Revision 51006
Added by Claudio Atzori over 6 years ago
| modules/dnet-mapreduce-jobs/branches/beta/src/test/java/eu/dnetlib/data/mapreduce/hbase/broker/EventFactoryTest.java | ||
|---|---|---|
| 2 | 2 |
|
| 3 | 3 |
import java.util.UUID; |
| 4 | 4 |
|
| 5 |
import eu.dnetlib.data.mapreduce.hbase.broker.model.EventWrapper; |
|
| 6 |
import eu.dnetlib.data.proto.FieldTypeProtos.StringField; |
|
| 5 | 7 |
import org.apache.commons.lang.math.RandomUtils; |
| 6 | 8 |
import org.apache.commons.logging.Log; |
| 7 | 9 |
import org.apache.commons.logging.LogFactory; |
| ... | ... | |
| 50 | 52 |
Result.newBuilder().setMetadata( |
| 51 | 53 |
Result.Metadata.newBuilder().setDateofacceptance(OafTest.sf(""))
|
| 52 | 54 |
.addTitle(sp("aaaa Title", "main title", "dnet:dataCite_title"))
|
| 55 |
.setDateofacceptance(sf("2017-07-13"))
|
|
| 53 | 56 |
.setResulttype(q("publication", "dnet:result_typologies")))
|
| 54 | 57 |
.addInstance( |
| 55 | 58 |
Instance.newBuilder() |
| ... | ... | |
| 61 | 64 |
log.info("oaf: " + JsonFormat.printToString(oaf));
|
| 62 | 65 |
|
| 63 | 66 |
final float trust = RandomUtils.nextFloat(); |
| 64 |
final EventMessage event = EventFactory.asEvent(oaf, Topic.ENRICH_MISSING_PUBLICATION_DATE, new OpenAireEventPayload(), source, trust); |
|
| 67 |
final OpenAireEventPayload openAireEventPayload = new OpenAireEventPayload(); |
|
| 68 |
final EventMessage event = EventFactory.asEvent(oaf, Topic.ENRICH_MISSING_PUBLICATION_DATE, openAireEventPayload, source, trust); |
|
| 65 | 69 |
|
| 66 | 70 |
final OpenAireEventPayload p = OpenAireEventPayloadFactory.fromOAF(oaf, source, trust); |
| 67 | 71 |
final OpenAireEventPayload payload = HighlightFactory.highlightEnrichPublicationDate(p, OafTest.sf("2012-01-01"));
|
| ... | ... | |
| 73 | 77 |
|
| 74 | 78 |
event.setPayload(payload.toJSON()); |
| 75 | 79 |
|
| 76 |
log.info("event: " + event.toString());
|
|
| 80 |
log.info("event message: " + event.toString());
|
|
| 77 | 81 |
|
| 82 |
final EventWrapper wrapper = new EventWrapper(event, "test"); |
|
| 83 |
|
|
| 84 |
log.info("broker event: " + wrapper.asBrokerEvent().toJson());
|
|
| 85 |
|
|
| 78 | 86 |
} |
| 79 | 87 |
|
| 80 | 88 |
private Qualifier.Builder q(final String classid, final String scheme) {
|
| ... | ... | |
| 85 | 93 |
return KeyValue.newBuilder().setKey(k).setValue(v); |
| 86 | 94 |
} |
| 87 | 95 |
|
| 96 |
private StringField.Builder sf(final String value) {
|
|
| 97 |
return StringField.newBuilder().setValue(value); |
|
| 98 |
} |
|
| 99 |
|
|
| 88 | 100 |
private StructuredProperty.Builder sp(final String v, final String classid, final String scheme) {
|
| 89 | 101 |
return OafTest.getStructuredproperty(v, classid, scheme); |
| 90 | 102 |
} |
| modules/dnet-mapreduce-jobs/branches/beta/src/main/java/eu/dnetlib/data/mapreduce/hbase/broker/model/Event.java | ||
|---|---|---|
| 4 | 4 |
import java.util.Map; |
| 5 | 5 |
|
| 6 | 6 |
import com.google.gson.Gson; |
| 7 |
import com.google.gson.GsonBuilder; |
|
| 7 | 8 |
|
| 8 | 9 |
public class Event {
|
| 9 | 10 |
|
| 11 |
public static final String DATE_FORMAT = "yyyy-MM-dd'T'HH:mm:ss.SSS'Z'"; |
