modification to the classes building the actionset to consider the h2020classification

This commit is contained in:
Miriam Baglioni 2020-09-23 17:31:15 +02:00
parent 3c7ef5ca04
commit 2cba3cb484
3 changed files with 203 additions and 46 deletions

View File

@ -4,11 +4,14 @@ package eu.dnetlib.dhp.actionmanager.project;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession; import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.HashMap; import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Optional; import java.util.Optional;
import org.apache.commons.io.IOUtils; import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils; import org.apache.commons.lang3.StringUtils;
import org.apache.spark.SparkConf; import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.function.MapFunction; import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.*; import org.apache.spark.sql.*;
import org.slf4j.Logger; import org.slf4j.Logger;
@ -66,49 +69,182 @@ public class PrepareProgramme {
HdfsSupport.remove(path, spark.sparkContext().hadoopConfiguration()); HdfsSupport.remove(path, spark.sparkContext().hadoopConfiguration());
} }
// private static void exec(SparkSession spark, String programmePath, String outputPath) {
// Dataset<CSVProgramme> programme = readPath(spark, programmePath, CSVProgramme.class);
//
// programme
// .toJavaRDD()
// .filter(p -> !p.getCode().contains("FP7"))
// .mapToPair(csvProgramme -> new Tuple2<>(csvProgramme.getCode(), csvProgramme))
// .reduceByKey((a, b) -> {
// if (StringUtils.isEmpty(a.getShortTitle())) {
// if (StringUtils.isEmpty(b.getShortTitle())) {
// if (StringUtils.isEmpty(a.getTitle())) {
// if (StringUtils.isNotEmpty(b.getTitle())) {
// a.setShortTitle(b.getTitle());
// a.setLanguage(b.getLanguage());
// }
// } else {// notIsEmpty a.getTitle
// if (StringUtils.isEmpty(b.getTitle())) {
// a.setShortTitle(a.getTitle());
// } else {
// if (b.getLanguage().equalsIgnoreCase("en")) {
// a.setShortTitle(b.getTitle());
// a.setLanguage(b.getLanguage());
// } else {
// a.setShortTitle(a.getTitle());
// }
// }
// }
// } else {// not isEmpty b.getShortTitle
// a.setShortTitle(b.getShortTitle());
// // a.setLanguage(b.getLanguage());
// }
// }
// return a;
//
// })
// .map(p -> {
// CSVProgramme csvProgramme = p._2();
// if (StringUtils.isEmpty(csvProgramme.getShortTitle())) {
// csvProgramme.setShortTitle(csvProgramme.getTitle());
// }
// return OBJECT_MAPPER.writeValueAsString(csvProgramme);
// })
// .saveAsTextFile(outputPath);
//
// }
private static void exec(SparkSession spark, String programmePath, String outputPath) { private static void exec(SparkSession spark, String programmePath, String outputPath) {
Dataset<CSVProgramme> programme = readPath(spark, programmePath, CSVProgramme.class); Dataset<CSVProgramme> programme = readPath(spark, programmePath, CSVProgramme.class);
programme JavaRDD<CSVProgramme> h2020Programmes = programme
.toJavaRDD() .toJavaRDD()
.filter(p -> !p.getCode().contains("FP7")) .filter(p -> !p.getCode().contains("FP7"))
.mapToPair(csvProgramme -> new Tuple2<>(csvProgramme.getCode(), csvProgramme)) .mapToPair(csvProgramme -> new Tuple2<>(csvProgramme.getCode(), csvProgramme))
.reduceByKey((a, b) -> { .reduceByKey((a, b) -> {
if (StringUtils.isEmpty(a.getShortTitle())) { if (!a.getLanguage().equals("en")) {
if (StringUtils.isEmpty(b.getShortTitle())) {
if (StringUtils.isEmpty(a.getTitle())) {
if (StringUtils.isNotEmpty(b.getTitle())) {
a.setShortTitle(b.getTitle());
a.setLanguage(b.getLanguage());
}
} else {// notIsEmpty a.getTitle
if (StringUtils.isEmpty(b.getTitle())) {
a.setShortTitle(a.getTitle());
} else {
if (b.getLanguage().equalsIgnoreCase("en")) { if (b.getLanguage().equalsIgnoreCase("en")) {
a.setShortTitle(b.getTitle()); a.setTitle(b.getTitle());
a.setLanguage(b.getLanguage()); a.setLanguage(b.getLanguage());
} else {
a.setShortTitle(a.getTitle());
