przemyslaw.jacewicz
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

refactoring

This commit is contained in:
Miriam Baglioni 2023-10-09 14:52:17 +02:00
parent 8448b9ebfb
commit a3d01ccb24
21 changed files with 268 additions and 239 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/opencitations/CreateActionSetSparkJob.java
View File

@ -7,7 +7,6 @@ import java.io.IOException;
import java.io.Serializable;
import java.util.*;
import eu.dnetlib.dhp.schema.oaf.utils.*;
import org.apache.commons.cli.ParseException;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.io.Text;
@ -30,6 +29,7 @@ import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.*;
import eu.dnetlib.dhp.utils.DHPUtils;
import scala.Tuple2;

16
dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/crossref/Crossref2Oaf.scala
View File

@ -31,9 +31,7 @@ case class mappingAuthor(
affiliation: Option[mappingAffiliation]
) {}
case class funderInfo(id:String,uri:String, name:String,synonym:List[String] ) {}
case class funderInfo(id: String, uri: String, name: String, synonym: List[String]) {}
case class mappingFunder(name: String, DOI: Option[String], award: Option[List[String]]) {}
@ -41,7 +39,9 @@ case object Crossref2Oaf {
val logger: Logger = LoggerFactory.getLogger(Crossref2Oaf.getClass)
val irishFunder: List[funderInfo] = {
val s = Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/doiboost/crossref/irish_funder.json")).mkString
val s = Source
.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/doiboost/crossref/irish_funder.json"))
.mkString
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
lazy val json: org.json4s.JValue = parse(s)
json.extract[List[funderInfo]]
@ -100,9 +100,11 @@ case object Crossref2Oaf {
"report" -> "0017 Report"
)
def getIrishId(doi:String):Option[String] = {
val id =doi.split("/").last
irishFunder.find(f => id.equalsIgnoreCase(f.id) || (f.synonym.nonEmpty && f.synonym.exists(s => s.equalsIgnoreCase(id)))).map(f => f.id)
def getIrishId(doi: String): Option[String] = {
val id = doi.split("/").last
irishFunder
.find(f => id.equalsIgnoreCase(f.id) || (f.synonym.nonEmpty && f.synonym.exists(s => s.equalsIgnoreCase(id))))
.map(f => f.id)
}
def mappingResult(result: Result, json: JValue, cobjCategory: String): Result = {

9
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/QueryCommunityAPI.java
View File

@ -21,7 +21,7 @@ import com.google.gson.Gson;
public class QueryCommunityAPI {
private static final String baseUrl = "https://services.openaire.eu/openaire/";
private static String get(String geturl) throws IOException{
private static String get(String geturl) throws IOException {
URL url = new URL(geturl);
HttpURLConnection conn = (HttpURLConnection) url.openConnection();
conn.setDoOutput(true);
@ -44,7 +44,7 @@ public class QueryCommunityAPI {
return get(baseUrl + "community/" + id);
}
public static String communityDatasource(String id)throws IOException{
public static String communityDatasource(String id) throws IOException {
return get(baseUrl + "community/" + id + "/contentproviders");
}
@ -53,8 +53,8 @@ public class QueryCommunityAPI {
return get(baseUrl + "community/" + id + "/propagationOrganizations");
}
public static String communityProjects(String id, String page, String size) throws IOException{
return get(baseUrl + "community/" + id +"/projects/" + page + "/" + size);
public static String communityProjects(String id, String page, String size) throws IOException {
return get(baseUrl + "community/" + id + "/projects/" + page + "/" + size);
}
@NotNull
@ -74,5 +74,4 @@ public class QueryCommunityAPI {
return body;
}
}

