2020-05-19 18:42:50 +02:00
|
|
|
|
|
|
|
|
|
package eu.dnetlib.dhp.actionmanager.project;
|
|
|
|
|
|
|
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
|
|
|
|
|
|
|
|
|
import java.util.HashMap;
|
2020-10-20 17:03:25 +02:00
|
|
|
|
import java.util.List;
|
2020-09-23 17:31:15 +02:00
|
|
|
|
import java.util.Map;
|
2020-05-19 18:42:50 +02:00
|
|
|
|
import java.util.Optional;
|
|
|
|
|
|
|
|
|
|
import org.apache.commons.io.IOUtils;
|
|
|
|
|
import org.apache.commons.lang3.StringUtils;
|
|
|
|
|
import org.apache.spark.SparkConf;
|
2020-09-23 17:31:15 +02:00
|
|
|
|
import org.apache.spark.api.java.JavaRDD;
|
2020-10-20 17:03:25 +02:00
|
|
|
|
import org.apache.spark.api.java.JavaSparkContext;
|
2020-05-19 18:42:50 +02:00
|
|
|
|
import org.apache.spark.api.java.function.MapFunction;
|
|
|
|
|
import org.apache.spark.sql.*;
|
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
|
|
|
|
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
|
|
|
|
2020-09-28 12:06:56 +02:00
|
|
|
|
import eu.dnetlib.dhp.actionmanager.project.utils.CSVProgramme;
|
2020-05-19 18:42:50 +02:00
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
|
|
|
import eu.dnetlib.dhp.common.HdfsSupport;
|
|
|
|
|
import scala.Tuple2;
|
|
|
|
|
|
2020-10-05 13:54:39 +02:00
|
|
|
|
/**
|
|
|
|
|
* Among all the programmes provided in the csv file, selects those in H2020 framework that have an english title.
|
|
|
|
|
*
|
|
|
|
|
* The title is then handled to get the programme description at a certain level. The set of programme titles will then
|
|
|
|
|
* be used to associate a classification for the programme.
|
|
|
|
|
*
|
|
|
|
|
* The programme code describes an hierarchy that can be exploited to provide the classification. To determine the hierarchy
|
|
|
|
|
* the code can be split by '.'. If the length of the splitted code is less than or equal to 2 it can be directly used
|
|
|
|
|
* as the classification: H2020-EU -> Horizon 2020 Framework Programme (It will never be repeated),
|
|
|
|
|
* H2020-EU.1. -> Excellent science, H2020-EU.2. -> Industrial leadership etc.
|
|
|
|
|
*
|
|
|
|
|
* The codes are ordered and for all of them the concatenation of all the titles (from the element in position 1 of
|
|
|
|
|
* the splitted code) handled as below is used to create the classification. For example:
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.1.1 -> Excellent science | European Research Council (ERC)
|
|
|
|
|
* from H2020-EU.1. -> Excellence science and H2020-EU.1.1. -> European Research Council (ERC)
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.3.1.3.1. -> Societal challenges | Health, demographic change and well-being | Treating and managing disease | Treating disease, including developing regenerative medicine
|
|
|
|
|
* from H2020-EU.3. -> Societal challenges,
|
|
|
|
|
* H2020-EU.3.1. -> Health, demographic change and well-being
|
|
|
|
|
* H2020-EU.3.1.3 -> Treating and managing disease
|
|
|
|
|
* H2020-EU.3.1.3.1. -> Treating disease, including developing regenerative medicine
|
|
|
|
|
*
|
|
|
|
|
* The classification up to level three, will be split in dedicated variables, while the complete classification will be stored
|
|
|
|
|
* in a variable called classification and provided as shown above.
|
|
|
|
|
*
|
|
|
|
|
* The programme title is not give in a standardized way:
|
|
|
|
|
*
|
|
|
|
|
* - Sometimes associated to the higher level in the hierarchy we can find Priority in title other times it is not the
|
|
|
|
|
* case. Since it is not uniform, we removed priority from the handled titles:
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.1. -> PRIORITY 'Excellent science'
|
|
|
|
|
* H2020-EU.2. -> PRIORITY 'Industrial leadership'
|
|
|
|
|
* H2020-EU.3. -> PRIORITY 'Societal challenges
|
|
|
|
|
*
|
|
|
|
|
* will become
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.1. -> Excellent science
|
|
|
|
|
* H2020-EU.2. -> Industrial leadership
|
|
|
|
|
* H2020-EU.3. -> Societal challenges
|
|
|
|
|
*
|
|
|
|
|
* - Sometimes the title of the parent is repeated in the title for the code, but it is not always the case, so, titles
|
|
|
|
|
* associated to previous levels in the hierarchy are removed from the code title.
