D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 10 Projects 2 Releases 2 Wiki Activity

master #59

Closed
claudio.atzori wants to merge 3221 commits from master into stable_ids
pull from: master
merge into: D-Net:stable_ids
Conversation 0 Commits 3221 Files Changed 2141 +43 -16
2 changed files with 43 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit c763aded70 - Show all commits

57
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java
View File

@ -206,11 +206,16 @@ public class SparkDedupTest implements Serializable {
.load(DedupUtility.createSimRelPath(testOutputBasePath, testActionSetId, "otherresearchproduct")) .load(DedupUtility.createSimRelPath(testOutputBasePath, testActionSetId, "otherresearchproduct"))
.count(); .count();
assertEquals(3082, orgs_simrel); assertEquals(3076, orgs_simrel);
assertEquals(7036, pubs_simrel); assertEquals(7040, pubs_simrel);
assertEquals(336, sw_simrel); assertEquals(336, sw_simrel);
assertEquals(442, ds_simrel); assertEquals(442, ds_simrel);
assertEquals(6750, orp_simrel); assertEquals(6784, orp_simrel);
// System.out.println("org_simrel = " + orgs_simrel);
// System.out.println("pub_simrel = " + pubs_simrel);
// System.out.println("sw_simrel = " + sw_simrel);
// System.out.println("ds_simrel = " + ds_simrel);
// System.out.println("orp_simrel = " + orp_simrel);
} }
@Test @Test
@ -258,10 +263,14 @@ public class SparkDedupTest implements Serializable {
.count(); .count();
// entities simrels supposed to be equal to the number of previous step (no rels in whitelist) // entities simrels supposed to be equal to the number of previous step (no rels in whitelist)
assertEquals(3082, orgs_simrel); assertEquals(3076, orgs_simrel);
assertEquals(7036, pubs_simrel); assertEquals(7040, pubs_simrel);
assertEquals(442, ds_simrel); assertEquals(442, ds_simrel);
assertEquals(6750, orp_simrel); assertEquals(6784, orp_simrel);
// System.out.println("org_simrel = " + orgs_simrel);
// System.out.println("pub_simrel = " + pubs_simrel);
// System.out.println("ds_simrel = " + ds_simrel);
// System.out.println("orp_simrel = " + orp_simrel);
// entities simrels to be different from the number of previous step (new simrels in the whitelist) // entities simrels to be different from the number of previous step (new simrels in the whitelist)
Dataset<Row> sw_simrel = spark Dataset<Row> sw_simrel = spark
@ -288,6 +297,7 @@ public class SparkDedupTest implements Serializable {
.count() > 0); .count() > 0);
assertEquals(338, sw_simrel.count()); assertEquals(338, sw_simrel.count());
// System.out.println("sw_simrel_count = " + sw_simrel.count());
} }
@ -435,11 +445,16 @@ public class SparkDedupTest implements Serializable {
.load(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel") .load(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel")
.count(); .count();
assertEquals(1272, orgs_mergerel); assertEquals(1268, orgs_mergerel);
assertEquals(1438, pubs_mergerel); assertEquals(1444, pubs_mergerel);
assertEquals(286, sw_mergerel); assertEquals(286, sw_mergerel);
assertEquals(472, ds_mergerel); assertEquals(472, ds_mergerel);
assertEquals(718, orp_mergerel); assertEquals(738, orp_mergerel);
// System.out.println("org_simrel = " + orgs_mergerel);
// System.out.println("pub_simrel = " + pubs_mergerel);
// System.out.println("sw_simrel = " + sw_mergerel);
// System.out.println("ds_simrel = " + ds_mergerel);
// System.out.println("orp_simrel = " + orp_mergerel);
} }
@ -483,11 +498,16 @@ public class SparkDedupTest implements Serializable {
testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_deduprecord") testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_deduprecord")
.count(); .count();
assertEquals(85, orgs_deduprecord); assertEquals(86, orgs_deduprecord);
assertEquals(65, pubs_deduprecord); assertEquals(67, pubs_deduprecord);
assertEquals(49, sw_deduprecord); assertEquals(49, sw_deduprecord);
assertEquals(97, ds_deduprecord); assertEquals(97, ds_deduprecord);
assertEquals(89, orp_deduprecord); assertEquals(92, orp_deduprecord);
// System.out.println("org_simrel = " + orgs_deduprecord);
// System.out.println("pub_simrel = " + pubs_deduprecord);
// System.out.println("sw_simrel = " + sw_deduprecord);
// System.out.println("ds_simrel = " + ds_deduprecord);
// System.out.println("orp_simrel = " + orp_deduprecord);
} }
@Test @Test
@ -566,13 +586,18 @@ public class SparkDedupTest implements Serializable {
.distinct() .distinct()
.count(); .count();
assertEquals(896, publications); assertEquals(898, publications);
assertEquals(838, organizations); assertEquals(839, organizations);
assertEquals(100, projects); assertEquals(100, projects);
assertEquals(100, datasource); assertEquals(100, datasource);
assertEquals(198, softwares); assertEquals(198, softwares);
assertEquals(389, dataset); assertEquals(389, dataset);
assertEquals(517, otherresearchproduct); assertEquals(520, otherresearchproduct);
// System.out.println("publications = " + publications);
// System.out.println("organizations = " + organizations);
// System.out.println("softwares = " + softwares);
// System.out.println("dataset = " + dataset);
// System.out.println("orp = " + otherresearchproduct);
long deletedOrgs = jsc long deletedOrgs = jsc
.textFile(testDedupGraphBasePath + "/organization") .textFile(testDedupGraphBasePath + "/organization")
@ -627,6 +652,8 @@ public class SparkDedupTest implements Serializable {
long relations = jsc.textFile(testDedupGraphBasePath + "/relation").count(); long relations = jsc.textFile(testDedupGraphBasePath + "/relation").count();
assertEquals(4860, relations); assertEquals(4860, relations);
// System.out.println("relations = " + relations);
// check deletedbyinference // check deletedbyinference
final Dataset<Relation> mergeRels = spark final Dataset<Relation> mergeRels = spark

2
pom.xml
View File

@ -801,7 +801,7 @@
<dnet-actionmanager-api.version>[4.0.3]</dnet-actionmanager-api.version> <dnet-actionmanager-api.version>[4.0.3]</dnet-actionmanager-api.version>
<dnet-actionmanager-common.version>[6.0.5]</dnet-actionmanager-common.version> <dnet-actionmanager-common.version>[6.0.5]</dnet-actionmanager-common.version>
<dnet-openaire-broker-common.version>[3.1.6]</dnet-openaire-broker-common.version> <dnet-openaire-broker-common.version>[3.1.6]</dnet-openaire-broker-common.version>
<dnet-pace-core.version>[4.1.7]</dnet-pace-core.version> <dnet-pace-core.version>[4.1.12]</dnet-pace-core.version>
<cnr-rmi-api.version>[2.6.1]</cnr-rmi-api.version> <cnr-rmi-api.version>[2.6.1]</cnr-rmi-api.version>
<solr.version>7.5.0</solr.version> <solr.version>7.5.0</solr.version>
<okhttp.version>4.7.2</okhttp.version> <okhttp.version>4.7.2</okhttp.version>