forked from D-Net/dnet-hadoop
removed unuseful classes
This commit is contained in:
parent
964b22d418
commit
3577219127
|
|
@ -1,333 +0,0 @@
|
||||||
package eu.dnetlib.dhp.countrypropagation;
|
|
||||||
|
|
||||||
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
||||||
import eu.dnetlib.dhp.PropagationConstant;
|
|
||||||
import eu.dnetlib.dhp.TypedRow;
|
|
||||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
||||||
import eu.dnetlib.dhp.schema.oaf.*;
|
|
||||||
import net.sf.saxon.expr.ContextMappingFunction;
|
|
||||||
import net.sf.saxon.expr.flwor.Tuple;
|
|
||||||
import net.sf.saxon.om.Item;
|
|
||||||
import org.apache.commons.io.IOUtils;
|
|
||||||
import org.apache.hadoop.fs.FileSystem;
|
|
||||||
import org.apache.hadoop.fs.Path;
|
|
||||||
import org.apache.hadoop.io.Text;
|
|
||||||
import org.apache.spark.SparkConf;
|
|
||||||
import org.apache.spark.api.java.JavaPairRDD;
|
|
||||||
import org.apache.spark.api.java.JavaRDD;
|
|
||||||
import org.apache.spark.api.java.JavaSparkContext;
|
|
||||||
import org.apache.spark.sql.*;
|
|
||||||
import org.apache.spark.sql.Dataset;
|
|
||||||
import org.codehaus.janino.Java;
|
|
||||||
import scala.Tuple2;
|
|
||||||
|
|
||||||
import javax.sql.DataSource;
|
|
||||||
import java.beans.Encoder;
|
|
||||||
import java.io.File;
|
|
||||||
import java.io.IOException;
|
|
||||||
import java.util.*;
|
|
||||||
|
|
||||||
import static eu.dnetlib.dhp.PropagationConstant.*;
|
|
||||||
|
|
||||||
public class SparkCountryPropagationJob {
|
|
||||||
|
|
||||||
public static void main(String[] args) throws Exception {
|
|
||||||
|
|
||||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(IOUtils.toString(SparkCountryPropagationJob.class.getResourceAsStream("/eu/dnetlib/dhp/countrypropagation/input_countrypropagation_parameters.json")));
|
|
||||||
parser.parseArgument(args);
|
|
||||||
SparkConf conf = new SparkConf();
|
|
||||||
conf.set("hive.metastore.uris", parser.get("hive_metastore_uris"));
|
|
||||||
final SparkSession spark = SparkSession
|
|
||||||
.builder()
|
|
||||||
.appName(SparkCountryPropagationJob.class.getSimpleName())
|
|
||||||
.master(parser.get("master"))
|
|
||||||
.config(conf)
|
|
||||||
.enableHiveSupport()
|
|
||||||
.getOrCreate();
|
|
||||||
|
|
||||||
|
|
||||||
final String inputPath = parser.get("sourcePath");
|
|
||||||
final String outputPath = "/tmp/provision/propagation/countrytoresultfrominstitutionalrepositories";
|
|
||||||
|
|
||||||
createOutputDirs(outputPath, FileSystem.get(spark.sparkContext().hadoopConfiguration()));
|
|
||||||
|
|
||||||
List<String> whitelist = Arrays.asList(parser.get("whitelist").split(";"));
|
|
||||||
List<String> allowedtypes = Arrays.asList(parser.get("allowedtypes").split(";"));
|
|
||||||
boolean writeUpdates = TRUE.equals(parser.get("writeUpdate"));
|
|
||||||
boolean saveGraph = TRUE.equals(parser.get("saveGraph"));
|
|
||||||
|
|
||||||
// datasource(spark, whitelist, outputPath, inputPath, "true".equals(parser.get("writeUpdate")),
|
|
||||||
// "true".equals(parser.get("saveGraph")), allowedtypes);
|
|
||||||
//
|
|
||||||
// }
|
|
||||||
//
|
|
||||||
//
|
|
||||||
// private static void datasource(SparkSession spark, List<String> whitelist, String outputPath, String inputPath,
|
|
||||||
// boolean writeUpdates, boolean saveGraph, List<String> allowedtypes){
|
|
||||||
|
|
||||||
String whitelisted = "";
|
|
||||||
for (String i : whitelist){
|
|
||||||
whitelisted += " OR id = '" + i + "'";
|
|
||||||
}
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
|
|
||||||
|
|
||||||
Dataset<Datasource> datasource = spark.createDataset(sc.textFile(inputPath + "/datasource")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Datasource.class)).rdd(), Encoders.bean(Datasource.class));
|
|
||||||
|
|
||||||
Dataset<Relation> relation = spark.createDataset(sc.textFile(inputPath + "/relation")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Relation.class)).rdd(), Encoders.bean(Relation.class));
