnetlib.dhp.schema.oaf.Relation;
 import scala.Tuple2;