80
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/Utils.java
View File

@ -1,16 +1,6 @@
package eu.dnetlib.dhp.api;
import com.amazonaws.util.StringUtils;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.common.collect.Maps;
import eu.dnetlib.dhp.api.model.*;
import eu.dnetlib.dhp.bulktag.community.Community;
import eu.dnetlib.dhp.bulktag.community.CommunityConfiguration;
import eu.dnetlib.dhp.bulktag.community.Provider;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolver;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolverFactory;
import javax.management.Query;
import java.io.IOException;
import java.io.Serializable;
import java.util.ArrayList;
@ -19,6 +9,19 @@ import java.util.Map;
import java.util.Objects;
import java.util.stream.Collectors;
import javax.management.Query;
import com.amazonaws.util.StringUtils;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.common.collect.Maps;
import eu.dnetlib.dhp.api.model.*;
import eu.dnetlib.dhp.bulktag.community.Community;
import eu.dnetlib.dhp.bulktag.community.CommunityConfiguration;
import eu.dnetlib.dhp.bulktag.community.Provider;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolver;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolverFactory;
/**
* @author miriam.baglioni
* @Date 09/10/23
@ -33,22 +36,24 @@ public class Utils implements Serializable {
getValidCommunities()
.forEach(community -> {
try {
CommunityModel cm = MAPPER.readValue(QueryCommunityAPI.community(community.getId()), CommunityModel.class);
CommunityModel cm = MAPPER
.readValue(QueryCommunityAPI.community(community.getId()), CommunityModel.class);
validCommunities.add(getCommunity(cm));
} catch (IOException e) {
throw new RuntimeException(e);
}
});
validCommunities.forEach(community ->{
validCommunities.forEach(community -> {
try {
DatasourceList dl = MAPPER.readValue(QueryCommunityAPI.communityDatasource(community.getId()), DatasourceList.class);
DatasourceList dl = MAPPER
.readValue(QueryCommunityAPI.communityDatasource(community.getId()), DatasourceList.class);
community.setProviders(dl.stream().map(d -> {
// if(d.getEnabled() == null || Boolean.FALSE.equals(d.getEnabled()))
// return null;
Provider p = new Provider();
p.setOpenaireId("10|" + d.getOpenaireId());
p.setSelectionConstraints(d.getSelectioncriteria());
if(p.getSelectionConstraints() != null)
if (p.getSelectionConstraints() != null)
p.getSelectionConstraints().setSelection(resolver);
return p;
})
@ -59,46 +64,50 @@ public class Utils implements Serializable {
}
});
validCommunities.forEach(community ->{
if(community.isValid())
validCommunities.forEach(community -> {
if (community.isValid())
communities.put(community.getId(), community);
});
return new CommunityConfiguration(communities);
}
private static Community getCommunity(CommunityModel cm){
private static Community getCommunity(CommunityModel cm) {
Community c = new Community();
c.setId(cm.getId());
c.setZenodoCommunities(cm.getOtherZenodoCommunities());
if(!StringUtils.isNullOrEmpty(cm.getZenodoCommunity()))
if (!StringUtils.isNullOrEmpty(cm.getZenodoCommunity()))
c.getZenodoCommunities().add(cm.getZenodoCommunity());
c.setSubjects(cm.getSubjects());
c.getSubjects().addAll(cm.getFos());
c.getSubjects().addAll(cm.getSdg());
c.setConstraints(cm.getAdvancedConstraints());
if(c.getConstraints()!=null)
if (c.getConstraints() != null)
c.getConstraints().setSelection(resolver);
c.setRemoveConstraints(cm.getRemoveConstraints());
if(c.getRemoveConstraints()!=null)
if (c.getRemoveConstraints() != null)
c.getRemoveConstraints().setSelection(resolver);
return c;
}
public static List<CommunityModel> getValidCommunities() throws IOException {
return MAPPER.readValue(QueryCommunityAPI.communities(), CommunitySummary.class)
return MAPPER
.readValue(QueryCommunityAPI.communities(), CommunitySummary.class)
.stream()
.filter(community -> !community.getStatus().equals("hidden") &&
.filter(
community -> !community.getStatus().equals("hidden") &&
(community.getType().equals("ri") || community.getType().equals("community")))
.collect(Collectors.toList());
}
public static CommunityEntityMap getCommunityOrganization() throws IOException {
CommunityEntityMap organizationMap = new CommunityEntityMap();
getValidCommunities()
.forEach(community -> {
String id = community.getId();
try {
List<String> associatedOrgs = MAPPER.readValue(QueryCommunityAPI.communityPropagationOrganization(id), OrganizationList.class);
if(associatedOrgs.size() >0){
List<String> associatedOrgs = MAPPER
.readValue(QueryCommunityAPI.communityPropagationOrganization(id), OrganizationList.class);
if (associatedOrgs.size() > 0) {
organizationMap.put(id, associatedOrgs);
}
} catch (IOException e) {
@ -108,28 +117,31 @@ public class Utils implements Serializable {
return organizationMap;
}
public static CommunityEntityMap getCommunityProjects()throws IOException{
public static CommunityEntityMap getCommunityProjects() throws IOException {
CommunityEntityMap projectMap = new CommunityEntityMap();
getValidCommunities()
.forEach(community ->{
.forEach(community -> {
int page = -1;
int size = 100;
ContentModel cm = new ContentModel();
List<String> projectList = new ArrayList<>();
do{
page ++;
do {
page++;
try {
cm = MAPPER.readValue( QueryCommunityAPI.communityProjects(community.getId(), String.valueOf(page), String.valueOf(size)), ContentModel.class);
if (cm.getContent().size() > 0){
cm = MAPPER
.readValue(
QueryCommunityAPI
.communityProjects(community.getId(), String.valueOf(page), String.valueOf(size)),
ContentModel.class);
if (cm.getContent().size() > 0) {
cm.getContent().forEach(p ->
projectList.add ("40|" + p.getOpenaireId()));
cm.getContent().forEach(p -> projectList.add("40|" + p.getOpenaireId()));
projectMap.put(community.getId(), projectList);
}
} catch (IOException e) {
throw new RuntimeException(e);
}
}while (!cm.getLast());
} while (!cm.getLast());
});
return projectMap;
}

