2022-04-20 14:02:05 +02:00
|
|
|
|
|
|
|
package eu.dnetlib.dhp.actionmanager.usagestats;
|
|
|
|
|
|
|
|
import static eu.dnetlib.dhp.actionmanager.Constants.*;
|
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkHiveSession;
|
|
|
|
|
|
|
|
import java.io.Serializable;
|
|
|
|
import java.util.Arrays;
|
|
|
|
import java.util.List;
|
|
|
|
import java.util.Optional;
|
|
|
|
|
2023-02-09 18:59:45 +01:00
|
|
|
import eu.dnetlib.dhp.schema.common.MainEntityType;
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.*;
|
2022-04-20 14:02:05 +02:00
|
|
|
import org.apache.commons.io.IOUtils;
|
2022-05-09 12:51:35 +02:00
|
|
|
import org.apache.hadoop.io.Text;
|
|
|
|
import org.apache.hadoop.mapred.SequenceFileOutputFormat;
|
2022-04-20 14:02:05 +02:00
|
|
|
import org.apache.spark.SparkConf;
|
|
|
|
import org.apache.spark.api.java.function.MapFunction;
|
|
|
|
import org.apache.spark.api.java.function.MapGroupsFunction;
|
|
|
|
import org.apache.spark.sql.Dataset;
|
|
|
|
import org.apache.spark.sql.Encoders;
|
|
|
|
import org.apache.spark.sql.SaveMode;
|
|
|
|
import org.apache.spark.sql.SparkSession;
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
|
|
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
|
|
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
|
|
import eu.dnetlib.dhp.common.HdfsSupport;
|
2022-05-09 14:43:27 +02:00
|
|
|
import eu.dnetlib.dhp.schema.action.AtomicAction;
|
2022-04-20 14:02:05 +02:00
|
|
|
import eu.dnetlib.dhp.schema.common.ModelConstants;
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
|
2022-05-09 12:51:35 +02:00
|
|
|
import scala.Tuple2;
|
2022-04-21 11:27:26 +02:00
|
|
|
|
2022-04-20 14:02:05 +02:00
|
|
|
/**
|
2022-04-21 12:14:03 +02:00
|
|
|
* created the Atomic Action for each type of results
|
2022-04-20 14:02:05 +02:00
|
|
|
*/
|
|
|
|
public class SparkAtomicActionUsageJob implements Serializable {
|
|
|
|
|
|
|
|
private static final Logger log = LoggerFactory.getLogger(SparkAtomicActionUsageJob.class);
|
|
|
|
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
|
|
|
|
|
|
|
|
public static <I extends Result> void main(String[] args) throws Exception {
|
|
|
|
|
|
|
|
String jsonConfiguration = IOUtils
|
|
|
|
.toString(
|
|
|
|
SparkAtomicActionUsageJob.class
|
|
|
|
.getResourceAsStream(
|
|
|
|
"/eu/dnetlib/dhp/actionmanager/usagestats/input_actionset_parameter.json"));
|
|
|
|
|
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
|
|
|
|
|
|
|
|
parser.parseArgument(args);
|
|
|
|
|
|
|
|
Boolean isSparkSessionManaged = Optional
|
|
|
|
.ofNullable(parser.get("isSparkSessionManaged"))
|
|
|
|
.map(Boolean::valueOf)
|
|
|
|
.orElse(Boolean.TRUE);
|
|
|
|
|
|
|
|
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
|
|
|
|
|
|
|
final String outputPath = parser.get("outputPath");
|
|
|
|
log.info("outputPath {}: ", outputPath);
|
|
|
|
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
conf.set("hive.metastore.uris", parser.get("hive_metastore_uris"));
|
|
|
|
|
2022-04-21 12:09:37 +02:00
|
|
|
final String dbname = parser.get("usagestatsdb");
|
2022-04-20 14:02:05 +02:00
|
|
|
|
|
|
|
final String workingPath = parser.get("workingPath");
|
|
|
|
|
|
|
|
runWithSparkHiveSession(
|
|
|
|
conf,
|
|
|
|
isSparkSessionManaged,
|
|
|
|
spark -> {
|
|
|
|
removeOutputDir(spark, outputPath);
|
2023-02-10 15:50:56 +01:00
|
|
|
prepareData(dbname, spark, workingPath + "/usageDb", "usage_stats", "result_id");
|
|
|
|
prepareData(dbname, spark, workingPath + "/projectDb", "project_stats", "id");
|
|
|
|
prepareData(dbname, spark, workingPath + "/datasourceDb", "datasource_stats", "repositor_id");
|
2022-05-09 12:51:35 +02:00
|
|
|
writeActionSet(spark, workingPath, outputPath);
|
2022-04-20 14:02:05 +02:00
|
|
|
});
|
|
|
|
}
|
|
|
|
|
2023-02-10 15:50:56 +01:00
|
|
|
private static void prepareData(String dbname, SparkSession spark, String workingPath, String tableName, String attribute_name) {
|
2022-04-20 14:02:05 +02:00
|
|
|
spark
|
2023-02-09 18:59:45 +01:00
|
|
|
.sql(
|
2023-02-10 15:50:56 +01:00
|
|
|
"Select " + attribute_name + " as id, downloads, views " +
|
2023-02-09 18:59:45 +01:00
|
|
|
"from " + dbname + "." + tableName)
|
|
|
|
.as(Encoders.bean(UsageStatsModel.class))
|
|
|
|
.write()
|
|
|
|
.mode(SaveMode.Overwrite)
|
|
|
|
.option("compression", "gzip")
|
|
|
|
.json(workingPath);
|
2022-04-20 14:02:05 +02:00
|
|
|
}
|
|
|
|
|
2023-02-09 18:59:45 +01:00
|
|
|
|
|
|
|
|
2022-05-09 12:51:35 +02:00
|
|
|
public static void writeActionSet(SparkSession spark, String inputPath, String outputPath) {
|
2023-02-09 18:59:45 +01:00
|
|
|
getFinalIndicatorsResult(spark, inputPath+ "/usageDb").