|
|
||||||
|
|
||||||
Dataset<Organization> organization = spark.createDataset(sc.textFile(inputPath + "/organization")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Organization.class)).rdd(), Encoders.bean(Organization.class));
|
|
||||||
|
|
||||||
Dataset<eu.dnetlib.dhp.schema.oaf.Dataset> dataset = spark.createDataset(sc.textFile(inputPath + "/dataset")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, eu.dnetlib.dhp.schema.oaf.Dataset.class)).rdd(),
|
|
||||||
Encoders.bean(eu.dnetlib.dhp.schema.oaf.Dataset.class));
|
|
||||||
|
|
||||||
Dataset<OtherResearchProduct> other = spark.createDataset(sc.textFile(inputPath + "/otherresearchproduct")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, eu.dnetlib.dhp.schema.oaf.OtherResearchProduct.class)).rdd(),
|
|
||||||
Encoders.bean(eu.dnetlib.dhp.schema.oaf.OtherResearchProduct.class));
|
|
||||||
|
|
||||||
Dataset<Software> software = spark.createDataset(sc.textFile(inputPath + "/software")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, eu.dnetlib.dhp.schema.oaf.Software.class)).rdd(),
|
|
||||||
Encoders.bean(eu.dnetlib.dhp.schema.oaf.Software.class));
|
|
||||||
|
|
||||||
Dataset<Publication> publication = spark.createDataset(sc.textFile(inputPath + "/publication")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, eu.dnetlib.dhp.schema.oaf.Publication.class)).rdd(),
|
|
||||||
Encoders.bean(eu.dnetlib.dhp.schema.oaf.Publication.class));
|
|
||||||
|
|
||||||
datasource.createOrReplaceTempView("datasource");
|
|
||||||
relation.createOrReplaceTempView("relation");
|
|
||||||
organization.createOrReplaceTempView("organization");
|
|
||||||
//todo add length(country.classid)>0
|
|
||||||
String query = "SELECT source ds, target org, country.classid country " +
|
|
||||||
"FROM ( SELECT id " +
|
|
||||||
"FROM datasource " +
|
|
||||||
"WHERE (datainfo.deletedbyinference = false " + whitelisted + ") " +
|
|
||||||
getConstraintList("datasourcetype.classid = '", allowedtypes) +
|
|
||||||
// "datasourcetype.classid = 'pubsrepository::institutional' " +
|
|
||||||
// "AND (datainfo.deletedbyinference = false " + whitelisted + ") ) d " +
|
|
||||||
"JOIN ( SELECT source, target " +
|
|
||||||
"FROM relation " +
|
|
||||||
"WHERE relclass = 'provides' " +
|
|
||||||
"AND datainfo.deletedbyinference = false ) rel " +
|
|
||||||
"ON d.id = rel.source " +
|
|
||||||
"JOIN (SELECT id, country " +
|
|
||||||
"FROM organization " +
|
|
||||||
"WHERE datainfo.deletedbyinference = false ) o " +
|
|
||||||
"ON o.id = rel.target";
|
|
||||||
|
|
||||||
//todo broadcast
|
|
||||||
Dataset<Row> rels = spark.sql(query);
|
|
||||||
rels.createOrReplaceTempView("rels");
|
|
||||||
|
|
||||||
software.createOrReplaceTempView("software");
|
|
||||||
final JavaRDD<Row> toupdateresultsoftware = propagateOnResult(spark, "software");
|
|
||||||
|
|
||||||
dataset.createOrReplaceTempView("dataset");
|
|
||||||
final JavaRDD<Row> toupdateresultdataset = propagateOnResult(spark, "dataset");
|
|
||||||
|
|
||||||
other.createOrReplaceTempView("other");
|
|
||||||
final JavaRDD<Row> toupdateresultother = propagateOnResult(spark, "other");
|
|
||||||
|
|
||||||
publication.createOrReplaceTempView("publication");
|
|
||||||
final JavaRDD<Row> toupdateresultpublication = propagateOnResult(spark, "publication");
|
|
||||||
|
|
||||||
if(writeUpdates){
|
|
||||||
writeUpdates(toupdateresultsoftware, toupdateresultdataset, toupdateresultother, toupdateresultpublication, outputPath);
|
|
||||||
}
|
|
||||||
|
|
||||||
if(saveGraph){
|
|
||||||
createUpdateForSoftwareDataset(toupdateresultsoftware, inputPath, spark)
|
|
||||||
.map(s -> new ObjectMapper().writeValueAsString(s))
|
|
||||||
.saveAsTextFile(outputPath + "/software");
|
|
||||||
|
|
||||||
createUpdateForDatasetDataset(toupdateresultdataset,inputPath,spark)
|
|
||||||
.map(d -> new ObjectMapper().writeValueAsString(d))
|
|
||||||
