diff --git a/dhp-build/dhp-build-assembly-resources/pom.xml b/dhp-build/dhp-build-assembly-resources/pom.xml index c837cd538..9b03536dd 100644 --- a/dhp-build/dhp-build-assembly-resources/pom.xml +++ b/dhp-build/dhp-build-assembly-resources/pom.xml @@ -12,6 +12,8 @@ dhp-build-assembly-resources jar + This module contains a set of scripts supporting the build lifecycle for the dnet-hadoop project + diff --git a/dhp-build/dhp-build-properties-maven-plugin/pom.xml b/dhp-build/dhp-build-properties-maven-plugin/pom.xml index 82228fe07..4d40edd99 100644 --- a/dhp-build/dhp-build-properties-maven-plugin/pom.xml +++ b/dhp-build/dhp-build-properties-maven-plugin/pom.xml @@ -12,6 +12,7 @@ dhp-build-properties-maven-plugin maven-plugin + This module is a maven plugin implementing custom properties substitutions in the build lifecycle @@ -33,7 +34,7 @@ org.kuali.maven.plugins properties-maven-plugin - 2.0.1 + ${properties.maven.plugin.version} com.google.code.findbugs @@ -83,41 +84,6 @@ - - - - - org.eclipse.m2e - lifecycle-mapping - 1.0.0 - - - - - - - org.apache.maven.plugins - - - maven-plugin-plugin - - - [3.2,) - - - descriptor - - - - - - - - - - - - diff --git a/dhp-build/dhp-code-style/pom.xml b/dhp-build/dhp-code-style/pom.xml index 830723c9f..5e896e7a5 100644 --- a/dhp-build/dhp-code-style/pom.xml +++ b/dhp-build/dhp-code-style/pom.xml @@ -11,6 +11,38 @@ jar + This module contains resources supporting common code style conventions + + + + dnet45-snapshots + DNet45 Snapshots + http://maven.research-infrastructures.eu/nexus/content/repositories/dnet45-snapshots + default + + + dnet45-releases + http://maven.research-infrastructures.eu/nexus/content/repositories/dnet45-releases + + + + + + + + org.apache.maven.plugins + maven-project-info-reports-plugin + 3.0.0 + + + org.apache.maven.plugins + maven-site-plugin + 3.7.1 + + + + + UTF-8 diff --git a/dhp-build/pom.xml b/dhp-build/pom.xml index ff6a858bb..041641fcf 100644 --- a/dhp-build/pom.xml +++ b/dhp-build/pom.xml @@ -8,6 +8,9 @@ dhp-build pom + + This module is a container for the build tools used in dnet-hadoop + dhp-code-style dhp-build-assembly-resources diff --git a/dhp-common/pom.xml b/dhp-common/pom.xml index 7b073397f..51af8d954 100644 --- a/dhp-common/pom.xml +++ b/dhp-common/pom.xml @@ -12,6 +12,8 @@ dhp-common jar + This module contains common utilities meant to be used across the dnet-hadoop submodules + diff --git a/dhp-schemas/pom.xml b/dhp-schemas/pom.xml index 8deb2eab2..4a123ceda 100644 --- a/dhp-schemas/pom.xml +++ b/dhp-schemas/pom.xml @@ -12,7 +12,7 @@ dhp-schemas jar - + This module contains common schema classes meant to be used across the dnet-hadoop submodules diff --git a/dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java b/dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java index 23b30cc7d..a0ae7bc3c 100644 --- a/dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java +++ b/dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java @@ -1,3 +1,4 @@ + package eu.dnetlib.dhp.oa.dedup; import static java.nio.file.Files.createTempDirectory; diff --git a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplication.java b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplication.java index d88325c9b..c0469735c 100644 --- a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplication.java +++ b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplication.java @@ -51,7 +51,7 @@ import eu.dnetlib.dhp.schema.oaf.Software; import eu.dnetlib.dhp.schema.oaf.StructuredProperty; public class MigrateDbEntitiesApplication extends AbstractMigrationApplication - implements Closeable { + implements Closeable { private static final Log log = LogFactory.getLog(MigrateDbEntitiesApplication.class); @@ -61,9 +61,10 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication 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"))); + IOUtils + .toString( + MigrateDbEntitiesApplication.class + .getResourceAsStream("/eu/dnetlib/dhp/oa/graph/migrate_db_entities_parameters.json"))); parser.parseArgument(args); @@ -76,7 +77,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication final boolean processClaims = parser.get("action") != null && parser.get("action").equalsIgnoreCase("claims"); try (final MigrateDbEntitiesApplication smdbe = new MigrateDbEntitiesApplication(hdfsPath, dbUrl, dbUser, - dbPassword)) { + dbPassword)) { if (processClaims) { log.info("Processing claims..."); smdbe.execute("queryClaims.sql", smdbe::processClaims); @@ -107,15 +108,15 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication } public MigrateDbEntitiesApplication( - final String hdfsPath, final String dbUrl, final String dbUser, final String dbPassword) - throws Exception { + final String hdfsPath, final String dbUrl, final String dbUser, final String dbPassword) + throws Exception { super(hdfsPath); this.dbClient = new DbClient(dbUrl, dbUser, dbPassword); this.lastUpdateTimestamp = new Date().getTime(); } public void execute(final String sqlFile, final Function> producer) - throws Exception { + throws Exception { final String sql = IOUtils.toString(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/sql/" + sqlFile)); final Consumer consumer = rs -> producer.apply(rs).forEach(oaf -> emitOaf(oaf)); @@ -134,7 +135,10 @@ 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"))); + .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 @@ -175,7 +179,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication 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 + .setJournal(prepareJournal(rs.getString("officialname"), rs.getString("journal"), info)); // Journal ds.setDataInfo(info); ds.setLastupdatetimestamp(lastUpdateTimestamp); @@ -195,7 +199,10 @@ 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"))); + .