|
| 12 |
|
|
| 10 | 13 |
private String eventId; |
| 11 | 14 |
|
| 12 | 15 |
private String producerId; |
| ... | ... | |
| 103 | 106 |
} |
| 104 | 107 |
|
| 105 | 108 |
public String toJson() {
|
| 106 |
return new Gson().toJson(this); |
|
| 109 |
|
|
| 110 |
final Gson gson = new GsonBuilder() |
|
| 111 |
.setDateFormat(DATE_FORMAT).create(); |
|
| 112 |
|
|
| 113 |
return gson.toJson(this); |
|
| 107 | 114 |
} |
| 108 | 115 |
} |
| modules/dnet-mapreduce-jobs/branches/beta/src/main/java/eu/dnetlib/data/mapreduce/hbase/broker/model/EventMessage.java | ||
|---|---|---|
| 3 | 3 |
import java.util.Map; |
| 4 | 4 |
|
| 5 | 5 |
import com.google.gson.Gson; |
| 6 |
|
|
| 7 | 6 |
import eu.dnetlib.data.mapreduce.hbase.broker.Topic; |
| 8 | 7 |
|
| 9 | 8 |
/** |
| modules/dnet-mapreduce-jobs/branches/beta/src/main/java/eu/dnetlib/data/mapreduce/hbase/broker/mapping/DateParser.java | ||
|---|---|---|
| 1 | 1 |
package eu.dnetlib.data.mapreduce.hbase.broker.mapping; |
| 2 | 2 |
|
| 3 | 3 |
import java.text.ParseException; |
| 4 |
import java.util.Date; |
|
| 5 | 4 |
|
| 6 | 5 |
import org.apache.commons.lang3.time.DateUtils; |
| 7 | 6 |
import org.apache.commons.logging.Log; |
| ... | ... | |
| 11 | 10 |
|
| 12 | 11 |
private static final Log log = LogFactory.getLog(DateParser.class); |
| 13 | 12 |
|
| 14 |
public static Date parse(final String s) {
|
|
| 13 |
public static Long parse(final String s) {
|
|
| 15 | 14 |
if (s == null) {
|
| 16 | 15 |
log.warn("Date is NULL");
|
| 17 | 16 |
return null; |
| 18 | 17 |
} |
| 19 | 18 |
try {
|
| 20 |
return DateUtils.parseDate(s, "yyyy-MM-dd", "yyyy-MM-dd HH:mm:ss"); |
|
| 19 |
return DateUtils.parseDate(s, "yyyy-MM-dd", "yyyy-MM-dd HH:mm:ss").getTime();
|
|
| 21 | 20 |
} catch (final ParseException e) {
|
| 22 | 21 |
log.warn("Invalid Date: " + s);
|
| 23 | 22 |
return null; |
| modules/dnet-mapreduce-jobs/branches/beta/pom.xml | ||
|---|---|---|
| 1 | 1 |
<?xml version="1.0" ?> |
| 2 |
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd"> |
|
| 2 |
<project xmlns="http://maven.apache.org/POM/4.0.0" |
|
| 3 |
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" |
|
| 4 |
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd"> |
|
| 3 | 5 |
<parent> |
| 4 | 6 |
<groupId>eu.dnetlib</groupId> |
| 5 | 7 |
<artifactId>dnet45-parent</artifactId> |
| 6 | 8 |
<version>1.0.0</version> |
| 7 |
<relativePath />
|
|
| 9 |
<relativePath/> |
|
| 8 | 10 |
</parent> |
| 9 | 11 |
<modelVersion>4.0.0</modelVersion> |
| 10 | 12 |
<groupId>eu.dnetlib</groupId> |
| ... | ... | |
| 148 | 150 |
<dependency> |
| 149 | 151 |
<groupId>eu.dnetlib</groupId> |
| 150 | 152 |
<artifactId>dnet-openaireplus-mapping-utils</artifactId> |
| 151 |
<version>[6.0.0,6.1.5-SNAPSHOT)</version>
|
|
| 153 |
<version>[6.0.0,7.0.0)</version>
|
|
| 152 | 154 |
</dependency> |
| 153 | 155 |
<dependency> |
| 154 | 156 |
<groupId>org.antlr</groupId> |
| ... | ... | |
| 190 | 192 |
<dependency> |
| 191 | 193 |
<groupId>org.elasticsearch</groupId> |
| 192 | 194 |
<artifactId>elasticsearch-hadoop-mr</artifactId> |
| 193 |
<version>2.4.5</version> |
|
| 195 |
