D-Net
/
dnet-hadoop
13
4
Fork 10
Code Issues 14 Pull Requests 6 Projects 2 Releases 1 Wiki Activity

master #59

Closed
claudio.atzori wants to merge 3221 commits from master into stable_ids
pull from: master
merge into: D-Net:stable_ids
Conversation 0 Commits 3221 Files Changed 2141 +253 -251
2 changed files with 253 additions and 251 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 0c1be41b30 - Show all commits

48
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/usagestats/SparkAtomicActionUsageJob.java
View File

@ -8,13 +8,12 @@ import java.io.Serializable;
import java.util.Arrays;
import java.util.List;
import java.util.Optional;
import eu.dnetlib.dhp.schema.oaf.*;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.SequenceFileOutputFormat;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SaveMode;
@ -28,6 +27,7 @@ import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
import scala.Tuple2;
@ -80,12 +80,12 @@ public class SparkAtomicActionUsageJob implements Serializable {
});
}
private static void prepareData(String dbname, SparkSession spark, String workingPath, String tableName, String attribute_name) {
private static void prepareData(String dbname, SparkSession spark, String workingPath, String tableName,
String attribute_name) {
spark
.sql(
"Select " + attribute_name + " as id, sum(downloads) as downloads, sum(views) as views " +
"from " + dbname + "." + tableName +
"group by " + attribute_name)
.sql(String.format(
"select %s as id, sum(downloads) as downloads, sum(views) as views " +
"from %s.%s group by %s", attribute_name, dbname, tableName, attribute_name))
.as(Encoders.bean(UsageStatsModel.class))
.write()
.mode(SaveMode.Overwrite)
@ -93,16 +93,16 @@ public class SparkAtomicActionUsageJob implements Serializable {
.json(workingPath);
}
public static void writeActionSet(SparkSession spark, String inputPath, String outputPath) {
getFinalIndicatorsResult(spark, inputPath+ "/usageDb").
toJavaRDD().
map(p -> new AtomicAction(p.getClass(),p))
.union(getFinalIndicatorsProject(spark, inputPath + "/projectDb")
getFinalIndicatorsResult(spark, inputPath + "/usageDb")
.toJavaRDD()
.map(p -> new AtomicAction(p.getClass(), p )))
.union(getFinalIndicatorsDatasource(spark, inputPath + "/datasourceDb")
.map(p -> new AtomicAction(p.getClass(), p))
.union(
getFinalIndicatorsProject(spark, inputPath + "/projectDb")
.toJavaRDD()
.map(p -> new AtomicAction(p.getClass(), p)))
.union(
getFinalIndicatorsDatasource(spark, inputPath + "/datasourceDb")
.toJavaRDD()
.map(p -> new AtomicAction(p.getClass(), p)))
.mapToPair(
@ -127,10 +127,10 @@ public class SparkAtomicActionUsageJob implements Serializable {
return readPath(spark, inputPath, UsageStatsModel.class)
.map((MapFunction<UsageStatsModel, Project>) usm -> {
Project r = new Project();
r.setId("40|" + usm.getId());
r.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
return r;
Project p = new Project();
p.setId("40|" + usm.getId());
p.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
return p;
}, Encoders.bean(Project.class));
}
@ -138,15 +138,13 @@ public class SparkAtomicActionUsageJob implements Serializable {
return readPath(spark, inputPath, UsageStatsModel.class)
.map((MapFunction<UsageStatsModel, Datasource>) usm -> {
Datasource r = new Datasource();
r.setId("10|" + usm.getId());
r.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
return r;
Datasource d = new Datasource();
d.setId("10|" + usm.getId());
d.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
return d;
}, Encoders.bean(Datasource.class));
}
private static List<Measure> getMeasure(Long downloads, Long views) {
DataInfo dataInfo = OafMapperUtils
.dataInfo(

70
dhp-workflows/dhp-aggregation/src/test/java/eu/dnetlib/dhp/actionmanager/usagestats/SparkAtomicActionCountJobTest.java
View File