|
|
|
|
toJavaRDD().
|
|
|
|
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)))
|
2022-05-09 14:43:27 +02:00
|
|
|
.mapToPair(
|
|
|
|
aa -> new Tuple2<>(new Text(aa.getClazz().getCanonicalName()),
|
|
|
|
new Text(OBJECT_MAPPER.writeValueAsString(aa))))
|
|
|
|
.saveAsHadoopFile(outputPath, Text.class, Text.class, SequenceFileOutputFormat.class);
|
2022-05-09 12:51:35 +02:00
|
|
|
|
2022-04-20 14:02:05 +02:00
|
|
|
}
|
|
|
|
|
2023-02-09 18:59:45 +01:00
|
|
|
private static Dataset<Result> getFinalIndicatorsResult(SparkSession spark, String inputPath) {
|
|
|
|
|
|
|
|
return getUsageStatsModelDataset(spark, inputPath)
|
|
|
|
.map((MapFunction<UsageStatsModel, Result>) usm -> {
|
|
|
|
Result r = new Result();
|
|
|
|
r.setId("50|" + usm.getId());
|
|
|
|
r.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
|
|
|
|
return r;
|
|
|
|
}, Encoders.bean(Result.class));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Dataset<Project> getFinalIndicatorsProject(SparkSession spark, String inputPath) {
|
|
|
|
|
|
|
|
return getUsageStatsModelDataset(spark, inputPath)
|
|
|
|
.map((MapFunction<UsageStatsModel, Project>) usm -> {
|
|
|
|
Project r = new Project();
|
|
|
|
r.setId("40|" + usm.getId());
|
|
|
|
r.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
|
|
|
|
return r;
|
|
|
|
}, Encoders.bean(Project.class));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Dataset<Datasource> getFinalIndicatorsDatasource(SparkSession spark, String inputPath) {
|
|
|
|
|
|
|
|
return getUsageStatsModelDataset(spark, inputPath)
|
|
|
|
.map((MapFunction<UsageStatsModel, Datasource>) usm -> {
|
|
|
|
Datasource r = new Datasource();
|
|
|
|
r.setId("10|" + usm.getId());
|
|
|
|
r.setMeasures(getMeasure(usm.getDownloads(), usm.getViews()));
|
|
|
|
return r;
|
|
|
|
}, Encoders.bean(Datasource.class));
|
|
|
|
}
|
|
|
|
|
|
|
|
private static Dataset<UsageStatsModel> getUsageStatsModelDataset(SparkSession spark, String inputPath) {
|
|
|
|
return readPath(spark, inputPath, UsageStatsModel.class)
|
|
|
|
.groupByKey((MapFunction<UsageStatsModel, String>) us -> us.getId(), Encoders.STRING())
|
|
|
|
.mapGroups((MapGroupsFunction<String, UsageStatsModel, UsageStatsModel>) (k, it) -> {
|
|
|
|
UsageStatsModel first = it.next();
|
|
|
|
it.forEachRemaining(us -> {
|
|
|
|
first.setDownloads(first.getDownloads() + us.getDownloads());
|
|
|
|
first.setViews(first.getViews() + us.getViews());
|
|
|
|
});
|
|
|
|
first.setId(k);
|
|
|
|
return first;
|
|
|
|
|
|
|
|
}, Encoders.bean(UsageStatsModel.class));
|
|
|
|
}
|
|
|
|
|
2022-04-20 14:02:05 +02:00
|
|
|
private static List<Measure> getMeasure(Long downloads, Long views) {
|
|
|
|
DataInfo dataInfo = OafMapperUtils
|
|
|
|
.dataInfo(
|
|
|
|
false,
|
|
|
|
UPDATE_DATA_INFO_TYPE,
|
|
|
|
true,
|
|
|
|
false,
|
|
|
|
OafMapperUtils
|
|
|
|
.qualifier(
|
|
|
|
UPDATE_MEASURE_USAGE_COUNTS_CLASS_ID,
|
|
|
|
UPDATE_CLASS_NAME,
|
|
|
|
ModelConstants.DNET_PROVENANCE_ACTIONS,
|
|
|
|
ModelConstants.DNET_PROVENANCE_ACTIONS),
|
|
|
|
"");
|
|
|
|
|
|
|
|
return Arrays
|
|
|
|
.asList(
|
2022-04-21 12:09:37 +02:00
|
|
|
OafMapperUtils
|
|
|
|
.newMeasureInstance("downloads", String.valueOf(downloads), UPDATE_KEY_USAGE_COUNTS, dataInfo),
|
|
|
|
OafMapperUtils.newMeasureInstance("views", String.valueOf(views), UPDATE_KEY_USAGE_COUNTS, dataInfo));
|
2022-04-20 14:02:05 +02:00
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
private static void removeOutputDir(SparkSession spark, String path) {
|
|
|
|
HdfsSupport.remove(path, spark.sparkContext().hadoopConfiguration());
|
|
|
|
}
|
|
|
|
|
|
|
|
public static <R> Dataset<R> readPath(
|
|
|
|
SparkSession spark, String inputPath, Class<R> clazz) {
|
|
|
|
return spark
|
|
|
|
.read()
|
|
|
|
.textFile(inputPath)
|
|
|
|
.map((MapFunction<String, R>) value -> OBJECT_MAPPER.readValue(value, clazz), Encoders.bean(clazz));
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|