<version>5.2.0</version> |
|
| 196 |
<exclusions> |
|
| 197 |
<exclusion> |
|
| 198 |
<groupId>tomcat</groupId> |
|
| 199 |
<artifactId>jasper-compiler</artifactId> |
|
| 200 |
</exclusion> |
|
| 201 |
<exclusion> |
|
| 202 |
<groupId>org.antlr</groupId> |
|
| 203 |
<artifactId>antlr-runtime</artifactId> |
|
| 204 |
</exclusion> |
|
| 205 |
<exclusion> |
|
| 206 |
<groupId>org.eclipse.jetty.aggregate</groupId> |
|
| 207 |
<artifactId>jetty-all</artifactId> |
|
| 208 |
</exclusion> |
|
| 209 |
<exclusion> |
|
| 210 |
<groupId>org.slf4j</groupId> |
|
| 211 |
<artifactId>slf4j-log4j12</artifactId> |
|
| 212 |
</exclusion> |
|
| 213 |
<exclusion> |
|
| 214 |
<groupId>org.glassfish.jersey.core</groupId> |
|
| 215 |
<artifactId>jersey-client</artifactId> |
|
| 216 |
</exclusion> |
|
| 217 |
<exclusion> |
|
| 218 |
<groupId>org.glassfish.jersey.core</groupId> |
|
| 219 |
<artifactId>jersey-common</artifactId> |
|
| 220 |
</exclusion> |
|
| 221 |
<exclusion> |
|
| 222 |
<groupId>org.glassfish.jersey.core</groupId> |
|
| 223 |
<artifactId>jersey-server</artifactId> |
|
| 224 |
</exclusion> |
|
| 225 |
<exclusion> |
|
| 226 |
<groupId>org.glassfish.jersey.containers</groupId> |
|
| 227 |
<artifactId>jersey-container-servlet</artifactId> |
|
| 228 |
</exclusion> |
|
| 229 |
<exclusion> |
|
| 230 |
<groupId>org.glassfish.jersey.containers</groupId> |
|
| 231 |
<artifactId>jersey-container-servlet-core</artifactId> |
|
| 232 |
</exclusion> |
|
| 233 |
|
|
| 234 |
<exclusion> |
|
| 235 |
<groupId>org.codehaus.groovy</groupId> |
|
| 236 |
<artifactId>groovy-all</artifactId> |
|
| 237 |
</exclusion> |
|
| 238 |
|
|
| 239 |
<exclusion> |
|
| 240 |
<groupId>org.apache.hive</groupId> |
|
| 241 |
<artifactId>hive-service</artifactId> |
|
| 242 |
</exclusion> |
|
| 243 |
|
|
| 244 |
<exclusion> |
|
| 245 |
<groupId>org.apache.spark</groupId> |
|
| 246 |
<artifactId>spark-core_2.10</artifactId> |
|
| 247 |
</exclusion> |
|
| 248 |
<exclusion> |
|
| 249 |
<groupId>org.apache.spark</groupId> |
|
| 250 |
<artifactId>spark-sql_2.10</artifactId> |
|
| 251 |
</exclusion> |
|
| 252 |
<exclusion> |
|
| 253 |
<groupId>org.apache.spark</groupId> |
|
| 254 |
<artifactId>spark-streaming_2.10</artifactId> |
|
| 255 |
</exclusion> |
|
| 256 |
<exclusion> |
|
| 257 |
<groupId>cascading</groupId> |
|
| 258 |
<artifactId>cascading-hadoop</artifactId> |
|
| 259 |
</exclusion> |
|
| 260 |
<exclusion> |
|
| 261 |
<groupId>cascading</groupId> |
|
| 262 |
<artifactId>cascading-local</artifactId> |
|
| 263 |
</exclusion> |
|
| 264 |
<exclusion> |
|
| 265 |
<groupId>org.apache.storm</groupId> |
|
| 266 |
<artifactId>storm-core</artifactId> |
|
| 267 |
</exclusion> |
|
| 268 |
<exclusion> |
|
| 269 |
<groupId>org.apache.pig</groupId> |
|
| 270 |
<artifactId>pig</artifactId> |
|
| 271 |
</exclusion> |
|
| 272 |
</exclusions> |
|
| 273 |
|
|
| 194 | 274 |
</dependency> |
| 195 | 275 |
<dependency> |
| 196 | 276 |
<groupId>org.mockito</groupId> |
Also available in: Unified diff
force gson to serialise dates in a format that can be undrestood by ElasticSearch, updated elasticsearch-hadoop-mr lib to version 5.2.0