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/CommunityContentprovider.java
View File

@ -1,10 +1,11 @@
package eu.dnetlib.dhp.api.model;
import com.fasterxml.jackson.annotation.JsonAutoDetect;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import com.google.gson.Gson;
import eu.dnetlib.dhp.bulktag.community.SelectionConstraints;
import eu.dnetlib.dhp.bulktag.community.SelectionConstraints;
@JsonAutoDetect
@JsonIgnoreProperties(ignoreUnknown = true)
@ -30,7 +31,6 @@ public class CommunityContentprovider {
this.openaireId = openaireId;
}
public SelectionConstraints getSelectioncriteria() {
return this.selectioncriteria;

2
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/CommunityModel.java
View File

@ -5,8 +5,8 @@ import java.io.Serializable;
import java.util.List;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import eu.dnetlib.dhp.bulktag.community.SelectionConstraints;
import eu.dnetlib.dhp.bulktag.community.SelectionConstraints;
/**
* @author miriam.baglioni

5
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/CommunitySummary.java
View File

@ -1,3 +1,4 @@
package eu.dnetlib.dhp.api.model;
import java.io.Serializable;
@ -11,6 +12,4 @@ public class CommunitySummary extends ArrayList<CommunityModel> implements Seria
public CommunitySummary() {
super();
}
}
}

5
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/ContentModel.java
View File

@ -1,10 +1,11 @@
package eu.dnetlib.dhp.api.model;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
package eu.dnetlib.dhp.api.model;
import java.io.Serializable;
import java.util.List;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
/**
* @author miriam.baglioni
* @Date 09/10/23

9
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/DatasourceList.java
View File

@ -1,12 +1,13 @@
package eu.dnetlib.dhp.api.model;
import eu.dnetlib.dhp.api.model.CommunityContentprovider;
import java.io.Serializable;
import java.util.ArrayList;
import eu.dnetlib.dhp.api.model.CommunityContentprovider;
public class DatasourceList extends ArrayList<CommunityContentprovider> implements Serializable {
public DatasourceList(){
public DatasourceList() {
super();
}
}

3
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/OrganizationList.java
View File

@ -1,3 +1,4 @@
package eu.dnetlib.dhp.api.model;
import java.io.Serializable;
@ -9,7 +10,7 @@ import java.util.ArrayList;
*/
public class OrganizationList extends ArrayList<String> implements Serializable {
public OrganizationList(){
public OrganizationList() {
super();
}
}

