2020-06-15 11:12:14 +02:00
|
|
|
|
2020-07-29 16:52:44 +02:00
|
|
|
package eu.dnetlib.dhp.oa.graph.dump.community;
|
2020-06-09 15:37:46 +02:00
|
|
|
|
2020-06-15 11:12:14 +02:00
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
|
|
|
|
|
|
|
import java.io.Serializable;
|
|
|
|
import java.util.Optional;
|
2020-06-09 15:37:46 +02:00
|
|
|
|
|
|
|
import org.apache.commons.io.IOUtils;
|
|
|
|
import org.apache.spark.SparkConf;
|
2020-11-24 14:40:22 +01:00
|
|
|
import org.apache.spark.api.java.function.MapFunction;
|
2020-06-09 15:37:46 +02:00
|
|
|
import org.apache.spark.sql.Dataset;
|
|
|
|
import org.apache.spark.sql.Encoders;
|
2020-06-09 15:39:03 +02:00
|
|
|
import org.apache.spark.sql.SaveMode;
|
2020-06-09 15:37:46 +02:00
|
|
|
import org.apache.spark.sql.SparkSession;
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
|
2020-06-15 11:12:14 +02:00
|
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
2020-06-09 15:37:46 +02:00
|
|
|
|
2020-06-15 11:12:14 +02:00
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
2020-07-29 16:52:44 +02:00
|
|
|
import eu.dnetlib.dhp.oa.graph.dump.Utils;
|
|
|
|
import eu.dnetlib.dhp.schema.dump.oaf.community.CommunityResult;
|
2020-11-24 14:40:22 +01:00
|
|
|
import scala.Tuple2;
|
2020-06-09 15:37:46 +02:00
|
|
|
|
2020-06-09 15:39:03 +02:00
|
|
|
public class SparkUpdateProjectInfo implements Serializable {
|
2020-06-09 15:37:46 +02:00
|
|
|
|
2020-06-15 11:12:14 +02:00
|
|
|
private static final Logger log = LoggerFactory.getLogger(SparkUpdateProjectInfo.class);
|
|
|
|
public static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
|
|
|
|
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
|
|
String jsonConfiguration = IOUtils
|
|
|
|
.toString(
|
|
|
|
SparkUpdateProjectInfo.class
|
|
|
|
.getResourceAsStream(
|
|
|
|
"/eu/dnetlib/dhp/oa/graph/dump/project_input_parameters.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 inputPath = parser.get("sourcePath");
|
|
|
|
log.info("inputPath: {}", inputPath);
|
|
|
|
|
|
|
|
final String outputPath = parser.get("outputPath");
|
|
|
|
log.info("outputPath: {}", outputPath);
|
|
|
|
|
|
|
|
final String preparedInfoPath = parser.get("preparedInfoPath");
|
|
|
|
log.info("preparedInfoPath: {}", preparedInfoPath);
|
|
|
|
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
|
|
|
|
runWithSparkSession(
|
|
|
|
conf,
|
|
|
|
isSparkSessionManaged,
|
|
|
|
spark -> {
|
|
|
|
Utils.removeOutputDir(spark, outputPath);
|
2021-08-11 12:13:22 +02:00
|
|
|
extend(spark, inputPath, outputPath, preparedInfoPath);
|
2020-06-15 11:12:14 +02:00
|
|
|
});
|
|
|
|
}
|
|
|
|
|
2020-06-18 11:22:58 +02:00
|
|
|
private static void extend(
|
2020-06-15 11:12:14 +02:00
|
|
|
SparkSession spark,
|
|
|
|
String inputPath,
|
|
|
|
String outputPath,
|
2020-07-29 16:52:44 +02:00
|
|
|
String preparedInfoPath) {
|
|
|
|
Dataset<CommunityResult> result = Utils.readPath(spark, inputPath, CommunityResult.class);
|
2020-06-15 11:12:14 +02:00
|
|
|
Dataset<ResultProject> resultProject = Utils.readPath(spark, preparedInfoPath, ResultProject.class);
|
|
|
|
result
|
|
|
|
.joinWith(
|
|
|
|
resultProject, result.col("id").equalTo(resultProject.col("resultId")),
|
|
|
|
"left")
|
2020-11-24 14:40:22 +01:00
|
|
|
.map((MapFunction<Tuple2<CommunityResult, ResultProject>, CommunityResult>) value -> {
|
2020-07-29 16:52:44 +02:00
|
|
|
CommunityResult r = value._1();
|
2021-08-11 12:13:22 +02:00
|
|
|
Optional.ofNullable(value._2()).ifPresent(rp -> r.setProjects(rp.getProjectsList()));
|
2020-06-15 11:12:14 +02:00
|
|
|
return r;
|
2020-07-29 16:52:44 +02:00
|
|
|
}, Encoders.bean(CommunityResult.class))
|
2020-06-15 11:12:14 +02:00
|
|
|
.write()
|
|
|
|
.option("compression", "gzip")
|
2020-06-18 11:22:58 +02:00
|
|
|
.mode(SaveMode.Append)
|
2020-06-15 11:12:14 +02:00
|
|
|
.json(outputPath);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|