forked from D-Net/dnet-hadoop
formatting
This commit is contained in:
parent
a5d7007005
commit
c43b4c8962
|
|
@ -1,3 +1,4 @@
|
|||
|
||||
package eu.dnetlib.dhp.oa.graph.raw;
|
||||
|
||||
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.asString;
|
||||
|
|
@ -10,6 +11,23 @@ import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.listKeyValues;
|
|||
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.qualifier;
|
||||
import static eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils.structuredProperty;
|
||||
|
||||
import java.io.Closeable;
|
||||
import java.io.IOException;
|
||||
import java.sql.Array;
|
||||
import java.sql.ResultSet;
|
||||
import java.sql.SQLException;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.Date;
|
||||
import java.util.List;
|
||||
import java.util.function.Consumer;
|
||||
import java.util.function.Function;
|
||||
|
||||
import org.apache.commons.io.IOUtils;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.apache.commons.logging.Log;
|
||||
import org.apache.commons.logging.LogFactory;
|
||||
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
||||
import eu.dnetlib.dhp.oa.graph.raw.common.AbstractMigrationApplication;
|
||||
import eu.dnetlib.dhp.oa.graph.raw.common.DbClient;
|
||||
|
|
@ -31,21 +49,6 @@ import eu.dnetlib.dhp.schema.oaf.Relation;
|
|||
import eu.dnetlib.dhp.schema.oaf.Result;
|
||||
import eu.dnetlib.dhp.schema.oaf.Software;
|
||||
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
|
||||
import java.io.Closeable;
|
||||
import java.io.IOException;
|
||||
import java.sql.Array;
|
||||
import java.sql.ResultSet;
|
||||
import java.sql.SQLException;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.Date;
|
||||
import java.util.List;
|
||||
import java.util.function.Consumer;
|
||||
import java.util.function.Function;
|
||||
import org.apache.commons.io.IOUtils;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.apache.commons.logging.Log;
|
||||
import org.apache.commons.logging.LogFactory;
|
||||
|
||||
public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
||||
implements Closeable {
|
||||
|
|
@ -57,11 +60,10 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
private final long lastUpdateTimestamp;
|
||||
|
||||
public static void main(final String[] args) throws Exception {
|
||||
final ArgumentApplicationParser parser =
|
||||
new ArgumentApplicationParser(
|
||||
IOUtils.toString(
|
||||
MigrateDbEntitiesApplication.class.getResourceAsStream(
|
||||
"/eu/dnetlib/dhp/oa/graph/migrate_db_entities_parameters.json")));
|
||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
||||
IOUtils
|
||||
.toString(MigrateDbEntitiesApplication.class
|
||||
.getResourceAsStream("/eu/dnetlib/dhp/oa/graph/migrate_db_entities_parameters.json")));
|
||||
|
||||
parser.parseArgument(args);
|
||||
|
||||
|
|
@ -71,11 +73,10 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
final String hdfsPath = parser.get("hdfsPath");
|
||||
|
||||
final boolean processClaims =
|
||||
parser.get("action") != null && parser.get("action").equalsIgnoreCase("claims");
|
||||
final boolean processClaims = parser.get("action") != null && parser.get("action").equalsIgnoreCase("claims");
|
||||
|
||||
try (final MigrateDbEntitiesApplication smdbe =
|
||||
new MigrateDbEntitiesApplication(hdfsPath, dbUrl, dbUser, dbPassword)) {
|
||||
try (final MigrateDbEntitiesApplication smdbe = new MigrateDbEntitiesApplication(hdfsPath, dbUrl, dbUser,
|
||||
dbPassword)) {
|
||||
if (processClaims) {
|
||||
log.info("Processing claims...");
|
||||
smdbe.execute("queryClaims.sql", smdbe::processClaims);
|
||||
|
|
@ -115,8 +116,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
public void execute(final String sqlFile, final Function<ResultSet, List<Oaf>> producer)