5
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/api/model/ProjectModel.java
View File

@ -1,9 +1,10 @@
package eu.dnetlib.dhp.api.model;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
import java.io.Serializable;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
/**
* @author miriam.baglioni
* @Date 09/10/23

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/SparkBulkTagJob.java
View File

@ -6,7 +6,6 @@ import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.*;
import eu.dnetlib.dhp.api.Utils;
import org.apache.commons.io.IOUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.FilterFunction;
@ -21,6 +20,7 @@ import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.gson.Gson;
import eu.dnetlib.dhp.api.Utils;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.bulktag.community.*;
import eu.dnetlib.dhp.schema.oaf.Datasource;
@ -87,7 +87,7 @@ public class SparkBulkTagJob {
if (isTest) {
cc = CommunityConfigurationFactory.newInstance(taggingConf);
} else {
cc = Utils.getCommunityConfiguration();//QueryInformationSystem.getCommunityConfiguration(parser.get("isLookUpUrl"));
cc = Utils.getCommunityConfiguration();// QueryInformationSystem.getCommunityConfiguration(parser.get("isLookUpUrl"));
}
runWithSparkSession(

1
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/Community.java
View File

@ -8,7 +8,6 @@ import java.util.Optional;
import com.google.gson.Gson;
/** Created by miriam on 01/08/2018. */
public class Community implements Serializable {

1
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/CommunityConfigurationFactory.java
View File

@ -5,6 +5,7 @@ import java.io.StringReader;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;

9
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/Constraint.java
View File

@ -4,9 +4,10 @@ package eu.dnetlib.dhp.bulktag.community;
import java.io.Serializable;
import java.lang.reflect.InvocationTargetException;
import org.apache.htrace.fasterxml.jackson.annotation.JsonIgnore;
import eu.dnetlib.dhp.bulktag.criteria.Selection;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolver;
import org.apache.htrace.fasterxml.jackson.annotation.JsonIgnore;
public class Constraint implements Serializable {
private String verb;
@ -39,11 +40,12 @@ public class Constraint implements Serializable {
public void setValue(String value) {
this.value = value;
}
//@JsonIgnore
//public void setSelection(Selection sel) {
// public void setSelection(Selection sel) {
// selection = sel;
// }
@JsonIgnore
@JsonIgnore
public void setSelection(VerbResolver resolver)
throws InvocationTargetException, NoSuchMethodException, InstantiationException,
IllegalAccessException {
@ -54,5 +56,4 @@ public class Constraint implements Serializable {
return selection.apply(metadata);
}
}

1
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/SelectionConstraints.java
View File

@ -12,6 +12,7 @@ import com.google.gson.Gson;
import com.google.gson.reflect.TypeToken;
import eu.dnetlib.dhp.bulktag.criteria.VerbResolver;
@JsonAutoDetect
public class SelectionConstraints implements Serializable {
private List<Constraints> criteria;

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttocommunityfromorganization/PrepareResultCommunitySet.java
View File

@ -6,8 +6,6 @@ import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkHiveSession;
import java.util.*;
import eu.dnetlib.dhp.api.Utils;
import eu.dnetlib.dhp.api.model.CommunityEntityMap;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.io.compress.GzipCodec;
import org.apache.spark.SparkConf;
@ -20,6 +18,8 @@ import org.slf4j.LoggerFactory;
import com.google.gson.Gson;
import eu.dnetlib.dhp.api.Utils;
import eu.dnetlib.dhp.api.model.CommunityEntityMap;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;

2
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/bulktag/CommunityConfigurationFactoryTest.java
View File

@ -47,7 +47,7 @@ class CommunityConfigurationFactoryTest {
sc.setVerb("not_contains");
sc.setField("contributor");
sc.setValue("DARIAH");
sc.setSelection(resolver);//.getSelectionCriteria(sc.getVerb(), sc.getValue()));
sc.setSelection(resolver);// .getSelectionCriteria(sc.getVerb(), sc.getValue()));
String metadata = "This work has been partially supported by DARIAH-EU infrastructure";
Assertions.assertFalse(sc.verifyCriteria(metadata));
}

