sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

merged conflicts with upstream branch

This commit is contained in:
Miriam Baglioni 2020-05-12 10:03:44 +02:00
parent 8ffc050b8a 527e8169a8
commit 7687519f00
2 changed files with 91 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

91
dhp-workflows/dhp-blacklist/src/main/resources/eu/dnetlib/dhp/blacklist/oozie_app/workflow.xml
View File

@ -33,10 +33,99 @@
<delete path="${outputPath}"/>
<mkdir path="${outputPath}"/>
</fs>
<ok to="read_blacklist"/>
<ok to="copy_entities"/>
<error to="Kill"/>
</action>
<fork name="copy_entities">
<path start="copy_publication"/>
<path start="copy_dataset"/>
<path start="copy_orp"/>
<path start="copy_software"/>
<path start="copy_datasource"/>
<path start="copy_project"/>
<path start="copy_organization"/>
</fork>
<action name="copy_publication">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/publication</arg>
<arg>${nameNode}/${outputPath}/publication</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_dataset">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/dataset</arg>
<arg>${nameNode}/${outputPath}/dataset</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_orp">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/otherresearchproduct</arg>
<arg>${nameNode}/${outputPath}/otherresearchproduct</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_software">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/software</arg>
<arg>${nameNode}/${outputPath}/software</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_organization">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/organization</arg>
<arg>${nameNode}/${outputPath}/organization</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_projects">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/project</arg>
<arg>${nameNode}/${outputPath}/project</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<action name="copy_datasources">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<arg>${nameNode}/${sourcePath}/datasource</arg>
<arg>${nameNode}/${outputPath}/datasource</arg>
</distcp>
<ok to="wait"/>
<error to="Kill"/>
</action>
<join name="wait" to="read_blacklist"/>
<action name="read_blacklist">
<java>
<job-tracker>${jobTracker}</job-tracker>

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

@ -74,7 +74,7 @@ public class CommunityConfigurationFactoryTest {
new ArrayList<>(
Collections
.singletonList(
"Pallino ha aiutato a scrivere il paper. Pallino lavora per DARIAH")));
"Author X helped to write the paper. X works for DARIAH")));
List<String> comm = cc
.getCommunityForDatasource(
"openaire____::1cfdb2e14977f31a98e0118283401f32", param);