.saveAsTextFile(outputPath + "/dataset");
|
|
||||||
|
|
||||||
createUpdateForOtherDataset(toupdateresultother, inputPath, spark)
|
|
||||||
.map(o -> new ObjectMapper().writeValueAsString(o))
|
|
||||||
.saveAsTextFile(outputPath + "/otherresearchproduct");
|
|
||||||
|
|
||||||
createUpdateForPublicationDataset(toupdateresultpublication, inputPath, spark)
|
|
||||||
.map(p -> new ObjectMapper().writeValueAsString(p))
|
|
||||||
.saveAsTextFile(outputPath + "/publication");
|
|
||||||
}
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
||||||
private static void writeUpdates(JavaRDD<Row> software, JavaRDD<Row> dataset, JavaRDD<Row> other , JavaRDD<Row> publication, String outputPath){
|
|
||||||
createUpdateForResultDatasetWrite(software, outputPath, "update_software");
|
|
||||||
createUpdateForResultDatasetWrite(dataset, outputPath, "update_dataset");
|
|
||||||
createUpdateForResultDatasetWrite(other, outputPath, "update_other");
|
|
||||||
createUpdateForResultDatasetWrite(publication, outputPath, "update_publication");
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaRDD<OtherResearchProduct> createUpdateForOtherDataset(JavaRDD<Row> toupdateresult, String inputPath, SparkSession spark) {
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
|
|
||||||
return sc.textFile(inputPath + "/otherresearchproduct")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, OtherResearchProduct.class))
|
|
||||||
.mapToPair(s -> new Tuple2<>(s.getId(), s)).leftOuterJoin(getStringResultJavaPairRDD(toupdateresult))
|
|
||||||
.map(c -> {
|
|
||||||
OtherResearchProduct oaf = c._2()._1();
|
|
||||||
List<Country> countryList = oaf.getCountry();
|
|
||||||
if (c._2()._2().isPresent()) {
|
|
||||||
HashSet<String> countries = new HashSet<>();
|
|
||||||
for (Qualifier country : countryList) {
|
|
||||||
countries.add(country.getClassid());
|
|
||||||
}
|
|
||||||
Result r = c._2()._2().get();
|
|
||||||
for (Country country : r.getCountry()) {
|
|
||||||
if (!countries.contains(country.getClassid())) {
|
|
||||||
countryList.add(country);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
oaf.setCountry(countryList);
|
|
||||||
}
|
|
||||||
return oaf;
|
|
||||||
});
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaRDD<Publication> createUpdateForPublicationDataset(JavaRDD<Row> toupdateresult, String inputPath, SparkSession spark) {
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
|
|
||||||
return sc.textFile(inputPath + "/publication")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Publication.class))
|
|
||||||
.mapToPair(s -> new Tuple2<>(s.getId(), s)).leftOuterJoin(getStringResultJavaPairRDD(toupdateresult))
|
|
||||||
.map(c -> {
|
|
||||||
Publication oaf = c._2()._1();
|
|
||||||
List<Country> countryList = oaf.getCountry();
|
|
||||||
if (c._2()._2().isPresent()) {
|
|
||||||
HashSet<String> countries = new HashSet<>();
|
|
||||||
for (Qualifier country : countryList) {
|
|
||||||
countries.add(country.getClassid());
|
|
||||||
}
|
|
||||||
Result r = c._2()._2().get();
|
|
||||||
for (Country country : r.getCountry()) {
|
|
||||||
if (!countries.contains(country.getClassid())) {
|
|
||||||
countryList.add(country);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
oaf.setCountry(countryList);
|
|
||||||
}
|
|
||||||
return oaf;
|
|
||||||
});
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaRDD<Software> createUpdateForSoftwareDataset(JavaRDD<Row> toupdateresult, String inputPath, SparkSession spark) {
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
|
|
||||||
return sc.textFile(inputPath + "/software")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Software.class))
|
|
||||||
.mapToPair(s -> new Tuple2<>(s.getId(), s)).leftOuterJoin(getStringResultJavaPairRDD(toupdateresult))
|
|
||||||
.map(c -> {
|
|
||||||
Software oaf = c._2()._1();
|
|
||||||
List<Country> countryList = oaf.getCountry();
|
|
||||||
if (c._2()._2().isPresent()) {