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.1.2. -> EXCELLENT SCIENCE - Future and Emerging Technologies (FET)
|
|
|
|
|
* H2020-EU.2.2. -> INDUSTRIAL LEADERSHIP - Access to risk finance
|
|
|
|
|
* H2020-EU.3.4. -> SOCIETAL CHALLENGES - Smart, Green And Integrated Transport
|
|
|
|
|
*
|
|
|
|
|
* will become
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.1.2. -> Future and Emerging Technologies (FET)
|
|
|
|
|
* H2020-EU.2.2. -> Access to risk finance
|
|
|
|
|
* H2020-EU.3.4. -> Smart, Green And Integrated Transport
|
|
|
|
|
*
|
|
|
|
|
* This holds at all levels in the hierarchy. Hence
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.2.1.2. -> INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies – Nanotechnologies
|
|
|
|
|
*
|
|
|
|
|
* will become
|
|
|
|
|
*
|
|
|
|
|
* H2020-EU.2.1.2. -> Nanotechnologies
|
|
|
|
|
*
|
|
|
|
|
* - Euratom is not given in the way the other programmes are: H2020-EU. but H2020-Euratom- . So we need to write
|
|
|
|
|
* specific code for it
|
|
|
|
|
*
|
|
|
|
|
*
|
|
|
|
|
*
|
|
|
|
|
*/
|
2020-05-19 18:42:50 +02:00
|
|
|
|
public class PrepareProgramme {
|
|
|
|
|
|
|
|
|
|
private static final Logger log = LoggerFactory.getLogger(PrepareProgramme.class);
|
|
|
|
|
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
|
|
|
|
|
|
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
|
|
|
|
|
|
|
|
String jsonConfiguration = IOUtils
|
|
|
|
|
.toString(
|
|
|
|
|
PrepareProgramme.class
|
|
|
|
|
.getResourceAsStream(
|
|
|
|
|
"/eu/dnetlib/dhp/actionmanager/project/prepare_programme_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 programmePath = parser.get("programmePath");
|
|
|
|
|
log.info("programmePath {}: ", programmePath);
|
|
|
|
|
|
|
|
|
|
final String outputPath = parser.get("outputPath");
|
|
|
|
|
log.info("outputPath {}: ", outputPath);
|
|
|
|
|
|
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
|
|
|
|
|
|
runWithSparkSession(
|
|
|
|
|
conf,
|
|
|
|
|
isSparkSessionManaged,
|
|
|
|
|
spark -> {
|
|
|
|
|
removeOutputDir(spark, outputPath);
|
|
|
|
|
exec(spark, programmePath, outputPath);
|
|
|
|
|
});
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
private static void removeOutputDir(SparkSession spark, String path) {
|
|
|
|
|
HdfsSupport.remove(path, spark.sparkContext().hadoopConfiguration());
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
private static void exec(SparkSession spark, String programmePath, String outputPath) {
|
|
|
|
|
Dataset<CSVProgramme> programme = readPath(spark, programmePath, CSVProgramme.class);
|
|
|
|
|
|
2020-09-23 17:31:15 +02:00
|
|
|
|
JavaRDD<CSVProgramme> h2020Programmes = programme
|
2020-05-19 18:42:50 +02:00
|
|
|
|
.toJavaRDD()
|
2020-09-24 15:19:31 +02:00
|
|
|
|
.filter(p -> p.getFrameworkProgramme().trim().equalsIgnoreCase("H2020"))
|
