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

modified the code to allow the dump for a single community (indip from its status)

This commit is contained in:
Miriam Baglioni 2021-04-23 12:41:32 +02:00
parent 3c3e3537e0
commit 952d6dc2fb
4 changed files with 57 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/dump/QueryInformationSystem.java
View File

@ -17,7 +17,7 @@ public class QueryInformationSystem {
private ISLookUpService isLookUp;
private static final String XQUERY = "for $x in collection('/db/DRIVER/ContextDSResources/ContextDSResourceType') "
private static final String XQUERY_ALL = "for $x in collection('/db/DRIVER/ContextDSResources/ContextDSResourceType') "
+
" where $x//CONFIGURATION/context[./@type='community' or ./@type='ri'] " +
" and ($x//context/param[./@name = 'status']/text() = 'all') "
@ -28,9 +28,22 @@ public class QueryInformationSystem {
"{$x//CONFIGURATION/context/@label}" +
"</community>";
public CommunityMap getCommunityMap()
private static final String XQUERY_CI = "for $x in collection('/db/DRIVER/ContextDSResources/ContextDSResourceType') "
+
" where $x//CONFIGURATION/context[./@type='community' or ./@type='ri'] " +
" and $x//CONFIGURATION/context[./@id=%s] "
+
" return " +
"<community> " +
"{$x//CONFIGURATION/context/@id}" +
"{$x//CONFIGURATION/context/@label}" +
"</community>";
public CommunityMap getCommunityMap(boolean singleCommunity, String community_id)
throws ISLookUpException, DocumentException {
return getMap(isLookUp.quickSearchProfile(XQUERY));
if (singleCommunity)
return getMap(isLookUp.quickSearchProfile(XQUERY_CI.replace("%s", "'" + community_id + "'")));
return getMap(isLookUp.quickSearchProfile(XQUERY_ALL));
}

12
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/dump/SaveCommunityMap.java
View File

@ -6,6 +6,7 @@ import java.io.IOException;
import java.io.OutputStreamWriter;
import java.io.Serializable;
import java.nio.charset.StandardCharsets;
import java.util.Optional;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
@ -71,14 +72,19 @@ public class SaveCommunityMap implements Serializable {
final String isLookUpUrl = parser.get("isLookUpUrl");
log.info("isLookUpUrl: {}", isLookUpUrl);
final Boolean singleCommunity = Optional.ofNullable(parser.get("singleDeposition"))
.map(Boolean::valueOf).orElse(false);
final String community_id = Optional.ofNullable(parser.get("communityId")).orElse(null);
final SaveCommunityMap scm = new SaveCommunityMap(outputPath, nameNode, isLookUpUrl);
scm.saveCommunityMap();
scm.saveCommunityMap(singleCommunity, community_id);
}
private void saveCommunityMap() throws ISLookUpException, IOException, DocumentException {
writer.write(Utils.OBJECT_MAPPER.writeValueAsString(queryInformationSystem.getCommunityMap()));
private void saveCommunityMap(boolean singleCommunity, String community_id) throws ISLookUpException, IOException, DocumentException {
writer.write(Utils.OBJECT_MAPPER.writeValueAsString(queryInformationSystem.getCommunityMap(singleCommunity, community_id)));
writer.close();
}
}

38
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/community/oozie_app/workflow.xml
View File

@ -134,29 +134,31 @@
<arg>--outputPath</arg><arg>${workingDir}/communityMap</arg>
<arg>--nameNode</arg><arg>${nameNode}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
</java>
<ok to="single_deposition"/>
<error to="Kill"/>
</action>
<decision name="single_deposition">
<switch>
<case to="remove_communities">${wf:conf('singleDeposition') eq true}</case>
<default to="fork_dump"/>
</switch>
</decision>
<action name="remove_communities">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.dump.community.RemoveCommunities</main-class>
<arg>--path</arg><arg>${workingDir}/communityMap</arg>
<arg>--nameNode</arg><arg>${nameNode}</arg>
<arg>--communityId</arg><arg>${communityId}</arg>
<arg>--singleDeposition</arg><arg>${singleDeposition}</arg>
<arg>--communityId></arg><arg>${communityId}</arg>
</java>
<ok to="fork_dump"/>
<error to="Kill"/>
</action>
<!-- <decision name="single_deposition">-->
<!-- <switch>-->
<!-- <case to="remove_communities">${wf:conf('singleDeposition') eq true}</case>-->
<!-- <default to="fork_dump"/>-->
<!-- </switch>-->
<!-- </decision>-->
<!-- <action name="remove_communities">-->
<!-- <java>-->
<!-- <main-class>eu.dnetlib.dhp.oa.graph.dump.community.RemoveCommunities</main-class>-->
<!-- <arg>&#45;&#45;path</arg><arg>${workingDir}/communityMap</arg>-->
<!-- <arg>&#45;&#45;nameNode</arg><arg>${nameNode}</arg>-->
<!-- <arg>&#45;&#45;communityId</arg><arg>${communityId}</arg>-->
<!-- </java>-->
<!-- <ok to="fork_dump"/>-->
<!-- <error to="Kill"/>-->
<!-- </action>-->
<fork name="fork_dump">
<path start="dump_publication"/>
<path start="dump_dataset"/>

12
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/input_cm_parameters.json
View File

@ -18,6 +18,18 @@
"paramLongName": "outputPath",
"paramDescription": "the path used to store temporary output files",
"paramRequired": true
},
{
"paramName": "sd",
"paramLongName": "singleDeposition",
"paramDescription": "true if the dump should be created for a single community",
"paramRequired": true
},
{
"paramName": "ci",
"paramLongName": "communityId",
"paramDescription": "the id of the community for which to create the dump",
"paramRequired": true
}
]