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"))); @@ -212,7 +219,7 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication 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)); + .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)); @@ -249,7 +256,10 @@ 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"))); + .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"))); @@ -264,12 +274,14 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication 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)); + .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)); + .setEcinternationalorganizationeurinterests( + field(Boolean.toString(rs.getBoolean("ecinternationalorganizationeurinterests")), info)); o - .setEcinternationalorganization(field(Boolean.toString(rs.getBoolean("ecinternationalorganization")), info)); + .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)); @@ -288,7 +300,8 @@ 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 collectedFrom = listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")); + final List collectedFrom = listKeyValues( + createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")); final Relation r1 = new Relation(); r1.setRelType("datasourceOrganization"); @@ -321,7 +334,8 @@ 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 collectedFrom = listKeyValues(createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")); + final List collectedFrom = listKeyValues( + createOpenaireId(10, rs.getString("collectedfromid"), true), rs.getString("collectedfromname")); final Relation r1 = new Relation(); r1.setRelType("projectOrganization"); @@ -351,10 +365,12 @@ public class MigrateDbEntitiesApplication extends AbstractMigrationApplication public List processClaims(final ResultSet rs) { - final DataInfo info = - dataInfo(false, null, false, false, qualifier("user:claim", "user:claim", "dnet:provenanceActions", "dnet:provenanceActions"), "0.9"); + final DataInfo info = dataInfo( + false, null, false, false, + qualifier("user:claim", "user:claim", "dnet:provenanceActions", "dnet:provenanceActions"), "0.9"); - final List collectedFrom = listKeyValues(createOpenaireId(10, "infrastruct_::openaire", true), "OpenAIRE"); + final List collectedFrom = listKeyValues( + createOpenaireId(10, "infrastruct_::openaire", true), "OpenAIRE"); try { @@ -440,11 +456,15 @@ 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; } @@ -458,19 +478,23 @@ 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 prepareListOfStructProps( - final Array array, - final DataInfo dataInfo) throws SQLException { + final Array array, + final DataInfo dataInfo) throws SQLException { final List res = new ArrayList<>(); if (array != null) { for (final String s : (String[]) array.getArray()) { @@ -489,8 +513,10 @@ 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); } diff --git a/dhp-workflows/dhp-stats-update/pom.xml b/dhp-workflows/dhp-stats-update/pom.xml index 4d6318c3e..0f5e18082 100644 --- a/dhp-workflows/dhp-stats-update/pom.xml +++ b/dhp-workflows/dhp-stats-update/pom.xml @@ -22,6 +22,7 @@ pl.project13.maven git-commit-id-plugin + 2.1.11 false diff --git a/dhp-workflows/dhp-worfklow-profiles/pom.xml b/dhp-workflows/dhp-worfklow-profiles/pom.xml index df90014ba..bad72a9ef 100644 --- a/dhp-workflows/dhp-worfklow-profiles/pom.xml +++ b/dhp-workflows/dhp-worfklow-profiles/pom.xml @@ -3,13 +3,27 @@ xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> - dhp + dhp-workflows eu.dnetlib.dhp 1.1.7-SNAPSHOT 4.0.0 dhp-worfklow-profiles + jar + \ No newline at end of file diff --git a/dhp-workflows/pom.xml b/dhp-workflows/pom.xml index cd5784c09..1645129b1 100644 --- a/dhp-workflows/pom.xml +++ b/dhp-workflows/pom.xml @@ -13,6 +13,8 @@ dhp-workflows pom + This module is the container for the oozie workflow definitions in dnet-hadoop project + dhp-worfklow-profiles dhp-aggregation @@ -182,7 +184,7 @@ org.kuali.maven.plugins properties-maven-plugin - 1.3.2 + ${properties.maven.plugin.version} eu.dnetlib.dhp @@ -562,60 +564,4 @@ - - - - - - org.eclipse.m2e - lifecycle-mapping - 1.0.0 - - - - - - - org.kuali.maven.plugins - - - properties-maven-plugin - - - [1.3.2,) - - - - read-project-properties - - - write-project-properties - - - - - - - - - - - org.apache.maven.plugins - maven-dependency-plugin - [1.0.0,) - - copy-dependencies - - - - - - - - - - - - - diff --git a/pom.xml b/pom.xml index 0bc01177e..483873219 100644 --- a/pom.xml +++ b/pom.xml @@ -6,14 +6,14 @@ 1.1.7-SNAPSHOT pom - http://www.d-net.research-infrastructures.eu - - The Apache Software License, Version 2.0 - http://www.apache.org/licenses/LICENSE-2.0.txt + GNU Affero General Public License v3.0 or later + https://spdx.org/licenses/AGPL-3.0-or-later.html#licenseText repo - A business-friendly OSS license + This program is free software: you can redistribute it and/or modify it under the terms of the + GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the + License, or (at your option) any later version. @@ -41,6 +41,8 @@ HEAD + This module is the root descriptor for the dnet-hadoop project + @@ -406,6 +408,18 @@ target/test-classes + + + org.apache.maven.plugins + maven-project-info-reports-plugin + 3.0.0 + + + org.apache.maven.plugins + maven-site-plugin + 3.7.1 + + org.apache.maven.plugins maven-compiler-plugin @@ -476,6 +490,14 @@ + + org.apache.maven.plugins + maven-site-plugin + + + org.apache.maven.plugins + maven-project-info-reports-plugin + net.revelc.code.formatter formatter-maven-plugin @@ -583,6 +605,7 @@ UTF-8 3.6.0 2.22.2 + 2.0.1 cdh5.9.2 2.6.0-${dhp.cdh.version} 4.1.0-${dhp.cdh.version}