|
|
||||||
HashSet<String> countries = new HashSet<>();
|
|
||||||
for (Qualifier country : countryList) {
|
|
||||||
countries.add(country.getClassid());
|
|
||||||
}
|
|
||||||
Result r = c._2()._2().get();
|
|
||||||
for (Country country : r.getCountry()) {
|
|
||||||
if (!countries.contains(country.getClassid())) {
|
|
||||||
countryList.add(country);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
oaf.setCountry(countryList);
|
|
||||||
}
|
|
||||||
return oaf;
|
|
||||||
});
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaRDD<eu.dnetlib.dhp.schema.oaf.Dataset> createUpdateForDatasetDataset(JavaRDD<Row> toupdateresult, String inputPath, SparkSession spark) {
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
|
|
||||||
return sc.textFile(inputPath + "/dataset")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, eu.dnetlib.dhp.schema.oaf.Dataset.class))
|
|
||||||
.mapToPair(d -> new Tuple2<>(d.getId(), d)).leftOuterJoin(getStringResultJavaPairRDD(toupdateresult))
|
|
||||||
.map(c -> {
|
|
||||||
eu.dnetlib.dhp.schema.oaf.Dataset oaf = c._2()._1();
|
|
||||||
List<Country> countryList = oaf.getCountry();
|
|
||||||
if (c._2()._2().isPresent()) {
|
|
||||||
HashSet<String> countries = new HashSet<>();
|
|
||||||
for (Qualifier country : countryList) {
|
|
||||||
countries.add(country.getClassid());
|
|
||||||
}
|
|
||||||
Result r = c._2()._2().get();
|
|
||||||
for (Country country : r.getCountry()) {
|
|
||||||
if (!countries.contains(country.getClassid())) {
|
|
||||||
countryList.add(country);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
oaf.setCountry(countryList);
|
|
||||||
}
|
|
||||||
return oaf;
|
|
||||||
});
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaRDD<Row> propagateOnResult(SparkSession spark, String result_type) {
|
|
||||||
String query;
|
|
||||||
query = "SELECT id, inst.collectedfrom.key cf , inst.hostedby.key hb " +
|
|
||||||
"FROM ( SELECT id, instance " +
|
|
||||||
"FROM " + result_type +
|
|
||||||
" WHERE datainfo.deletedbyinference = false) ds " +
|
|
||||||
"LATERAL VIEW EXPLODE(instance) i AS inst";
|
|
||||||
Dataset<Row> cfhb = spark.sql(query);
|
|
||||||
cfhb.createOrReplaceTempView("cfhb");
|
|
||||||
|
|
||||||
return countryPropagationAssoc(spark, "cfhb").toJavaRDD();
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
||||||
private static Dataset<Row> countryPropagationAssoc(SparkSession spark, String cfhbTable){
|
|
||||||
String query = "SELECT id, collect_set(country) country "+
|
|
||||||
"FROM ( SELECT id, country " +
|
|
||||||
"FROM rels " +
|
|
||||||
"JOIN " + cfhbTable +
|
|
||||||
" ON cf = ds " +
|
|
||||||
"UNION ALL " +
|
|
||||||
"SELECT id , country " +
|
|
||||||
"FROM rels " +
|
|
||||||
"JOIN " + cfhbTable +
|
|
||||||
" ON hb = ds ) tmp " +
|
|
||||||
"GROUP BY id";
|
|
||||||
return spark.sql(query);
|
|
||||||
}
|
|
||||||
|
|
||||||
private static JavaPairRDD<String, Result> getStringResultJavaPairRDD(JavaRDD<Row> toupdateresult) {
|
|
||||||
return toupdateresult.map(c -> {
|
|
||||||
List<Country> countryList = new ArrayList<>();
|
|
||||||
List<String> tmp = c.getList(1);
|
|
||||||
for (String country : tmp) {
|
|
||||||
countryList.add(getCountry(country));
|
|
||||||
}
|
|
||||||
Result r = new Result();
|
|
||||||
r.setId(c.getString(0));
|
|
||||||
r.setCountry(countryList);
|
|
||||||
return r;
|
|
||||||
}).mapToPair(r -> new Tuple2<>(r.getId(), r));
|
|
||||||
}
|
|
||||||
|
|
||||||
private static void createUpdateForResultDatasetWrite(JavaRDD<Row> toupdateresult, String outputPath, String type){
|
|
||||||
toupdateresult.map(c -> {
|
|
||||||
List<Country> countryList = new ArrayList<>();
|
|
||||||
List<String> tmp = c.getList(1);
|
|
||||||
for (String country : tmp) {
|
|
||||||
countryList.add(getCountry(country));
|
|
||||||
}
|
|
||||||
Result r = new Result();
|
|
||||||
r.setId(c.getString(0));
|
|
||||||
r.setCountry(countryList);
|
|
||||||
return r;
|