} }
} }
} if (StringUtils.isEmpty(a.getShortTitle())) {
} else {// not isEmpty b.getShortTitle if (!StringUtils.isEmpty(b.getShortTitle())) {
a.setShortTitle(b.getShortTitle()); a.setShortTitle(b.getShortTitle());
// a.setLanguage(b.getLanguage());
} }
} }
return a; return a;
}) })
.map(p -> { .map(p -> {
CSVProgramme csvProgramme = p._2(); CSVProgramme csvProgramme = p._2();
if (StringUtils.isEmpty(csvProgramme.getShortTitle())) { String programmeTitle = csvProgramme.getTitle().trim();
csvProgramme.setShortTitle(csvProgramme.getTitle()); if (programmeTitle.length() > 8 && programmeTitle.substring(0, 8).equalsIgnoreCase("PRIORITY")) {
programmeTitle = programmeTitle.substring(9);
if (programmeTitle.charAt(0) == '\'') {
programmeTitle = programmeTitle.substring(1);
} }
if (programmeTitle.charAt(programmeTitle.length() - 1) == '\'') {
programmeTitle = programmeTitle.substring(0, programmeTitle.length() - 1);
}
csvProgramme.setTitle(programmeTitle);
}
return csvProgramme;
});
Object[] codedescription = h2020Programmes
.map(value -> new Tuple2<>(value.getCode(), value.getTitle()))
.collect()
.toArray();
for (int i = 0; i < codedescription.length - 1; i++) {
for (int j = i + 1; j < codedescription.length; j++) {
Tuple2<String, String> t2i = (Tuple2<String, String>) codedescription[i];
Tuple2<String, String> t2j = (Tuple2<String, String>) codedescription[j];
if (t2i._1().compareTo(t2j._1()) > 0) {
Tuple2<String, String> temp = t2i;
codedescription[i] = t2j;
codedescription[j] = temp;
}
}
}
Map<String, String> map = new HashMap<>();
for (int j = 0; j < codedescription.length; j++) {
Tuple2<String, String> entry = (Tuple2<String, String>) codedescription[j];
String ent = entry._1();
if (ent.contains("Euratom-")) {
ent = ent.replace("-Euratom-", ".Euratom.");
}
String[] tmp = ent.split("\\.");
if (tmp.length <= 2) {
map.put(entry._1(), entry._2());
} else {
if (ent.endsWith(".")) {
ent = ent.substring(0, ent.length() - 1);
}
String key = ent.substring(0, ent.lastIndexOf(".") + 1);
if (key.contains("Euratom")) {
key = key.replace(".Euratom.", "-Euratom-");
ent = ent.replace(".Euratom.", "-Euratom-");
if (key.endsWith("-")) {
key = key.substring(0, key.length() - 1);
}
}
String current = entry._2();
if (!ent.contains("Euratom")) {
String parent;
String tmp_key = tmp[0] + ".";
for (int i = 1; i < tmp.length - 1; i++) {
tmp_key += tmp[i] + ".";
parent = map.get(tmp_key).toLowerCase().trim();
if (parent.contains("|")) {
parent = parent.substring(parent.lastIndexOf("|") + 1).trim();
}
if (current.trim().length() > parent.length()
&& current.toLowerCase().trim().substring(0, parent.length()).equals(parent)) {
current = current.substring(parent.length() + 1);
if (current.trim().charAt(0) == '-') {
current = current.trim().substring(1).trim();
}
}
}
}
map.put(ent + ".", map.get(key) + " | " + current);
// String current = entry._2();
// String parent;
// String tmp_key = tmp[0] + ".";
// for (int i = 1; i< tmp.length -1; i++){
// tmp_key += tmp[i] + ".";
// parent = map.get(tmp_key).toLowerCase().trim();
// if (current.trim().length() > parent.length() && current.toLowerCase().trim().substring(0, parent.length()).equals(parent)){
// current = current.substring(parent.length()+1);
// if(current.trim().charAt(0) == '-'){
// current = current.trim().substring(1).trim();
// }
//
// }
// }
//
// map.put(ent + ".", map.get(key) + " $ " + current);
}
}
h2020Programmes.map(csvProgramme -> {
if (!csvProgramme.getCode().endsWith(".") && !csvProgramme.getCode().contains("Euratom")
&& !csvProgramme.getCode().equals("H2020-EC"))
csvProgramme.setClassification(map.get(csvProgramme.getCode() + "."));
else
csvProgramme.setClassification(map.get(csvProgramme.getCode()));
return OBJECT_MAPPER.writeValueAsString(csvProgramme); return OBJECT_MAPPER.writeValueAsString(csvProgramme);
}) }).saveAsTextFile(outputPath);
.saveAsTextFile(outputPath);
} }

