From 13f28fa225d248e080c13d99694c3069826b3184 Mon Sep 17 00:00:00 2001 From: Enrico Ottonello Date: Thu, 12 Nov 2020 17:43:32 +0100 Subject: [PATCH] moved AuthorData to dhp-schemas; added other names to author data --- .../dnetlib/dhp/schema/orcid}/AuthorData.java | 18 +- .../orcid/SparkGenerateDoiAuthorList.java | 5 +- .../doiboost/orcid/SummariesDecompressor.java | 7 +- .../doiboost/orcid/xml/XMLRecordParser.java | 8 +- .../SparkGenEnrichedOrcidWorks.java | 38 +- .../doiboost/orcidnodoi/json/JsonWriter.java | 19 +- .../orcidnodoi/model/Contributor.java | 2 +- .../orcidnodoi/similarity/AuthorMatcher.java | 11 +- .../oozie_app/config-default.xml | 2 +- .../orcid/xml/XMLRecordParserTest.java | 21 +- .../orcidnodoi/xml/OrcidNoDoiTest.java | 12 +- .../orcid/xml/record_8888-8888-8888-8880.xml | 770 ++++++++++++++++++ .../summary_0000-0001-5109-1000_othername.xml | 196 +++++ 13 files changed, 1053 insertions(+), 56 deletions(-) rename {dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/model => dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/orcid}/AuthorData.java (71%) create mode 100644 dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/record_8888-8888-8888-8880.xml create mode 100644 dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/summary_0000-0001-5109-1000_othername.xml diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/model/AuthorData.java b/dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/orcid/AuthorData.java similarity index 71% rename from dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/model/AuthorData.java rename to dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/orcid/AuthorData.java index e0624509b..6c94cdb13 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/model/AuthorData.java +++ b/dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/orcid/AuthorData.java @@ -1,7 +1,10 @@ -package eu.dnetlib.doiboost.orcid.model; +package eu.dnetlib.dhp.schema.orcid; import java.io.Serializable; +import java.util.List; + +import com.google.common.collect.Lists; /** * This class models the data that are retrieved from orcid publication @@ -13,8 +16,8 @@ public class AuthorData implements Serializable { private String name; private String surname; private String creditName; - private String otherName; private String errorCode; + private List otherNames; public String getErrorCode() { return errorCode; @@ -56,11 +59,14 @@ public class AuthorData implements Serializable { this.oid = oid; } - public String getOtherName() { - return otherName; + public List getOtherNames() { + return otherNames; } - public void setOtherName(String otherName) { - this.otherName = otherName; + public void setOtherNames(List otherNames) { + if (this.otherNames == null) { + this.otherNames = Lists.newArrayList(); + } + this.otherNames = otherNames; } } diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SparkGenerateDoiAuthorList.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SparkGenerateDoiAuthorList.java index b4239bba2..011c153ec 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SparkGenerateDoiAuthorList.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SparkGenerateDoiAuthorList.java @@ -13,9 +13,6 @@ import java.util.stream.Stream; import org.apache.commons.io.IOUtils; import org.apache.hadoop.io.Text; -import org.apache.hadoop.io.compress.GzipCodec; -import org.apache.hadoop.mapreduce.Job; -import org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat; import org.apache.spark.SparkConf; import org.apache.spark.api.java.JavaPairRDD; import org.apache.spark.api.java.JavaRDD; @@ -33,7 +30,7 @@ import com.google.gson.JsonElement; import com.google.gson.JsonParser; import eu.dnetlib.dhp.application.ArgumentApplicationParser; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcid.model.WorkData; import scala.Tuple2; diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SummariesDecompressor.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SummariesDecompressor.java index 29d72ed0b..d1b2a1d73 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SummariesDecompressor.