2020-05-19 18:42:50 +02:00
|
|
|
|
.mapToPair(csvProgramme -> new Tuple2<>(csvProgramme.getCode(), csvProgramme))
|
|
|
|
|
.reduceByKey((a, b) -> {
|
2020-09-23 17:31:15 +02:00
|
|
|
|
if (!a.getLanguage().equals("en")) {
|
|
|
|
|
if (b.getLanguage().equalsIgnoreCase("en")) {
|
|
|
|
|
a.setTitle(b.getTitle());
|
|
|
|
|
a.setLanguage(b.getLanguage());
|
|
|
|
|
}
|
|
|
|
|
}
|
2020-05-19 18:42:50 +02:00
|
|
|
|
if (StringUtils.isEmpty(a.getShortTitle())) {
|
2020-09-23 17:31:15 +02:00
|
|
|
|
if (!StringUtils.isEmpty(b.getShortTitle())) {
|
2020-05-19 18:42:50 +02:00
|
|
|
|
a.setShortTitle(b.getShortTitle());
|
|
|
|
|
}
|
|
|
|
|
}
|
2020-09-23 17:31:15 +02:00
|
|
|
|
|
2020-05-19 18:42:50 +02:00
|
|
|
|
return a;
|
|
|
|
|
|
|
|
|
|
})
|
|
|
|
|
.map(p -> {
|
|
|
|
|
CSVProgramme csvProgramme = p._2();
|
2020-09-23 17:31:15 +02:00
|
|
|
|
String programmeTitle = csvProgramme.getTitle().trim();
|
|
|
|
|
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);
|
2020-05-19 18:42:50 +02:00
|
|
|
|
}
|
2020-09-23 17:31:15 +02:00
|
|
|
|
return csvProgramme;
|
|
|
|
|
});
|
|
|
|
|
|
2020-10-20 17:03:25 +02:00
|
|
|
|
// prepareClassification(h2020Programmes);
|
2020-10-05 13:54:39 +02:00
|
|
|
|
|
2020-10-20 17:03:25 +02:00
|
|
|
|
JavaSparkContext jsc = new JavaSparkContext(spark.sparkContext());
|
|
|
|
|
|
|
|
|
|
JavaRDD<CSVProgramme> rdd = jsc.parallelize(prepareClassification(h2020Programmes), 1);
|
|
|
|
|
rdd
|
|
|
|
|
.map(csvProgramme -> {
|
|
|
|
|
String tmp = OBJECT_MAPPER.writeValueAsString(csvProgramme);
|
|
|
|
|
return tmp;
|
|
|
|
|
})
|
2020-10-07 13:22:54 +02:00
|
|
|
|
.saveAsTextFile(outputPath);
|
2020-10-05 13:54:39 +02:00
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
2020-10-20 17:03:25 +02:00
|
|
|
|
private static List<CSVProgramme> prepareClassification(JavaRDD<CSVProgramme> h2020Programmes) {
|
2020-09-23 17:31:15 +02:00
|
|
|
|
Object[] codedescription = h2020Programmes
|
2020-10-20 17:03:25 +02:00
|
|
|
|
.map(
|
|
|
|
|
value -> new Tuple2<>(value.getCode(),
|
|
|
|
|
new Tuple2<String, String>(value.getTitle(), value.getShortTitle())))
|
2020-09-23 17:31:15 +02:00
|
|
|
|
.collect()
|
|
|
|
|
.toArray();
|
|
|
|
|
|
|
|
|
|
for (int i = 0; i < codedescription.length - 1; i++) {
|
|
|
|
|
for (int j = i + 1; j < codedescription.length; j++) {
|
2020-10-20 17:03:25 +02:00
|
|
|
|
Tuple2<String, Tuple2<String, String>> t2i = (Tuple2<String, Tuple2<String, String>>) codedescription[i];
|
|
|
|
|
Tuple2<String, Tuple2<String, String>> t2j = (Tuple2<String, Tuple2<String, String>>) codedescription[j];
|
2020-09-23 17:31:15 +02:00
|
|
|
|
if (t2i._1().compareTo(t2j._1()) > 0) {
|
2020-10-20 17:03:25 +02:00
|
|
|
|
Tuple2<String, Tuple2<String, String>> temp = t2i;
|
2020-09-23 17:31:15 +02:00
|
|
|
|
codedescription[i] = t2j;
|
|
|
|
|
