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

minor changes

This commit is contained in:
Claudio Atzori 2020-11-05 11:24:12 +01:00
parent 4625b7486e
commit 2148a51fae
2 changed files with 52 additions and 59 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

79
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/oozie_app/workflow.xml
View File

@ -1,18 +1,18 @@
<workflow-app name="dump_community_products" xmlns="uri:oozie:workflow:0.5"> <workflow-app name="dump_community_products" xmlns="uri:oozie:workflow:0.5">
<parameters> <parameters>
<property> <property>
<name>sourcePath</name> <name>sourcePath</name>
<description>the source path</description> <description>the source path</description>
</property> </property>
<property> <property>
<name>isLookUpUrl</name> <name>isLookUpUrl</name>
<description>the isLookup service endpoint</description> <description>the isLookup service endpoint</description>
</property> </property>
<property> <property>
<name>outputPath</name> <name>outputPath</name>
<description>the output path</description> <description>the output path</description>
</property> </property>
<property> <property>
<name>accessToken</name> <name>accessToken</name>
<description>the access token used for the deposition in Zenodo</description> <description>the access token used for the deposition in Zenodo</description>
@ -320,6 +320,7 @@
<ok to="join_extend"/> <ok to="join_extend"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="extend_orp"> <action name="extend_orp">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
@ -344,6 +345,7 @@
<ok to="join_extend"/> <ok to="join_extend"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="extend_software"> <action name="extend_software">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
@ -371,43 +373,42 @@
<join name="join_extend" to="splitForCommunities"/> <join name="join_extend" to="splitForCommunities"/>
<action name="splitForCommunities"> <action name="splitForCommunities">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
<mode>cluster</mode> <mode>cluster</mode>
<name>Split dumped result for community</name> <name>Split dumped result for community</name>
<class>eu.dnetlib.dhp.oa.graph.dump.community.SparkSplitForCommunity</class> <class>eu.dnetlib.dhp.oa.graph.dump.community.SparkSplitForCommunity</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar> <jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts> <spark-opts>
--executor-memory=${sparkExecutorMemory} --executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores} --executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory} --driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners} --conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
</spark-opts> </spark-opts>
<arg>--sourcePath</arg><arg>${workingDir}/ext</arg> <arg>--sourcePath</arg><arg>${workingDir}/ext</arg>
<arg>--outputPath</arg><arg>${workingDir}/split</arg> <arg>--outputPath</arg><arg>${workingDir}/split</arg>
<arg>--communityMapPath</arg><arg>${workingDir}/communityMap</arg> <arg>--communityMapPath</arg><arg>${workingDir}/communityMap</arg>
</spark> </spark>
<ok to="make_archive"/> <ok to="make_archive"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="make_archive"> <action name="make_archive">
<java> <java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.MakeTar</main-class> <main-class>eu.dnetlib.dhp.oa.graph.dump.MakeTar</main-class>
<arg>--hdfsPath</arg><arg>${outputPath}</arg> <arg>--hdfsPath</arg><arg>${outputPath}</arg>
<arg>--nameNode</arg><arg>${nameNode}</arg> <arg>--nameNode</arg><arg>${nameNode}</arg>
<arg>--sourcePath</arg><arg>${workingDir}/split</arg> <arg>--sourcePath</arg><arg>${workingDir}/split</arg>
</java> </java>
<ok to="send_zenodo"/> <ok to="send_zenodo"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="send_zenodo"> <action name="send_zenodo">
<java> <java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class> <main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class>
@ -424,8 +425,6 @@
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<end name="End"/> <end name="End"/>
</workflow-app> </workflow-app>

32
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump_whole/oozie_app/workflow.xml
View File

@ -1,18 +1,18 @@
<workflow-app name="dump_whole_graph" xmlns="uri:oozie:workflow:0.5"> <workflow-app name="dump_whole_graph" xmlns="uri:oozie:workflow:0.5">
<parameters> <parameters>
<property> <property>
<name>sourcePath</name> <name>sourcePath</name>
<description>the source path</description> <description>the source path</description>
</property> </property>
<property> <property>
<name>isLookUpUrl</name> <name>isLookUpUrl</name>
<description>the isLookup service endpoint</description> <description>the isLookup service endpoint</description>
</property> </property>
<property> <property>
<name>outputPath</name> <name>outputPath</name>
<description>the output path</description> <description>the output path</description>
</property> </property>
<property> <property>
<name>resultAggregation</name> <name>resultAggregation</name>
<description>true if all the result type have to be dumped under result. false otherwise</description> <description>true if all the result type have to be dumped under result. false otherwise</description>
@ -357,10 +357,8 @@
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<join name="join_dump" to="fork_context"/> <join name="join_dump" to="fork_context"/>
<fork name="fork_context"> <fork name="fork_context">
<path start="create_entities_fromcontext"/> <path start="create_entities_fromcontext"/>
<path start="create_relation_fromcontext"/> <path start="create_relation_fromcontext"/>
@ -389,7 +387,6 @@
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="create_relation_fromorgs"> <action name="create_relation_fromorgs">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
@ -418,7 +415,6 @@
<join name="join_context" to="fork_extract_relations"/> <join name="join_context" to="fork_extract_relations"/>
<fork name="fork_extract_relations"> <fork name="fork_extract_relations">
<path start="rels_from_pubs"/> <path start="rels_from_pubs"/>
<path start="rels_from_dats"/> <path start="rels_from_dats"/>
@ -530,7 +526,6 @@
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<join name="join_extract_relations" to="collect_and_save"/> <join name="join_extract_relations" to="collect_and_save"/>
<action name="collect_and_save"> <action name="collect_and_save">
@ -568,8 +563,7 @@
<ok to="send_zenodo"/> <ok to="send_zenodo"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="send_zenodo"> <action name="send_zenodo">
<java> <java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class> <main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class>