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/SummariesDecompressor.java @@ -19,7 +19,7 @@ import org.apache.hadoop.io.compress.CompressionCodec; import org.apache.hadoop.io.compress.CompressionCodecFactory; import org.mortbay.log.Log; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcid.xml.XMLRecordParser; import eu.dnetlib.doiboost.orcidnodoi.json.JsonWriter; @@ -56,6 +56,7 @@ public class SummariesDecompressor { int nameFound = 0; int surnameFound = 0; int creditNameFound = 0; + int otherNamesFound = 0; int errorFromOrcidFound = 0; int xmlParserErrorFound = 0; try (TarArchiveInputStream tais = new TarArchiveInputStream(gzipInputStream)) { @@ -117,6 +118,9 @@ public class SummariesDecompressor { if (authorData.getCreditName() != null) { creditNameFound += 1; } + if (authorData.getOtherNames() != null && authorData.getOtherNames().size() > 1) { + otherNamesFound += authorData.getOtherNames().size(); + } } else { Log.warn("Data not retrievable [" + entry.getName() + "] " + buffer.toString()); @@ -152,6 +156,7 @@ public class SummariesDecompressor { Log.info("Name found: " + nameFound); Log.info("Surname found: " + surnameFound); Log.info("Credit name found: " + creditNameFound); + Log.info("Other names found: " + otherNamesFound); Log.info("Error from Orcid found: " + errorFromOrcidFound); Log.info("Error parsing xml record found: " + xmlParserErrorFound); } diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParser.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParser.java index 2e43f4d3e..a807cf132 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParser.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParser.java @@ -14,7 +14,7 @@ import com.ximpleware.VTDNav; import eu.dnetlib.dhp.parser.utility.VtdException; import eu.dnetlib.dhp.parser.utility.VtdUtilityParser; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcid.model.WorkData; public class XMLRecordParser { @@ -81,6 +81,12 @@ public class XMLRecordParser { if (!creditNames.isEmpty()) { authorData.setCreditName(creditNames.get(0)); } + + final List otherNames = VtdUtilityParser.getTextValue(ap, vn, "//other-name:content"); + if (!otherNames.isEmpty()) { + authorData.setOtherNames(otherNames); + } + return authorData; } diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/SparkGenEnrichedOrcidWorks.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/SparkGenEnrichedOrcidWorks.java index 7f715fa7d..cc65b0b4f 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/SparkGenEnrichedOrcidWorks.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/SparkGenEnrichedOrcidWorks.java @@ -1,18 +1,12 @@ package eu.dnetlib.doiboost.orcidnodoi; -import com.fasterxml.jackson.databind.ObjectMapper; -import com.google.gson.Gson; -import com.google.gson.JsonElement; -import com.google.gson.JsonParser; -import eu.dnetlib.dhp.application.ArgumentApplicationParser; -import eu.dnetlib.dhp.schema.action.AtomicAction; -import eu.dnetlib.dhp.schema.oaf.Publication; -import eu.dnetlib.doiboost.orcid.json.JsonHelper; -import eu.dnetlib.doiboost.orcid.model.AuthorData; -import eu.dnetlib.doiboost.orcidnodoi.model.WorkDataNoDoi; -import eu.dnetlib.doiboost.orcidnodoi.oaf.PublicationToOaf; -import eu.dnetlib.doiboost.orcidnodoi.similarity.AuthorMatcher; +import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession; + +import java.io.IOException; +import java.util.Objects; +import java.util.Optional; + import org.apache.commons.io.IOUtils; import org.apache.hadoop.io.Text; import org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat; @@ -26,14 +20,22 @@ import org.apache.spark.sql.Encoders; import org.apache.spark.util.LongAccumulator; import org.slf4j.Logger; import org.slf4j.LoggerFactory; + +import com.fasterxml.jackson.databind.ObjectMapper; +import com.google.gson.Gson; +import com.google.gson.JsonElement; +import com.google.gson.JsonParser; + +import eu.dnetlib.dhp.application.ArgumentApplicationParser; +import eu.dnetlib.dhp.schema.action.AtomicAction; +import eu.dnetlib.dhp.schema.oaf.Publication; +import eu.dnetlib.dhp.schema.orcid.AuthorData; +import eu.dnetlib.doiboost.orcid.json.JsonHelper; +import eu.dnetlib.doiboost.orcidnodoi.model.WorkDataNoDoi; +import eu.dnetlib.doiboost.orcidnodoi.oaf.PublicationToOaf; +import eu.dnetlib.doiboost.orcidnodoi.similarity.AuthorMatcher; import scala.Tuple2; -import java.io.IOException; -import java.util.Objects; -import java.util.Optional; - -import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession; - /** * This spark job generates one parquet file, containing orcid publications dataset */ diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/json/JsonWriter.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/json/JsonWriter.java index 363cb13e6..982fb6316 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/json/JsonWriter.