codedescription[j] = temp;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2020-10-20 17:03:25 +02:00
|
|
|
|
Map<String, Tuple2<String, String>> map = new HashMap<>();
|
2020-09-23 17:31:15 +02:00
|
|
|
|
for (int j = 0; j < codedescription.length; j++) {
|
2020-10-20 17:03:25 +02:00
|
|
|
|
Tuple2<String, Tuple2<String, String>> entry = (Tuple2<String, Tuple2<String, String>>) codedescription[j];
|
2020-09-23 17:31:15 +02:00
|
|
|
|
String ent = entry._1();
|
|
|
|
|
if (ent.contains("Euratom-")) {
|
|
|
|
|
ent = ent.replace("-Euratom-", ".Euratom.");
|
|
|
|
|
}
|
|
|
|
|
String[] tmp = ent.split("\\.");
|
|
|
|
|
if (tmp.length <= 2) {
|
2020-10-20 17:03:25 +02:00
|
|
|
|
if (StringUtils.isEmpty(entry._2()._2())) {
|
|
|
|
|
map.put(entry._1(), new Tuple2<String, String>(entry._2()._1(), entry._2()._1()));
|
|
|
|
|
} else {
|
|
|
|
|
map.put(entry._1(), entry._2());
|
|
|
|
|
}
|
2020-09-23 17:31:15 +02:00
|
|
|
|
} 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);
|
|
|
|
|
}
|
|
|
|
|
}
|
2020-10-20 17:03:25 +02:00
|
|
|
|
String current = entry._2()._1();
|
2020-09-23 17:31:15 +02:00
|
|
|
|
if (!ent.contains("Euratom")) {
|
|
|
|
|
|
|
|
|
|
String parent;
|
|
|
|
|
String tmp_key = tmp[0] + ".";
|
|
|
|
|
for (int i = 1; i < tmp.length - 1; i++) {
|
|
|
|
|
tmp_key += tmp[i] + ".";
|
2020-10-20 17:03:25 +02:00
|
|
|
|
parent = map.get(tmp_key)._1().toLowerCase().trim();
|
2020-09-23 17:31:15 +02:00
|
|
|
|
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);
|
2020-10-01 11:51:11 +02:00
|
|
|
|
if (current.trim().charAt(0) == '-' || current.trim().charAt(0) == '–') {
|
2020-09-23 17:31:15 +02:00
|
|
|
|
current = current.trim().substring(1).trim();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
}
|
2020-10-20 17:03:25 +02:00
|
|
|
|
String shortTitle = entry._2()._2();
|
|
|
|
|
if (StringUtils.isEmpty(shortTitle)) {
|
|
|
|
|
shortTitle = current;
|
|
|
|
|
}
|
|
|
|
|
Tuple2<String, String> newEntry = new Tuple2<>(map.get(key)._1() + " | " + current,
|
|
|
|
|
map.get(key)._2() + " | " + shortTitle);
|
|
|
|
|
map.put(ent + ".", newEntry);
|
2020-09-28 12:06:56 +02:00
|
|
|
|
|
2020-09-23 17:31:15 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
}
|
2020-10-20 17:03:25 +02:00
|
|
|
|
return h2020Programmes.map(csvProgramme -> {
|
|
|
|
|
|
|
|
|
|
String code = csvProgramme.getCode();
|
|
|
|
|
if (!code.endsWith(".") && !code.contains("Euratom")
|
|
|
|
|
&& !code.equals("H2020-EC"))
|
|
|
|
|
code += ".";
|
|
|
|
|
|
|
|
|
|
csvProgramme.setClassification(map.get(code)._1());
|
|
|
|
|
csvProgramme.setClassification_short(map.get(code)._2());
|
|
|
|
|
|
|
|
|
|
return csvProgramme;
|
|
|
|
|
}).collect();
|
2020-05-19 18:42:50 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
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));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
}
|