|
||||||
|
|
||||||
}).map(r ->new ObjectMapper().writeValueAsString(r))
|
|
||||||
.saveAsTextFile(outputPath+"/"+type);
|
|
||||||
}
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
@ -1,118 +0,0 @@
|
||||||
package eu.dnetlib.dhp.projecttoresult;
|
|
||||||
|
|
||||||
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
||||||
import eu.dnetlib.dhp.TypedRow;
|
|
||||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
||||||
import eu.dnetlib.dhp.schema.oaf.*;
|
|
||||||
import org.apache.commons.io.IOUtils;
|
|
||||||
import org.apache.hadoop.io.Text;
|
|
||||||
import org.apache.spark.api.java.JavaPairRDD;
|
|
||||||
import org.apache.spark.api.java.JavaRDD;
|
|
||||||
import org.apache.spark.api.java.JavaSparkContext;
|
|
||||||
import org.apache.spark.sql.SparkSession;
|
|
||||||
import scala.Tuple2;
|
|
||||||
|
|
||||||
import java.io.File;
|
|
||||||
import java.util.ArrayList;
|
|
||||||
import java.util.Arrays;
|
|
||||||
import java.util.HashSet;
|
|
||||||
import java.util.List;
|
|
||||||
|
|
||||||
import static eu.dnetlib.dhp.PropagationConstant.*;
|
|
||||||
import static eu.dnetlib.dhp.PropagationConstant.toPair;
|
|
||||||
|
|
||||||
public class SparkResultToProjectThroughSemRelJob {
|
|
||||||
public static void main(String[] args) throws Exception {
|
|
||||||
|
|
||||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(IOUtils.toString(SparkResultToProjectThroughSemRelJob.class.getResourceAsStream("/eu/dnetlib/dhp/projecttoresult/input_projecttoresult_parameters.json")));
|
|
||||||
parser.parseArgument(args);
|
|
||||||
final SparkSession spark = SparkSession
|
|
||||||
.builder()
|
|
||||||
.appName(SparkResultToProjectThroughSemRelJob.class.getSimpleName())
|
|
||||||
.master(parser.get("master"))
|
|
||||||
.enableHiveSupport()
|
|
||||||
.getOrCreate();
|
|
||||||
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
final String inputPath = parser.get("sourcePath");
|
|
||||||
final String outputPath = "/tmp/provision/propagation/projecttoresult";
|
|
||||||
|
|
||||||
final List<String> allowedsemrel = Arrays.asList(parser.get("allowedsemrels").split(";"));
|
|
||||||
|
|
||||||
File directory = new File(outputPath);
|
|
||||||
|
|
||||||
if (!directory.exists()) {
|
|
||||||
directory.mkdirs();
|
|
||||||
}
|
|
||||||
|
|
||||||
JavaRDD<Relation> relations = sc.sequenceFile(inputPath + "/relation", Text.class, Text.class)
|
|
||||||
.map(item -> new ObjectMapper().readValue(item._2().toString(), Relation.class)).cache();
|
|
||||||
|
|
||||||
JavaPairRDD<String, TypedRow> result_result = getResultResultSemRel(allowedsemrel, relations);
|
|
||||||
|
|
||||||
JavaPairRDD<String, TypedRow> result_project = relations
|
|
||||||
.filter(r -> !r.getDataInfo().getDeletedbyinference())
|
|
||||||
.filter(r -> RELATION_RESULT_PROJECT_REL_CLASS.equals(r.getRelClass()) && RELATION_RESULTPROJECT_REL_TYPE.equals(r.getRelType()))
|
|
||||||
.map(r -> new TypedRow().setSourceId(r.getSource()).setTargetId(r.getTarget()))
|
|
||||||
.mapToPair(toPair());
|
|
||||||
|
|
||||||
//relationships from project to result. One pair for each relationship for results having allowed semantics relation with another result
|
|
||||||
JavaPairRDD<String, TypedRow> project_result = result_project.join(result_result)
|
|
||||||
.map(c -> {
|
|
||||||
String projectId = c._2()._1().getTargetId();
|
|
||||||
String resultId = c._2()._2().getTargetId();
|
|
||||||
return new TypedRow().setSourceId(projectId).setTargetId(resultId);
|
|
||||||
})
|
|
||||||
.mapToPair(toPair());
|
|
||||||
|
|
||||||
//relationships from project to result. One Pair for each project => project id list of results related to the project
|
|
||||||
JavaPairRDD<String, TypedRow> project_results = relations
|
|
||||||
.filter(r -> !r.getDataInfo().getDeletedbyinference())
|
|
||||||
.filter(r -> RELATION_PROJECT_RESULT_REL_CLASS.equals(r.getRelClass()) && RELATION_RESULTPROJECT_REL_TYPE.equals(r.getRelType()))
