From b33dd58be4dbaeb841363c61d95f3f9bab890d25 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Fri, 22 May 2020 08:50:06 +0200
Subject: [PATCH] replaced parameter 'reuseRecords' with 'resumeFrom', allowing
 to restart the provision workflow execution from any step, useful for manual
 submissions or debugging

---
 .../dhp/oa/provision/oozie_app/workflow.xml   | 22 +++++++++----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml b/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
index fd8f5ba89a..6983ecf53c 100644
--- a/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-graph-provision/src/main/resources/eu/dnetlib/dhp/oa/provision/oozie_app/workflow.xml
@@ -84,7 +84,6 @@
             <name>sparkNetworkTimeout</name>
             <description>configures spark.network.timeout</description>
         </property>
-
     </parameters>
 
     <global>
@@ -98,12 +97,16 @@
         </configuration>
     </global>
 
-    <start to="reuse_records"/>
+    <start to="resume_from"/>
 
-    <decision name="reuse_records">
+    <decision name="resume_from">
         <switch>
-            <case to="prepare_relations">${wf:conf('reuseRecords') eq false}</case>
-            <case to="to_solr_index">${wf:conf('reuseRecords') eq true}</case>
+            <case to="prepare_relations">${wf:conf('resumeFrom') eq 'prepare_relations'}</case>
+            <case to="fork_join_related_entities">${wf:conf('resumeFrom') eq 'fork_join_related_entities'}</case>
+            <case to="join_all_entities">${wf:conf('resumeFrom') eq 'join_all_entities'}</case>
+            <case to="adjancency_lists">${wf:conf('resumeFrom') eq 'adjancency_lists'}</case>
+            <case to="convert_to_xml">${wf:conf('resumeFrom') eq 'convert_to_xml'}</case>
+            <case to="to_solr_index">${wf:conf('resumeFrom') eq 'to_solr_index'}</case>
             <default to="prepare_relations"/>
         </switch>
     </decision>
@@ -131,9 +134,7 @@
             </spark-opts>
             <arg>--inputRelationsPath</arg><arg>${inputGraphRootPath}/relation</arg>
             <arg>--outputPath</arg><arg>${workingDir}/relation</arg>
-            <arg>--relPartitions</arg><arg>${relPartitions}</arg>
-            <arg>--relationFilter</arg><arg>${relationFilter}</arg>
-            <arg>--maxRelations</arg><arg>${maxRelations}</arg>
+            <arg>--relPartitions</arg><arg>3000</arg>
         </spark>
         <ok to="fork_join_related_entities"/>
         <error to="Kill"/>
@@ -340,7 +341,6 @@
 
     <join name="wait_joins" to="join_all_entities"/>
 
-
     <action name="join_all_entities">
         <spark xmlns="uri:oozie:spark-action:0.2">
             <master>yarn</master>
@@ -362,7 +362,7 @@
             <arg>--inputGraphRootPath</arg><arg>${inputGraphRootPath}</arg>
             <arg>--inputRelatedEntitiesPath</arg><arg>${workingDir}/join_partial</arg>
             <arg>--outputPath</arg><arg>${workingDir}/join_entities</arg>
-            <arg>--numPartitions</arg><arg>24000</arg>
+            <arg>--numPartitions</arg><arg>12000</arg>
         </spark>
         <ok to="adjancency_lists"/>
         <error to="Kill"/>
@@ -386,7 +386,7 @@
                 --conf spark.sql.shuffle.partitions=7680
                 --conf spark.network.timeout=${sparkNetworkTimeout}
             </spark-opts>
-            <arg>--inputPath</arg><arg>${workingDir}/join_entities</arg>
+            <arg>--inputPath</arg> <arg>${workingDir}/join_entities</arg>
             <arg>--outputPath</arg><arg>${workingDir}/joined</arg>
         </spark>
         <ok to="convert_to_xml"/>