49
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/bulktag/QueryCommunityAPITest.java
View File

@ -1,14 +1,8 @@
package eu.dnetlib.dhp.bulktag;
import eu.dnetlib.dhp.api.Utils;
import eu.dnetlib.dhp.api.model.CommunityEntityMap;
import eu.dnetlib.dhp.bulktag.community.Community;
import eu.dnetlib.dhp.bulktag.community.CommunityConfiguration;
import java.util.List;
import eu.dnetlib.dhp.api.model.CommunityModel;
import eu.dnetlib.dhp.api.model.CommunitySummary;
import eu.dnetlib.dhp.api.model.DatasourceList;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.Test;
@ -16,8 +10,13 @@ import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.api.QueryCommunityAPI;
import java.util.List;
import eu.dnetlib.dhp.api.Utils;
import eu.dnetlib.dhp.api.model.CommunityEntityMap;
import eu.dnetlib.dhp.api.model.CommunityModel;
import eu.dnetlib.dhp.api.model.CommunitySummary;
import eu.dnetlib.dhp.api.model.DatasourceList;
import eu.dnetlib.dhp.bulktag.community.Community;
import eu.dnetlib.dhp.bulktag.community.CommunityConfiguration;
/**
* @author miriam.baglioni
@ -43,9 +42,12 @@ public class QueryCommunityAPITest {
void community() throws Exception {
String id = "dh-ch";
String body = QueryCommunityAPI.community(id);
System.out.println(new ObjectMapper().writeValueAsString(new ObjectMapper()
.readValue(body, CommunityModel.class)))
;
System.out
.println(
new ObjectMapper()
.writeValueAsString(
new ObjectMapper()
.readValue(body, CommunityModel.class)));
}
@Test
@ -54,7 +56,7 @@ public class QueryCommunityAPITest {
String body = QueryCommunityAPI.communityDatasource(id);
new ObjectMapper()
.readValue(body, DatasourceList.class)
.forEach(ds-> {
.forEach(ds -> {
try {
System.out.println(new ObjectMapper().writeValueAsString(ds));
} catch (JsonProcessingException e) {
@ -68,15 +70,20 @@ public class QueryCommunityAPITest {
void validCommunities() throws Exception {
CommunityConfiguration cc = Utils.getCommunityConfiguration();
System.out.println(cc.getCommunities().keySet());
Community community =cc.getCommunities().get("aurora");
Community community = cc.getCommunities().get("aurora");
Assertions.assertEquals(0, community.getSubjects().size());
Assertions.assertEquals(null, community.getConstraints());
Assertions.assertEquals(null, community.getRemoveConstraints());
Assertions.assertEquals(2, community.getZenodoCommunities().size());
Assertions.assertTrue(community.getZenodoCommunities().stream().anyMatch(c -> c.equals("aurora-universities-network")));
Assertions.assertTrue(community.getZenodoCommunities().stream().anyMatch(c -> c.equals("university-of-innsbruck")));
Assertions
.assertTrue(
community.getZenodoCommunities().stream().anyMatch(c -> c.equals("aurora-universities-network")));
Assertions
.assertTrue(community.getZenodoCommunities().stream().anyMatch(c -> c.equals("university-of-innsbruck")));
Assertions.assertEquals(35, community.getProviders().size());
Assertions.assertEquals(35, community.getProviders().stream().filter(p->p.getSelectionConstraints()==null).count());
Assertions
.assertEquals(
35, community.getProviders().stream().filter(p -> p.getSelectionConstraints() == null).count());
}
@Test
@ -84,8 +91,12 @@ public class QueryCommunityAPITest {
CommunityEntityMap projectMap = Utils.getCommunityProjects();
Assertions.assertFalse(projectMap.containsKey("mes"));
Assertions.assertEquals(33, projectMap.size());
Assertions.assertTrue(projectMap.keySet().stream().allMatch(k -> projectMap.get(k).stream().allMatch(p -> p.startsWith("40|"))));
Assertions
.assertTrue(
projectMap
.keySet()
.stream()
.allMatch(k -> projectMap.get(k).stream().allMatch(p -> p.startsWith("40|"))));
}
}