@ -8,7 +8,6 @@ import java.nio.file.Files;
import java.nio.file.Path;
import java.util.stream.Collectors;
import eu.dnetlib.dhp.schema.oaf.OafEntity;
import org.apache.commons.io.FileUtils;
import org.apache.hadoop.io.Text;
import org.apache.spark.SparkConf;
@ -25,6 +24,7 @@ import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.dhp.schema.oaf.OafEntity;
import eu.dnetlib.dhp.schema.oaf.Result;
public class SparkAtomicActionCountJobTest {
@ -79,16 +79,16 @@ public class SparkAtomicActionCountJobTest {
JavaRDD<AtomicAction> tmp = sc
.sequenceFile(workingDir.toString() + "/actionSet", Text.class, Text.class)
.map(usm -> OBJECT_MAPPER.readValue(usm._2.getBytes(), AtomicAction.class));
//.map(aa -> (Result) aa.getPayload());
// .map(aa -> (Result) aa.getPayload());
Assertions.assertEquals(9,tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("50|")).count());
Assertions.assertEquals(9,tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("10|")).count());
Assertions.assertEquals(9,tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("40|")).count());
Assertions.assertEquals(9, tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("50|")).count());
Assertions.assertEquals(9, tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("10|")).count());
Assertions.assertEquals(9, tmp.filter(aa -> ((OafEntity) aa.getPayload()).getId().startsWith("40|")).count());
tmp.foreach(r -> Assertions.assertEquals(2, ((OafEntity)r.getPayload()).getMeasures().size()));
tmp.foreach(r -> Assertions.assertEquals(2, ((OafEntity) r.getPayload()).getMeasures().size()));
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
@ -98,14 +98,14 @@ public class SparkAtomicActionCountJobTest {
.forEach(u -> Assertions.assertFalse(u.getDataInfo().getDeletedbyinference()))));
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
m -> m.getUnit().stream().forEach(u -> Assertions.assertTrue(u.getDataInfo().getInferred()))));
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
@ -116,7 +116,7 @@ public class SparkAtomicActionCountJobTest {
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
@ -130,7 +130,7 @@ public class SparkAtomicActionCountJobTest {
u.getDataInfo().getProvenanceaction().getClassid()))));
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
@ -145,7 +145,7 @@ public class SparkAtomicActionCountJobTest {
tmp
.foreach(
r -> ((OafEntity)r.getPayload())
r -> ((OafEntity) r.getPayload())
.getMeasures()
.stream()
.forEach(
@ -160,13 +160,19 @@ public class SparkAtomicActionCountJobTest {
Assertions
.assertEquals(
1, tmp.filter(r -> ((OafEntity)r.getPayload()).getId().equals("50|dedup_wf_001::53575dc69e9ace947e02d47ecd54a7a6")).count());
1,
tmp
.filter(
r -> ((OafEntity) r.getPayload())
.getId()
.equals("50|dedup_wf_001::53575dc69e9ace947e02d47ecd54a7a6"))
.count());
Assertions
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|dedup_wf_001::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -182,7 +188,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"5",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|dedup_wf_001::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -199,7 +205,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|doi_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -215,7 +221,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"1",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|doi_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -232,7 +238,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"2",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|doi_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)
@ -248,7 +254,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"6",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("50|doi_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)
@ -261,12 +267,11 @@ public class SparkAtomicActionCountJobTest {
.get(0)
.getValue());
Assertions
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f1__________::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -282,7 +287,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"5",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f1__________::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -299,7 +304,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f11_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -315,7 +320,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"1",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f11_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -332,7 +337,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"2",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f12_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)
@ -348,7 +353,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"6",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("40|f12_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)
@ -361,12 +366,11 @@ public class SparkAtomicActionCountJobTest {
.get(0)
.getValue());
Assertions
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d1__________::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -382,7 +386,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"5",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d1__________::53575dc69e9ace947e02d47ecd54a7a6"))
.collect()
.get(0)
@ -399,7 +403,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"0",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d11_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -415,7 +419,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"1",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d11_________::17eda2ff77407538fbe5d3d719b9d1c0"))
.collect()
.get(0)
@ -432,7 +436,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"2",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d12_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)
@ -448,7 +452,7 @@ public class SparkAtomicActionCountJobTest {
.assertEquals(
"6",
tmp
.map(r -> ((OafEntity)r.getPayload()))
.map(r -> ((OafEntity) r.getPayload()))
.filter(r -> r.getId().equals("10|d12_________::3085e4c6e051378ca6157fe7f0430c1f"))
.collect()
.get(0)