|
|
||||||
.map(r -> new TypedRow().setSourceId(r.getSource()).setTargetId(r.getTarget()))
|
|
||||||
.mapToPair(toPair())
|
|
||||||
.reduceByKey((a, b) -> {
|
|
||||||
if (a == null) {
|
|
||||||
return b;
|
|
||||||
}
|
|
||||||
if (b == null) {
|
|
||||||
return a;
|
|
||||||
}
|
|
||||||
a.addAll(b.getAccumulator());
|
|
||||||
return a;
|
|
||||||
});
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
JavaRDD<Relation> newRels = project_result.join(project_results)
|
|
||||||
.flatMap(c -> {
|
|
||||||
String resId = c._2()._1().getTargetId();
|
|
||||||
|
|
||||||
if (c._2()._2().getAccumulator().contains(resId)) {
|
|
||||||
return null;
|
|
||||||
}
|
|
||||||
String progId = c._2()._1().getSourceId();
|
|
||||||
List<Relation> rels = new ArrayList();
|
|
||||||
|
|
||||||
rels.add(getRelation(progId, resId, RELATION_PROJECT_RESULT_REL_CLASS,
|
|
||||||
RELATION_RESULTPROJECT_REL_TYPE, RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID, PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
rels.add(getRelation(resId, progId, RELATION_RESULT_PROJECT_REL_CLASS,
|
|
||||||
RELATION_RESULTPROJECT_REL_TYPE, RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID, PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
return rels.iterator();
|
|
||||||
})
|
|
||||||
.cache();
|
|
||||||
|
|
||||||
newRels.map(p -> new ObjectMapper().writeValueAsString(p))
|
|
||||||
.saveAsTextFile(outputPath + "/relation_new");
|
|
||||||
|
|
||||||
newRels.union(relations).map(p -> new ObjectMapper().writeValueAsString(p))
|
|
||||||
.saveAsTextFile(outputPath + "/relation");
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
@ -1,222 +0,0 @@
|
||||||
package eu.dnetlib.dhp.projecttoresult;
|
|
||||||
|
|
||||||
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
||||||
import eu.dnetlib.dhp.TypedRow;
|
|
||||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
||||||
import eu.dnetlib.dhp.schema.oaf.Relation;
|
|
||||||
import org.apache.commons.io.IOUtils;
|
|
||||||
import org.apache.hadoop.fs.FileSystem;
|
|
||||||
import org.apache.spark.SparkConf;
|
|
||||||
import org.apache.spark.api.java.JavaPairRDD;
|
|
||||||
import org.apache.spark.api.java.JavaRDD;
|
|
||||||
import org.apache.spark.api.java.JavaSparkContext;
|
|
||||||
import org.apache.spark.sql.Dataset;
|
|
||||||
import org.apache.spark.sql.Encoders;
|
|
||||||
import org.apache.spark.sql.Row;
|
|
||||||
import org.apache.spark.sql.SparkSession;
|
|
||||||
import scala.Tuple2;
|
|
||||||
|
|
||||||
import java.util.ArrayList;
|
|
||||||
import java.util.Arrays;
|
|
||||||
import java.util.List;
|
|
||||||
import java.util.Set;
|
|
||||||
|
|
||||||
import static eu.dnetlib.dhp.PropagationConstant.*;
|
|
||||||
|
|
||||||
public class SparkResultToProjectThroughSemRelJob2 {
|
|
||||||
public static void main(String[] args) throws Exception {
|
|
||||||
|
|
||||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
|
||||||
IOUtils.toString(SparkResultToProjectThroughSemRelJob2.class
|
|
||||||
.getResourceAsStream("/eu/dnetlib/dhp/projecttoresult/input_projecttoresult_parameters.json")));
|
|
||||||
parser.parseArgument(args);
|
|
||||||
|
|
||||||
parser.getObjectMap().keySet().stream().forEach(k -> System.out.println(k + " = " + parser.getObjectMap().get(k)));
|
|
||||||
SparkConf conf = new SparkConf();
|
|
||||||
conf.set("hive.metastore.uris", parser.get("hive_metastore_uris"));
|
|
||||||
final SparkSession spark = SparkSession
|
|
||||||
.builder()
|
|
||||||
.appName(SparkResultToProjectThroughSemRelJob2.class.getSimpleName())
|
|
||||||
.master(parser.get("master"))
|
|
||||||
.config(conf)
|
|
||||||
.enableHiveSupport()
|
|
||||||
.getOrCreate();
|
|
||||||
|
|
||||||
|
|
||||||
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
||||||
final String inputPath = parser.get("sourcePath");
|
|
||||||
final String outputPath = "/tmp/provision/propagation/projecttoresult";
|
|
||||||
