D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

Fix workflow application path

This commit is contained in:
Serafeim Chatzopoulos 2023-05-16 16:28:48 +03:00
parent 26328e2a0d
commit 8ef718c363
3 changed files with 21 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
dhp-workflows/dhp-impact-indicators/README.md
View File

@ -24,3 +24,13 @@ mvn package -Poozie-package,deploy,run -Dworkflow.source.dir=eu/dnetlib/dhp/oa/g
``` ```
Note: edit the property `bip.ranker.tag` of the `pom.xml` file to specify the tag of [BIP-Ranker](https://github.com/athenarc/Bip-Ranker) that you want to use. Note: edit the property `bip.ranker.tag` of the `pom.xml` file to specify the tag of [BIP-Ranker](https://github.com/athenarc/Bip-Ranker) that you want to use.
Job info and logs:
```
export OOZIE_URL=http://iis-cdh5-test-m3:11000/oozie
oozie job -info <jobId>
oozie job -log <jobId>
```
where `jobId` is the id of the job returned by the `run_workflow.sh` script.

9
dhp-workflows/dhp-impact-indicators/src/main/resources/eu/dnetlib/dhp/oa/graph/impact_indicators/job.properties
View File

@ -76,7 +76,7 @@ bipScorePath=${workingDir}/openaire_universe_scores/
checkpointDir=${nameNode}/${workingDir}/check/ checkpointDir=${nameNode}/${workingDir}/check/
# The directory for the doi-based bip graph # The directory for the doi-based bip graph
bipGraphFilePath=${nameNode}/${workingDir}/bipdbv8_graph # bipGraphFilePath=${nameNode}/${workingDir}/bipdbv8_graph
# The folder from which synonyms of openaire-ids are read # The folder from which synonyms of openaire-ids are read
# openaireDataInput=${nameNode}/tmp/beta_provision/graph/21_graph_cleaned/ # openaireDataInput=${nameNode}/tmp/beta_provision/graph/21_graph_cleaned/
@ -89,9 +89,12 @@ synonymFolder=${nameNode}/${workingDir}/openaireid_to_dois/
openaireGraphInputPath=${nameNode}/${workingDir}/openaire_id_graph openaireGraphInputPath=${nameNode}/${workingDir}/openaire_id_graph
# The workflow application path # The workflow application path
wfAppPath=${nameNode}/${oozieWorkflowPath} wfAppPath=${oozieTopWfApplicationPath}
# The following is needed as a property of a workflow # The following is needed as a property of a workflow
oozie.wf.application.path=${wfAppPath} #oozie.wf.application.path=${wfAppPath}
oozie.wf.application.path=${oozieTopWfApplicationPath}
# Path where the final output should be? # Path where the final output should be?
actionSetOutputPath=${workingDir}/bip_actionsets/ actionSetOutputPath=${workingDir}/bip_actionsets/

10
dhp-workflows/dhp-impact-indicators/src/main/resources/eu/dnetlib/dhp/oa/graph/impact_indicators/oozie_app/workflow.xml
View File

@ -126,7 +126,7 @@
<!-- number of partitions to be used on joins --> <!-- number of partitions to be used on joins -->
<arg>${sparkShufflePartitions}</arg> <arg>${sparkShufflePartitions}</arg>
<!-- This needs to point to the file on the hdfs i think --> <!-- This needs to point to the file on the hdfs i think -->
<file>${wfAppPath}/CC.py#CC.py</file> <file>${wfAppPath}/bip-ranker/CC.py#CC.py</file>
</spark> </spark>
<!-- Do this after finishing okay --> <!-- Do this after finishing okay -->
@ -171,7 +171,7 @@
<arg>${sparkShufflePartitions}</arg> <arg>${sparkShufflePartitions}</arg>
<arg>${checkpointDir}</arg> <arg>${checkpointDir}</arg>
<!-- This needs to point to the file on the hdfs i think --> <!-- This needs to point to the file on the hdfs i think -->
<file>${wfAppPath}/TAR.py#TAR.py</file> <file>${wfAppPath}/bip-ranker/TAR.py#TAR.py</file>
</spark> </spark>
<!-- Do this after finishing okay --> <!-- Do this after finishing okay -->
@ -216,7 +216,7 @@
<arg>${sparkShufflePartitions}</arg> <arg>${sparkShufflePartitions}</arg>
<arg>3</arg> <arg>3</arg>
<!-- This needs to point to the file on the hdfs i think --> <!-- This needs to point to the file on the hdfs i think -->
<file>${wfAppPath}/CC.py#CC.py</file> <file>${wfAppPath}/bip-ranker/CC.py#CC.py</file>
</spark> </spark>
<!-- Do this after finishing okay --> <!-- Do this after finishing okay -->
@ -274,7 +274,7 @@
<arg>${sparkShufflePartitions}</arg> <arg>${sparkShufflePartitions}</arg>
<arg>dfs</arg> <arg>dfs</arg>
<!-- This needs to point to the file on the hdfs i think --> <!-- This needs to point to the file on the hdfs i think -->
<file>${wfAppPath}/PageRank.py#PageRank.py</file> <file>${wfAppPath}/bip-ranker/PageRank.py#PageRank.py</file>
</spark> </spark>
<!-- Do this after finishing okay --> <!-- Do this after finishing okay -->
@ -324,7 +324,7 @@
<arg>${sparkShufflePartitions}</arg> <arg>${sparkShufflePartitions}</arg>
<arg>dfs</arg> <arg>dfs</arg>
<!-- This needs to point to the file on the hdfs i think --> <!-- This needs to point to the file on the hdfs i think -->
<file>${wfAppPath}/AttRank.py#AttRank.py</file> <file>${wfAppPath}/bip-ranker/AttRank.py#AttRank.py</file>
</spark> </spark>
<!-- Do this after finishing okay --> <!-- Do this after finishing okay -->