|
||||
throws Exception {
|
||||
final String sql =
|
||||
IOUtils.toString(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/sql/" + sqlFile));
|
||||
final String sql = IOUtils.toString(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/sql/" + sqlFile));
|
||||
|
||||
final Consumer<ResultSet> consumer = rs -> producer.apply(rs).forEach(oaf -> emitOaf(oaf));
|
||||
|
||||
|
|
@ -133,10 +133,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
ds.setId(createOpenaireId(10, rs.getString("datasourceid"), true));
|
||||
ds.setOriginalId(Arrays.asList(rs.getString("datasourceid")));
|
||||
ds.setCollectedfrom(
|
||||
listKeyValues(
|
||||
createOpenaireId(10, rs.getString("collectedfromid"), true),
|
||||
rs.getString("collectedfromname")));
|
||||
ds
|
||||
.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
|
||||
ds.setPid(new ArrayList<>());
|
||||
ds.setDateofcollection(asString(rs.getDate("dateofcollection")));
|
||||
ds.setDateoftransformation(null); // Value not returned by the SQL query
|
||||
|
|
@ -176,8 +174,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
ds.setPidsystems(field(rs.getString("pidsystems"), info));
|
||||
ds.setCertificates(field(rs.getString("certificates"), info));
|
||||
ds.setPolicies(new ArrayList<>()); // The sql query returns an empty array
|
||||
ds.setJournal(
|
||||
prepareJournal(rs.getString("officialname"), rs.getString("journal"), info)); // Journal
|
||||
ds
|
||||
.setJournal(prepareJournal(rs.getString("officialname"), rs.getString("journal"), info)); // Journal
|
||||
ds.setDataInfo(info);
|
||||
ds.setLastupdatetimestamp(lastUpdateTimestamp);
|
||||
|
||||
|
|
@ -196,10 +194,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
p.setId(createOpenaireId(40, rs.getString("projectid"), true));
|
||||
p.setOriginalId(Arrays.asList(rs.getString("projectid")));
|
||||
p.setCollectedfrom(
|
||||
listKeyValues(
|
||||
createOpenaireId(10, rs.getString("collectedfromid"), true),
|
||||
rs.getString("collectedfromname")));
|
||||
p
|
||||
.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
|
||||
p.setPid(new ArrayList<>());
|
||||
p.setDateofcollection(asString(rs.getDate("dateofcollection")));
|
||||
p.setDateoftransformation(asString(rs.getDate("dateoftransformation")));
|
||||
|
|
@ -215,8 +211,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
p.setKeywords(field(rs.getString("keywords"), info));
|
||||
p.setDuration(field(Integer.toString(rs.getInt("duration")), info));
|
||||
p.setEcsc39(field(Boolean.toString(rs.getBoolean("ecsc39")), info));
|
||||
p.setOamandatepublications(
|
||||
field(Boolean.toString(rs.getBoolean("oamandatepublications")), info));
|
||||
p
|
||||
.setOamandatepublications(field(Boolean.toString(rs.getBoolean("oamandatepublications")), info));
|
||||
p.setEcarticle29_3(field(Boolean.toString(rs.getBoolean("ecarticle29_3")), info));
|
||||
p.setSubjects(prepareListOfStructProps(rs.getArray("subjects"), info));
|
||||
p.setFundingtree(prepareListFields(rs.getArray("fundingtree"), info));
|
||||
|
|
@ -252,10 +248,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
o.setId(createOpenaireId(20, rs.getString("organizationid"), true));
|
||||
o.setOriginalId(Arrays.asList(rs.getString("organizationid")));
|
||||
o.setCollectedfrom(
|
||||
listKeyValues(
|
||||
createOpenaireId(10, rs.getString("collectedfromid"), true),
|
||||
rs.getString("collectedfromname")));
|
||||
o
|
||||
.setCollectedfrom(listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")));
|
||||
o.setPid(new ArrayList<>());
|
||||
o.setDateofcollection(asString(rs.getDate("dateofcollection")));
|
||||
o.setDateoftransformation(asString(rs.getDate("dateoftransformation")));
|
||||
|
|
@ -269,13 +263,13 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