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/json/JsonWriter.java @@ -1,9 +1,12 @@ package eu.dnetlib.doiboost.orcidnodoi.json; +import com.fasterxml.jackson.annotation.JsonInclude; +import com.fasterxml.jackson.core.JsonProcessingException; +import com.fasterxml.jackson.databind.ObjectMapper; import com.google.gson.JsonObject; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcid.model.WorkData; /** @@ -12,15 +15,11 @@ import eu.dnetlib.doiboost.orcid.model.WorkData; public class JsonWriter { - public static String create(AuthorData authorData) { - JsonObject author = new JsonObject(); - author.addProperty("oid", authorData.getOid()); - author.addProperty("name", authorData.getName()); - author.addProperty("surname", authorData.getSurname()); - if (authorData.getCreditName() != null) { - author.addProperty("creditname", authorData.getCreditName()); - } - return author.toString(); + public static final com.fasterxml.jackson.databind.ObjectMapper OBJECT_MAPPER = new ObjectMapper() + .setSerializationInclusion(JsonInclude.Include.NON_NULL);; + + public static String create(AuthorData authorData) throws JsonProcessingException { + return OBJECT_MAPPER.writeValueAsString(authorData); } public static String create(WorkData workData) { diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/model/Contributor.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/model/Contributor.java index 9a8651c85..9222c1cc4 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/model/Contributor.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/model/Contributor.java @@ -3,7 +3,7 @@ package eu.dnetlib.doiboost.orcidnodoi.model; import java.io.Serializable; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; /** * This class models the data related to a contributor, that are retrieved from an orcid publication diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/similarity/AuthorMatcher.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/similarity/AuthorMatcher.java index 2f86820fb..c0f617868 100644 --- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/similarity/AuthorMatcher.java +++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcidnodoi/similarity/AuthorMatcher.java @@ -18,7 +18,7 @@ import com.ximpleware.XPathEvalException; import com.ximpleware.XPathParseException; import eu.dnetlib.dhp.parser.utility.VtdException; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcidnodoi.model.Contributor; import eu.dnetlib.doiboost.orcidnodoi.model.WorkDataNoDoi; @@ -47,7 +47,7 @@ public class AuthorMatcher { .forEach(c -> { if (simpleMatch(c.getCreditName(), author.getName()) || simpleMatch(c.getCreditName(), author.getSurname()) || - simpleMatch(c.getCreditName(), author.getOtherName())) { + simpleMatchOnOtherNames(c.getCreditName(), author.getOtherNames())) { matchCounters.set(0, matchCounters.get(0) + 1); c.setSimpleMatch(true); } @@ -91,6 +91,13 @@ public class AuthorMatcher { } + public static boolean simpleMatchOnOtherNames(String name, List otherNames) { + if (otherNames == null || (otherNames != null && otherNames.isEmpty())) { + return false; + } + return otherNames.stream().filter(o -> simpleMatch(name, o)).count() > 0; + } + public static boolean simpleMatch(String name, String searchValue) { if (searchValue == null) { return false; diff --git a/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_summaries/oozie_app/config-default.xml b/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_summaries/oozie_app/config-default.xml index e1829e847..191654378 100644 --- a/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_summaries/oozie_app/config-default.xml +++ b/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_summaries/oozie_app/config-default.xml @@ -21,6 +21,6 @@ oozie.launcher.mapreduce.map.java.opts - -Xmx16g + -Xmx8g \ No newline at end of file diff --git a/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParserTest.java b/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParserTest.java index 5bf6f27b9..b7be5e5cd 100644 --- a/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParserTest.java +++ b/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcid/xml/XMLRecordParserTest.java @@ -2,12 +2,14 @@ package eu.dnetlib.doiboost.orcid.xml; import static org.junit.jupiter.api.Assertions.assertNotNull; +import static org.junit.jupiter.api.Assertions.assertTrue; import org.apache.commons.io.IOUtils; import org.junit.jupiter.api.Test; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcid.model.WorkData; +import eu.dnetlib.doiboost.orcidnodoi.json.JsonWriter; public class XMLRecordParserTest { @@ -55,4 +57,21 @@ public class XMLRecordParserTest { assertNotNull(workData.getDoi()); System.out.println("doi: " + workData.getDoi()); } + + @Test + public void testOrcidOtherNamesXMLParser() throws