boolean writeUpdates = "true".equals(parser.get("writeUpdate"));
|
|
||||||
boolean saveGraph = "true".equals(parser.get("saveGraph"));
|
|
||||||
|
|
||||||
final List<String> allowedsemrel = Arrays.asList(parser.get("allowedsemrels").split(";"));
|
|
||||||
|
|
||||||
createOutputDirs(outputPath, FileSystem.get(spark.sparkContext().hadoopConfiguration()));
|
|
||||||
|
|
||||||
Dataset<Relation> relation = spark.createDataset(sc.textFile(inputPath + "/relation")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Relation.class)).rdd(), Encoders.bean(Relation.class));
|
|
||||||
|
|
||||||
relation.createOrReplaceTempView("relation");
|
|
||||||
|
|
||||||
String query = "Select source, target " +
|
|
||||||
"from relation " +
|
|
||||||
"where datainfo.deletedbyinference = false and relClass = '" + RELATION_RESULT_PROJECT_REL_CLASS + "'";
|
|
||||||
|
|
||||||
Dataset<Row> resproj_relation = spark.sql(query);
|
|
||||||
|
|
||||||
query = "Select source, target " +
|
|
||||||
"from relation " +
|
|
||||||
"where datainfo.deletedbyinference = false " + getConstraintList(" relClass = '", allowedsemrel );
|
|
||||||
|
|
||||||
Dataset<Row> resres_relation = spark.sql(query);
|
|
||||||
resres_relation.createOrReplaceTempView("resres_relation");
|
|
||||||
resproj_relation.createOrReplaceTempView("resproj_relation");
|
|
||||||
|
|
||||||
query ="SELECT proj, collect_set(r1target) result_set " +
|
|
||||||
"FROM (" +
|
|
||||||
" SELECT r1.source as source, r1.target as r1target, r2.target as proj " +
|
|
||||||
" FROM resres_relation r1 " +
|
|
||||||
" JOIN resproj_relation r2 " +
|
|
||||||
" ON r1.source = r2.source " +
|
|
||||||
" ) tmp " +
|
|
||||||
"GROUP BY proj ";
|
|
||||||
|
|
||||||
Dataset<Row> toaddrelations = spark.sql(query);
|
|
||||||
|
|
||||||
query = "select target, collect_set(source) result_list from " +
|
|
||||||
"resproj_relation " +
|
|
||||||
"group by target";
|
|
||||||
|
|
||||||
Dataset<Row> project_resultlist = spark.sql(query);
|
|
||||||
|
|
||||||
//if (writeUpdaes){
|
|
||||||
toaddrelations.toJavaRDD().map(r->new ObjectMapper().writeValueAsString(r))
|
|
||||||
.saveAsTextFile(outputPath + "/toupdaterelations");
|
|
||||||
//}
|
|
||||||
|
|
||||||
if(saveGraph){
|
|
||||||
JavaRDD<Relation> new_relations = toaddrelations.toJavaRDD().mapToPair(r -> new Tuple2<>(r.getString(0), r.getList(1)))
|
|
||||||
.leftOuterJoin(project_resultlist.toJavaRDD().mapToPair(pr -> new Tuple2<>(pr.getString(0), pr.getList(1))))
|
|
||||||
.flatMap(c -> {
|
|
||||||
List<Object> toAddRel = new ArrayList<>();
|
|
||||||
toAddRel.addAll(c._2()._1());
|
|
||||||
if (c._2()._2().isPresent()) {
|
|
||||||
List<Object> originalRels = c._2()._2().get();
|
|
||||||
for (Object o : originalRels) {
|
|
||||||
if (toAddRel.contains(o)) {
|
|
||||||
toAddRel.remove(o);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
}
|
|
||||||
List<Relation> relationList = new ArrayList<>();
|
|
||||||
String projId = c._1();
|
|
||||||
for (Object r : toAddRel) {
|
|
||||||
String rId = (String) r;
|
|
||||||
relationList.add(getRelation(rId, projId, RELATION_RESULT_PROJECT_REL_CLASS, RELATION_RESULTPROJECT_REL_TYPE,
|
|
||||||
RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
relationList.add(getRelation(projId, rId, RELATION_PROJECT_RESULT_REL_CLASS, RELATION_RESULTPROJECT_REL_TYPE,
|
|
||||||
RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID,
|
|
||||||
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
|
|
||||||
}
|
|
||||||
if(relationList.size()==0){
|
|
||||||
return null;
|
|
||||||
}
|
|
||||||
return relationList.iterator();
|
|
||||||
}).filter(r -> !(r==null));
|
|
||||||
|
|
||||||
|
|
||||||
new_relations.map(r-> new ObjectMapper().writeValueAsString(r))
|
|
||||||
.saveAsTextFile(outputPath + "/new_relations" );
|
|
||||||
|
|
||||||
sc.textFile(inputPath + "/relation")
|
|
||||||
.map(item -> new ObjectMapper().readValue(item, Relation.class))
|
|
||||||