o.setEclegalbody(field(Boolean.toString(rs.getBoolean("eclegalbody")), info));
|
||||
o.setEclegalperson(field(Boolean.toString(rs.getBoolean("eclegalperson")), info));
|
||||
o.setEcnonprofit(field(Boolean.toString(rs.getBoolean("ecnonprofit")), info));
|
||||
o.setEcresearchorganization(
|
||||
field(Boolean.toString(rs.getBoolean("ecresearchorganization")), info));
|
||||
o
|
||||
.setEcresearchorganization(field(Boolean.toString(rs.getBoolean("ecresearchorganization")), info));
|
||||
o.setEchighereducation(field(Boolean.toString(rs.getBoolean("echighereducation")), info));
|
||||
o.setEcinternationalorganizationeurinterests(
|
||||
field(Boolean.toString(rs.getBoolean("ecinternationalorganizationeurinterests")), info));
|
||||
o.setEcinternationalorganization(
|
||||
field(Boolean.toString(rs.getBoolean("ecinternationalorganization")), info));
|
||||
o
|
||||
.setEcinternationalorganizationeurinterests(field(Boolean.toString(rs.getBoolean("ecinternationalorganizationeurinterests")), info));
|
||||
o
|
||||
.setEcinternationalorganization(field(Boolean.toString(rs.getBoolean("ecinternationalorganization")), info));
|
||||
o.setEcenterprise(field(Boolean.toString(rs.getBoolean("ecenterprise")), info));
|
||||
o.setEcsmevalidated(field(Boolean.toString(rs.getBoolean("ecsmevalidated")), info));
|
||||
o.setEcnutscode(field(Boolean.toString(rs.getBoolean("ecnutscode")), info));
|
||||
|
|
@ -294,10 +288,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
final DataInfo info = prepareDataInfo(rs);
|
||||
final String orgId = createOpenaireId(20, rs.getString("organization"), true);
|
||||
final String dsId = createOpenaireId(10, rs.getString("datasource"), true);
|
||||
final List<KeyValue> collectedFrom =
|
||||
listKeyValues(
|
||||
createOpenaireId(10, rs.getString("collectedfromid"), true),
|
||||
rs.getString("collectedfromname"));
|
||||
final List<KeyValue> collectedFrom = listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname"));
|
||||
|
||||
final Relation r1 = new Relation();
|
||||
r1.setRelType("datasourceOrganization");
|
||||
|
|
@ -330,10 +321,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
final DataInfo info = prepareDataInfo(rs);
|
||||
final String orgId = createOpenaireId(20, rs.getString("resporganization"), true);
|
||||
final String projectId = createOpenaireId(40, rs.getString("project"), true);
|
||||
final List<KeyValue> collectedFrom =
|
||||
listKeyValues(
|
||||
createOpenaireId(10, rs.getString("collectedfromid"), true),
|
||||
rs.getString("collectedfromname"));
|
||||
final List<KeyValue> collectedFrom = listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname"));
|
||||
|
||||
final Relation r1 = new Relation();
|
||||
r1.setRelType("projectOrganization");
|
||||
|
|
@ -364,17 +352,9 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
public List<Oaf> processClaims(final ResultSet rs) {
|
||||
|
||||
final DataInfo info =
|
||||
dataInfo(
|
||||
false,
|
||||
null,
|
||||
false,
|
||||
false,
|
||||
qualifier(
|
||||
"user:claim", "user:claim", "dnet:provenanceActions", "dnet:provenanceActions"),
|
||||
"0.9");
|
||||
dataInfo(false, null, false, false, qualifier("user:claim", "user:claim", "dnet:provenanceActions", "dnet:provenanceActions"), "0.9");
|
||||
|
||||
final List<KeyValue> collectedFrom =
|
||||
listKeyValues(createOpenaireId(10, "infrastruct_::openaire", true), "OpenAIRE");
|
||||
final List<KeyValue> collectedFrom = listKeyValues(createOpenaireId(10, "infrastruct_::openaire", true), "OpenAIRE");
|
||||
|
||||
try {
|
||||
|
||||
|
|
@ -402,10 +382,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
|
||||
return Arrays.asList(r);
|
||||
} else {
|
||||
final String sourceId =
|
||||
