From 77a0951b322c0acb736ddad86cbcb5ed9970243a Mon Sep 17 00:00:00 2001
From: "miriam.baglioni" <miriam.baglioni@isti.cnr.it>
Date: Tue, 11 Aug 2020 15:32:32 +0200
Subject: [PATCH] added the make archive step in the workflow

---
 .../dhp/oa/graph/dump/oozie_app/workflow.xml  | 20 ++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)

diff --git a/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/oozie_app/workflow.xml b/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/oozie_app/workflow.xml
index 18c8bb682..d5b674715 100644
--- a/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/oozie_app/workflow.xml
@@ -118,7 +118,7 @@
 
     <action name="save_community_map">
         <java>
-            <main-class>eu.dnetlib.dhp.oa.graph.dump.community.SaveCommunityMap</main-class>
+            <main-class>eu.dnetlib.dhp.oa.graph.dump.SaveCommunityMap</main-class>
             <arg>--outputPath</arg><arg>${workingDir}/communityMap</arg>
             <arg>--nameNode</arg><arg>${nameNode}</arg>
             <arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
@@ -389,13 +389,25 @@
                     --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir}
                 </spark-opts>
                 <arg>--sourcePath</arg><arg>${workingDir}/ext</arg>
-                <arg>--outputPath</arg><arg>${outputPath}</arg>
+                <arg>--outputPath</arg><arg>${workingDir}/split</arg>
                 <arg>--communityMapPath</arg><arg>${workingDir}/communityMap</arg>
             </spark>
-            <ok to="send_zenodo"/>
+            <ok to="make_archive"/>
             <error to="Kill"/>
         </action>
 
+    <action name="make_archive">
+        <java>
+            <main-class>eu.dnetlib.dhp.oa.graph.dump.MakeTar</main-class>
+            <arg>--hdfsPath</arg><arg>${outputPath}</arg>
+            <arg>--nameNode</arg><arg>${nameNode}</arg>
+           <arg>--sourcePath</arg><arg>${workingDir}/split</arg>
+        </java>
+        <ok to="send_zenodo"/>
+        <error to="Kill"/>
+    </action>
+
+
     <action name="send_zenodo">
         <java>
             <main-class>eu.dnetlib.dhp.oa.graph.dump.SendToZenodoHDFS</main-class>
@@ -412,6 +424,8 @@
         <error to="Kill"/>
     </action>
 
+
+
     <end name="End"/>
 
 </workflow-app>
\ No newline at end of file