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

updated dedup tests following the dnet-pace-core library update

This commit is contained in:
Claudio Atzori 2020-10-02 14:30:53 +02:00
parent 4fddd18403
commit 23f64d9eb4
2 changed files with 17 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java
View File

@ -194,10 +194,10 @@ public class SparkDedupTest implements Serializable {
.textFile(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_simrel") .textFile(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_simrel")
.count(); .count();
assertEquals(3432, orgs_simrel); assertEquals(3082, orgs_simrel);
assertEquals(7152, pubs_simrel); assertEquals(7036, pubs_simrel);
assertEquals(344, sw_simrel); assertEquals(344, sw_simrel);
assertEquals(458, ds_simrel); assertEquals(442, ds_simrel);
assertEquals(6750, orp_simrel); assertEquals(6750, orp_simrel);
} }
@ -343,8 +343,8 @@ public class SparkDedupTest implements Serializable {
.load(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel") .load(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_mergerel")
.count(); .count();
assertEquals(1276, orgs_mergerel); assertEquals(1272, orgs_mergerel);
assertEquals(1442, pubs_mergerel); assertEquals(1438, pubs_mergerel);
assertEquals(288, sw_mergerel); assertEquals(288, sw_mergerel);
assertEquals(472, ds_mergerel); assertEquals(472, ds_mergerel);
assertEquals(718, orp_mergerel); assertEquals(718, orp_mergerel);
@ -390,10 +390,10 @@ public class SparkDedupTest implements Serializable {
testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_deduprecord") testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_deduprecord")
.count(); .count();
assertEquals(82, orgs_deduprecord); assertEquals(85, orgs_deduprecord);
assertEquals(66, pubs_deduprecord); assertEquals(65, pubs_deduprecord);
assertEquals(51, sw_deduprecord); assertEquals(51, sw_deduprecord);
assertEquals(96, ds_deduprecord); assertEquals(97, ds_deduprecord);
assertEquals(89, orp_deduprecord); assertEquals(89, orp_deduprecord);
} }
@ -473,12 +473,12 @@ public class SparkDedupTest implements Serializable {
.distinct() .distinct()
.count(); .count();
assertEquals(897, publications); assertEquals(896, publications);
assertEquals(835, organizations); assertEquals(838, organizations);
assertEquals(100, projects); assertEquals(100, projects);
assertEquals(100, datasource); assertEquals(100, datasource);
assertEquals(200, softwares); assertEquals(200, softwares);
assertEquals(388, dataset); assertEquals(389, dataset);
assertEquals(517, otherresearchproduct); assertEquals(517, otherresearchproduct);
long deletedOrgs = jsc long deletedOrgs = jsc
@ -533,7 +533,7 @@ public class SparkDedupTest implements Serializable {
long relations = jsc.textFile(testDedupGraphBasePath + "/relation").count(); long relations = jsc.textFile(testDedupGraphBasePath + "/relation").count();
assertEquals(4866, relations); assertEquals(4858, relations);
// check deletedbyinference // check deletedbyinference
final Dataset<Relation> mergeRels = spark final Dataset<Relation> mergeRels = spark

10
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkStatsTest.java
View File

@ -168,10 +168,10 @@ public class SparkStatsTest implements Serializable {
.textFile(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_blockstats") .textFile(testOutputBasePath + "/" + testActionSetId + "/otherresearchproduct_blockstats")
.count(); .count();
assertEquals(121, orgs_blocks); assertEquals(549, orgs_blocks);
assertEquals(110, pubs_blocks); assertEquals(299, pubs_blocks);
assertEquals(21, sw_blocks); assertEquals(122, sw_blocks);
assertEquals(67, ds_blocks); assertEquals(186, ds_blocks);
assertEquals(55, orp_blocks); assertEquals(170, orp_blocks);
} }
} }