createOpenaireId(rs.getString("source_type"), rs.getString("source_id"), false);
|
||||
final String targetId =
|
||||
createOpenaireId(rs.getString("target_type"), rs.getString("target_id"), false);
|
||||
final String sourceId = createOpenaireId(rs.getString("source_type"), rs.getString("source_id"), false);
|
||||
final String targetId = createOpenaireId(rs.getString("target_type"), rs.getString("target_id"), false);
|
||||
|
||||
final Relation r1 = new Relation();
|
||||
final Relation r2 = new Relation();
|
||||
|
|
@ -462,19 +440,11 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
final String inferenceprovenance = rs.getString("inferenceprovenance");
|
||||
final Boolean inferred = rs.getBoolean("inferred");
|
||||
final String trust = rs.getString("trust");
|
||||
return dataInfo(
|
||||
deletedbyinference,
|
||||
inferenceprovenance,
|
||||
inferred,
|
||||
false,
|
||||
MigrationConstants.ENTITYREGISTRY_PROVENANCE_ACTION,
|
||||
trust);
|
||||
return dataInfo(deletedbyinference, inferenceprovenance, inferred, false, MigrationConstants.ENTITYREGISTRY_PROVENANCE_ACTION, trust);
|
||||
}
|
||||
|
||||
private Qualifier prepareQualifierSplitting(final String s) {
|
||||
if (StringUtils.isBlank(s)) {
|
||||
return null;
|
||||
}
|
||||
if (StringUtils.isBlank(s)) { return null; }
|
||||
final String[] arr = s.split("@@@");
|
||||
return arr.length == 4 ? qualifier(arr[0], arr[1], arr[2], arr[3]) : null;
|
||||
}
|
||||
|
|
@ -488,22 +458,19 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
}
|
||||
|
||||
private StructuredProperty prepareStructProp(final String s, final DataInfo dataInfo) {
|
||||
if (StringUtils.isBlank(s)) {
|
||||
return null;
|
||||
}
|
||||
if (StringUtils.isBlank(s)) { return null; }
|
||||
final String[] parts = s.split("###");
|
||||
if (parts.length == 2) {
|
||||
final String value = parts[0];
|
||||
final String[] arr = parts[1].split("@@@");
|
||||
if (arr.length == 4) {
|
||||
return structuredProperty(value, arr[0], arr[1], arr[2], arr[3], dataInfo);
|
||||
}
|
||||
if (arr.length == 4) { return structuredProperty(value, arr[0], arr[1], arr[2], arr[3], dataInfo); }
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
private List<StructuredProperty> prepareListOfStructProps(
|
||||
final Array array, final DataInfo dataInfo) throws SQLException {
|
||||
final Array array,
|
||||
final DataInfo dataInfo) throws SQLException {
|
||||
final List<StructuredProperty> res = new ArrayList<>();
|
||||
if (array != null) {
|
||||
for (final String s : (String[]) array.getArray()) {
|
||||
|
|
@ -522,10 +489,8 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication
|
|||
final String[] arr = sj.split("@@@");
|
||||
if (arr.length == 3) {
|
||||
final String issn = StringUtils.isNotBlank(arr[0]) ? arr[0] : null;
|
||||
final String eissn = StringUtils.isNotBlank(arr[1]) ? arr[1] : null;
|
||||
;
|
||||
final String lissn = StringUtils.isNotBlank(arr[2]) ? arr[2] : null;
|
||||
;
|
||||
final String eissn = StringUtils.isNotBlank(arr[1]) ? arr[1] : null;;
|
||||
final String lissn = StringUtils.isNotBlank(arr[2]) ? arr[2] : null;;
|
||||
if (issn != null || eissn != null || lissn != null) {
|
||||
return journal(name, issn, eissn, eissn, null, null, null, null, null, null, null, info);
|
||||
}
|
||||
|
|
|
|||
|
|
@ -3,10 +3,10 @@
|
|||
<parent>
|
||||
<artifactId>dhp-workflows</artifactId>
|
||||
<groupId>eu.dnetlib.dhp</groupId>
|
||||
<version>1.1.6-SNAPSHOT</version>
|
||||
<version>1.1.7-SNAPSHOT</version>
|
||||
</parent>
|
||||
<modelVersion>4.0.0</modelVersion>
|
||||
<artifactId>dhp-statistics-workflow</artifactId>
|
||||
<artifactId>dhp-stats-update</artifactId>
|
||||
<dependencies>
|
||||
<dependency>
|
||||
<groupId>org.apache.spark</groupId>
|
||||
|
|
|
|||
Loading…
Reference in New Issue