Exception { + + String xml = IOUtils + .toString( + this.getClass().getResourceAsStream("summary_0000-0001-5109-1000_othername.xml")); + + XMLRecordParser p = new XMLRecordParser(); + + AuthorData authorData = XMLRecordParser.VTDParseAuthorData(xml.getBytes()); + assertNotNull(authorData); + assertNotNull(authorData.getOtherNames()); + assertTrue(authorData.getOtherNames().get(0).equals("Andrew C. Porteus")); + String jsonData = JsonWriter.create(authorData); + assertNotNull(jsonData); + } } diff --git a/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcidnodoi/xml/OrcidNoDoiTest.java b/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcidnodoi/xml/OrcidNoDoiTest.java index c2c4ed5e1..948e5b094 100644 --- a/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcidnodoi/xml/OrcidNoDoiTest.java +++ b/dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/orcidnodoi/xml/OrcidNoDoiTest.java @@ -5,34 +5,24 @@ import static org.junit.jupiter.api.Assertions.assertNotNull; import static org.junit.jupiter.api.Assertions.assertTrue; import java.io.IOException; -import java.text.Normalizer; import java.util.*; -import javax.validation.constraints.AssertTrue; - import org.apache.commons.io.IOUtils; import org.apache.commons.lang3.StringUtils; -import org.apache.commons.text.similarity.JaccardSimilarity; -import org.apache.commons.text.similarity.JaroWinklerSimilarity; import org.junit.jupiter.api.Test; -import org.mortbay.log.Log; import org.slf4j.Logger; import org.slf4j.LoggerFactory; -import com.google.gson.Gson; -import com.google.gson.GsonBuilder; import com.ximpleware.NavException; import com.ximpleware.ParseException; import com.ximpleware.XPathEvalException; import com.ximpleware.XPathParseException; import eu.dnetlib.dhp.parser.utility.VtdException; -import eu.dnetlib.dhp.schema.oaf.Author; -import eu.dnetlib.doiboost.orcid.model.AuthorData; +import eu.dnetlib.dhp.schema.orcid.AuthorData; import eu.dnetlib.doiboost.orcidnodoi.model.Contributor; import eu.dnetlib.doiboost.orcidnodoi.model.WorkDataNoDoi; import eu.dnetlib.doiboost.orcidnodoi.similarity.AuthorMatcher; -import jdk.nashorn.internal.ir.annotations.Ignore; public class OrcidNoDoiTest { diff --git a/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/record_8888-8888-8888-8880.xml b/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/record_8888-8888-8888-8880.xml new file mode 100644 index 000000000..7abc2f35a --- /dev/null +++ b/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/record_8888-8888-8888-8880.xml @@ -0,0 +1,770 @@ + + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + zh_CN + + + API + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + true + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + 2001-12-31T12:00:00 + true + true + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + give-names + family-name + credit-name + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + other-name-1 + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + biography + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + url-name-1 + http://url.com/ + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + user1@email.com + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + US + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + keyword1 + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + type-1 + value-1 + http://url.com/1 + self + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + distinction:department-name + distinction:role-title + + 1948 + 02 + 02 + + + 1948 + 02 + 02 + + + distinction-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-distinction + GRID + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + education:department-name + education:role-title + + 1948 + 02 + 02 + + + 2019 + 01 + 01 + + + education-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-education + GRID + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + employment:department-name + employment:role-title + + 1948 + 02 + 02 + + + 2025 + + + employment-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-employment + GRID + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + grant_number + external-id-value-1 + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + + common:title + common:translated-title + + + + grant_number + external-id-value-1 + http://tempuri.org + self + + + grant + + 1948 + 02 + 02 + + + 1948 + 02 + 02 + + + common:name + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-funding + FUNDREF + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + invited-position:department-name + invited-position:role-title + + 2019 + 01 + 01 + + + 2025 + 01 + 01 + + + invited-position-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-invited-position + GRID + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + membership:department-name + membership:role-title + + 1948 + 02 + 02 + + + membership-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-membership + RINGGOLD + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + something + external-id-value + http://orcid.org + self + + + + 2001-12-31T12:00:00 + + + something + external-id-value + + http://orcid.org + self + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/client/APP-9999999999999901 + APP-9999999999999901 + orcid.org + + + + reviewer + + + something + external-id-value + http://orcid.org + self + + + http://orcid.org + review + + 1948 + 02 + 02 + + orcid-generated:12345 + + common:name + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-peer-review + RINGGOLD + + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + qualification:department-name + qualification:role-title + + 1948 + 02 + 02 + + + 2025 + 12 + + + qualification-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-qualification + RINGGOLD + + + + + + + 2017-01-18T15:06:05.147-06:00 + + 2017-01-18T15:03:56.856-06:00 + + + proposal_id + 123456 + self + + + + + 2015-06-25T16:01:12.718Z + 2017-09-08T13:31:19.987Z + + + + https://orcid.org/0000-0000-0000-0000 + 0000-0000-0000-0000 + orcid.org + + XSEDE ORCID integration + + + + + Giant Laser Award + + + + XSEDE + + city + region + US + + + XX + grid + + + + + + proposal_id + 123456 + self + + + + 1999 + 02 + 02 + + + 2012 + 02 + 02 + + http://xsede.org/GiantLaserAward + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + + agr + external-id-value + + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + service:department-name + service:role-title + + 1948 + 02 + 02 + + + service-org + + common:city + common:region + AF + + + common:disambiguated-organization-identifier-service + RINGGOLD + + + + + + + 2001-12-31T12:00:00 + + 2001-12-31T12:00:00 + + + agr + external-id-value + http://orcid.org + part-of + + + + 2001-12-31T12:00:00 + 2001-12-31T12:00:00 + + + https://orcid.org/client/8888-8888-8888-8880 + 8888-8888-8888-8880 + orcid.org + + + + + common:title + + common:translated-title + + + + agr + external-id-value + http://tempuri.org + self + + + artistic-performance + + 1948 + 02 + 02 + + Procedia Computer Science + + + + + \ No newline at end of file diff --git a/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/summary_0000-0001-5109-1000_othername.xml b/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/summary_0000-0001-5109-1000_othername.xml new file mode 100644 index 000000000..43bc96b8c --- /dev/null +++ b/dhp-workflows/dhp-doiboost/src/test/resources/eu/dnetlib/doiboost/orcid/xml/summary_0000-0001-5109-1000_othername.xml @@ -0,0 +1,196 @@ + + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + + en + + + Member-referred + 2019-05-01T13:04:57.507Z + 2019-05-01T13:59:54.268Z + true + true + true + + + 2019-05-01T13:45:47.727Z + + 2019-05-01T13:04:57.507Z + 2019-05-01T13:04:57.740Z + Andrew + Porteus + + + 2019-05-01T13:44:57.072Z + + 2019-05-01T13:44:57.072Z + 2019-05-01T13:44:57.072Z + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + Andrew Porteus + + Andrew C. Porteus + + + + 2019-05-01T13:59:54.263Z + 2019-05-01T13:59:54.263Z + Retired Librarian + + + 2019-05-01T13:45:47.727Z + + 2019-05-01T13:45:47.727Z + 2019-05-01T13:45:47.727Z + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + Andrew Porteus + + Niagara Falls Poetry Project + http://niagarapoetry.ca + + + + + 2019-05-01T13:45:09.764Z + + 2019-05-01T13:45:09.764Z + 2019-05-01T13:45:09.764Z + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + Andrew Porteus + + CA + + + + + + + 2019-05-01T13:57:45.787Z + + + 2019-05-01T13:15:26.102Z + + 2019-05-01T13:15:26.102Z + + + 2019-05-01T13:15:26.102Z + 2019-05-01T13:15:26.102Z + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + Andrew Porteus + + Library Technician Diploma + + 1976 + 09 + + + 1978 + 05 + + + Niagara College + + Welland + ON + CA + + + 125147 + RINGGOLD + + + + + + + + + + + + 2019-05-01T13:19:49.021Z + + 2019-05-01T13:19:49.021Z + + + 2019-05-01T13:19:49.021Z + 2019-05-01T13:19:49.021Z + + + https://orcid.org/0000-0001-5109-1000 + 0000-0001-5109-1000 + orcid.org + + Andrew Porteus + + Communication, Film & Popular Culture + Master's Candidate + + 2018 + 09 + + + Brock University + + Saint Catharines + ON + CA + + + 7497 + RINGGOLD + + + + + + + \ No newline at end of file