View File

@ -3,28 +3,20 @@ package eu.dnetlib.dhp.actionmanager.project;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession; import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.io.IOException;
import java.util.Arrays; import java.util.Arrays;
import java.util.HashMap; import java.util.HashMap;
import java.util.Objects; import java.util.Objects;
import java.util.Optional; import java.util.Optional;
import java.util.function.Consumer;
import org.apache.commons.io.IOUtils; import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration; import org.apache.commons.lang3.StringUtils;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.SequenceFile;
import org.apache.hadoop.io.Text; import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.SequenceFileOutputFormat; import org.apache.hadoop.mapred.SequenceFileOutputFormat;
import org.apache.hadoop.mapred.TextOutputFormat;
import org.apache.spark.SparkConf; import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.MapFunction; import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.api.java.function.MapGroupsFunction; import org.apache.spark.api.java.function.MapGroupsFunction;
import org.apache.spark.rdd.SequenceFileRDDFunctions;
import org.apache.spark.sql.Dataset; import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders; import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SaveMode;
import org.apache.spark.sql.SparkSession; import org.apache.spark.sql.SparkSession;
import org.slf4j.Logger; import org.slf4j.Logger;
import org.slf4j.LoggerFactory; import org.slf4j.LoggerFactory;
@ -37,12 +29,11 @@ import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport; import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.schema.action.AtomicAction; import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.dhp.schema.common.ModelSupport; import eu.dnetlib.dhp.schema.common.ModelSupport;
import eu.dnetlib.dhp.schema.oaf.Programme; import eu.dnetlib.dhp.schema.oaf.H2020Classification;
import eu.dnetlib.dhp.schema.oaf.H2020Programme;
import eu.dnetlib.dhp.schema.oaf.Project; import eu.dnetlib.dhp.schema.oaf.Project;
import eu.dnetlib.dhp.utils.DHPUtils; import eu.dnetlib.dhp.utils.DHPUtils;
import scala.Function1;
import scala.Tuple2; import scala.Tuple2;
import scala.runtime.BoxedUnit;
public class SparkAtomicActionJob { public class SparkAtomicActionJob {
private static final Logger log = LoggerFactory.getLogger(SparkAtomicActionJob.class); private static final Logger log = LoggerFactory.getLogger(SparkAtomicActionJob.class);
@ -105,20 +96,30 @@ public class SparkAtomicActionJob {
project project
.joinWith(programme, project.col("programme").equalTo(programme.col("code")), "left") .joinWith(programme, project.col("programme").equalTo(programme.col("code")), "left")
.map(c -> { .map((MapFunction<Tuple2<CSVProject, CSVProgramme>, Project>) c -> {
CSVProject csvProject = c._1(); CSVProject csvProject = c._1();
Optional<CSVProgramme> csvProgramme = Optional.ofNullable(c._2()); Optional<CSVProgramme> ocsvProgramme = Optional.ofNullable(c._2());
if (csvProgramme.isPresent()) { if (ocsvProgramme.isPresent()) {
Project p = new Project(); Project p = new Project();
p p
.setId( .setId(
createOpenaireId( createOpenaireId(
ModelSupport.entityIdPrefix.get("project"), ModelSupport.entityIdPrefix.get("project"),
"corda__h2020", csvProject.getId())); "corda__h2020", csvProject.getId()));
Programme pm = new Programme(); p.setH2020topiccode(csvProject.getTopics());
H2020Programme pm = new H2020Programme();
H2020Classification h2020classification = new H2020Classification();
pm.setCode(csvProject.getProgramme()); pm.setCode(csvProject.getProgramme());
pm.setDescription(csvProgramme.get().getShortTitle()); CSVProgramme csvProgramme = ocsvProgramme.get();
p.setProgramme(Arrays.asList(pm)); if (StringUtils.isNotEmpty(csvProgramme.getShortTitle())) {
pm.setDescription(csvProgramme.getShortTitle());
} else {
pm.setDescription(csvProgramme.getTitle());
}
h2020classification.setClassification(ocsvProgramme.get().getClassification());
setLevels(h2020classification, ocsvProgramme.get().getClassification());
h2020classification.setH2020Programme(pm);
p.setH2020classification(Arrays.asList(h2020classification));
return p; return p;
} }
@ -144,6 +145,17 @@ public class SparkAtomicActionJob {
} }
private static void setLevels(H2020Classification h2020Classification, String classification) {
String[] tmp = classification.split(" | ");
h2020Classification.setLevel1(tmp[0]);
if (tmp.length > 1) {
h2020Classification.setLevel2(tmp[1]);
}
if (tmp.length > 2) {
h2020Classification.setLevel3(tmp[2]);
}
}
public static <R> Dataset<R> readPath( public static <R> Dataset<R> readPath(
SparkSession spark, String inputPath, Class<R> clazz) { SparkSession spark, String inputPath, Class<R> clazz) {
return spark return spark

View File

@ -9,6 +9,15 @@ public class CSVProgramme implements Serializable {
private String title; private String title;
private String shortTitle; private String shortTitle;
private String language; private String language;
private String classification;
public String getClassification() {
return classification;
}
public void setClassification(String classification) {
this.classification = classification;
}
public String getRcn() { public String getRcn() {
return rcn; return rcn;