diff --git a/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/provision.xml b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_construction.xml similarity index 99% rename from dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/provision.xml rename to dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_construction.xml index 28cbde70d..819b3e12d 100644 --- a/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/provision.xml +++ b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_construction.xml @@ -7,7 +7,7 @@ - Data Provision [OCEAN] + Graph Construction [OCEAN] Data Provision 30 diff --git a/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_to_hiveDB.xml b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_to_hiveDB.xml new file mode 100644 index 000000000..0ace12ea3 --- /dev/null +++ b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/graph_to_hiveDB.xml @@ -0,0 +1,73 @@ + +
+ + + + + +
+ + Graph to HiveDB [OCEAN] + Data Provision + 30 + + + Set the path containing the AGGREGATOR graph + + inputPath + + + + + + + + Set the target path to store the RAW graph + + hiveDbName + + + + + + + + + wait configurations + + + + + + + create the AGGREGATOR graph + + executeOozieJob + IIS + + { + 'inputPath' : 'inputPath', + 'hiveDbName' : 'hiveDbName' + } + + + { + 'oozie.wf.application.path' : '/lib/dnet/oa/graph/hive/oozie_app' + } + + build-report + + + + + + + + + wf_20200615_163630_609 + 2020-06-15T17:08:00+00:00 + SUCCESS + + + +
\ No newline at end of file diff --git a/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_solr.xml b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_solr.xml new file mode 100644 index 000000000..8a7738bcf --- /dev/null +++ b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_solr.xml @@ -0,0 +1,98 @@ + +
+ + + + + +
+ + Update Solr [OCEAN] + Data Provision + 30 + + + Set the path containing the AGGREGATOR graph + + inputGraphRootPath + + + + + + + + Set the target path to store the RAW graph + + format + TMF + + + + + + + Set the lookup address + + isLookupUrl + http://beta.services.openaire.eu:8280/is/services/isLookUp?wsdl + + + + + + + + wait configurations + + + + + + + create the AGGREGATOR graph + + executeOozieJob + IIS + + { + 'inputGraphRootPath' : 'inputGraphRootPath', + 'isLookupUrl' : 'isLookupUrl', + 'format' : 'format' + } + + + { + 'oozie.wf.application.path' : '/lib/dnet/oa/provision/oozie_app', + 'maxRelations' : '100', + 'relPartitions' : '3000', + 'batchSize' : '2000', + 'relationFilter' : 'isAuthorInstitutionOf,produces,hasAmongTopNSimilarDocuments', + 'otherDsTypeId' : 'scholarcomminfra,infospace,pubsrepository::mock,entityregistry,entityregistry::projects,entityregistry::repositories,websource', + 'resumeFrom' : 'prepare_relations', + 'sparkDriverMemoryForJoining' : '3G', + 'sparkExecutorMemoryForJoining' : '7G', + 'sparkExecutorCoresForJoining' : '4', + 'sparkDriverMemoryForIndexing' : '2G', + 'sparkExecutorMemoryForIndexing' : '2G', + 'sparkExecutorCoresForIndexing' : '64', + 'sparkNetworkTimeout' : '600', + 'workingDir' : '/tmp/beta_provision/working_dir/update_solr' + } + + build-report + + + + + + + + + wf_20200615_163630_609 + 2020-06-15T17:08:00+00:00 + SUCCESS + + + +
\ No newline at end of file diff --git a/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_stats.xml b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_stats.xml new file mode 100644 index 000000000..a91b6302e --- /dev/null +++ b/dhp-workflows/dhp-worfklow-profiles/src/main/resources/eu/dnetlib/dhp/wf/profiles/update_stats.xml @@ -0,0 +1,74 @@ + +
+ + + + + +
+ + Update Stats [OCEAN] + Data Provision + 30 + + + Set the path containing the AGGREGATOR graph + + openaire_db_name + + + + + + + + Set the target path to store the RAW graph + + stats_db_name + + + + + + + + + wait configurations + + + + + + + create the AGGREGATOR graph + + executeOozieJob + IIS + + { + 'openaire_db_name' : 'openaire_db_name', + 'stats_db_name' : 'stats_db_name' + } + + + { + 'oozie.wf.application.path' : '/lib/dnet/oa/graph/stats/oozie_app', + 'hive_timeout' : '3000' + } + + build-report + + + + + + + + + wf_20200615_163630_609 + 2020-06-15T17:08:00+00:00 + SUCCESS + + + +
\ No newline at end of file