.union(new_relations)
|
|
||||||
.map(r -> new ObjectMapper().writeValueAsString(r))
|
|
||||||
.saveAsTextFile(outputPath + "/relation");
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
||||||
|
|
||||||
//JavaPairRDD<String, TypedRow> result_result = getResultResultSemRel(allowedsemrel, relations);
|
|
||||||
|
|
||||||
// JavaPairRDD<String, TypedRow> result_project = relations
|
|
||||||
// .filter(r -> !r.getDataInfo().getDeletedbyinference())
|
|
||||||
// .filter(r -> RELATION_RESULT_PROJECT_REL_CLASS.equals(r.getRelClass())
|
|
||||||
// && RELATION_RESULTPROJECT_REL_TYPE.equals(r.getRelType()))
|
|
||||||
// .map(rel ->{
|
|
||||||
//
|
|
||||||
// TypedRow tr = new TypedRow();
|
|
||||||
// tr.setSourceId(rel.getSource());
|
|
||||||
// tr.setTargetId(rel.getTarget());
|
|
||||||
// return tr;
|
|
||||||
// })
|
|
||||||
// .mapToPair(toPair());
|
|
||||||
//
|
|
||||||
// //relationships from project to result. One pair for each relationship for results having allowed semantics relation with another result
|
|
||||||
// JavaPairRDD<String, TypedRow> project_result = result_project.join(result_result)
|
|
||||||
// .map(c -> {
|
|
||||||
// String projectId = c._2()._1().getTargetId();
|
|
||||||
// String resultId = c._2()._2().getTargetId();
|
|
||||||
// TypedRow tr = new TypedRow(); tr.setSourceId(projectId); tr.setTargetId(resultId);
|
|
||||||
// return tr;
|
|
||||||
// })
|
|
||||||
// .mapToPair(toPair());
|
|
||||||
//
|
|
||||||
// //relationships from project to result. One Pair for each project => project id list of results related to the project
|
|
||||||
// JavaPairRDD<String, TypedRow> project_results = relations
|
|
||||||
// .filter(r -> !r.getDataInfo().getDeletedbyinference())
|
|
||||||
// .filter(r -> RELATION_PROJECT_RESULT_REL_CLASS.equals(r.getRelClass()) && RELATION_RESULTPROJECT_REL_TYPE.equals(r.getRelType()))
|
|
||||||
// .map(r -> {
|
|
||||||
// TypedRow tr = new TypedRow(); tr.setSourceId(r.getSource()); tr.setTargetId(r.getTarget());
|
|
||||||
// return tr;
|
|
||||||
// })
|
|
||||||
// .mapToPair(toPair())
|
|
||||||
// .reduceByKey((a, b) -> {
|
|
||||||
// if (a == null) {
|
|
||||||
// return b;
|
|
||||||
// }
|
|
||||||
// if (b == null) {
|
|
||||||
// return a;
|
|
||||||
// }
|
|
||||||
// a.addAll(b.getAccumulator());
|
|
||||||
// return a;
|
|
||||||
// });
|
|
||||||
//
|
|
||||||
//
|
|
||||||
//
|
|
||||||
// JavaRDD<Relation> newRels = project_result.join(project_results)
|
|
||||||
// .flatMap(c -> {
|
|
||||||
// String resId = c._2()._1().getTargetId();
|
|
||||||
//
|
|
||||||
// if (c._2()._2().getAccumulator().contains(resId)) {
|
|
||||||
// return null;
|
|
||||||
// }
|
|
||||||
// String progId = c._2()._1().getSourceId();
|
|
||||||
// List<Relation> rels = new ArrayList();
|
|
||||||
//
|
|
||||||
// rels.add(getRelation(progId, resId, RELATION_PROJECT_RESULT_REL_CLASS,
|
|
||||||
// RELATION_RESULTPROJECT_REL_TYPE, RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
// PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID, PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
// rels.add(getRelation(resId, progId, RELATION_RESULT_PROJECT_REL_CLASS,
|
|
||||||
// RELATION_RESULTPROJECT_REL_TYPE, RELATION_RESULTPROJECT_SUBREL_TYPE, PROPAGATION_DATA_INFO_TYPE,
|
|
||||||
// PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID, PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME));
|
|
||||||
// return rels.iterator();
|
|
||||||
// })
|
|
||||||
// .cache();
|
|
||||||
//
|
|
||||||
// newRels.map(p -> new ObjectMapper().writeValueAsString(p))
|
|
||||||
// .saveAsTextFile(outputPath + "/relation_new");
|
|
||||||
//
|
|
||||||
// newRels.union(relations).map(p -> new ObjectMapper().writeValueAsString(p))
|
|
||||||
// .saveAsTextFile(outputPath + "/relation");
|
|
||||||
|
|
||||||
}
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
}
|
|
||||||
Loading…
Reference in New Issue