2020-04-08 13:04:04 +02:00
|
|
|
package eu.dnetlib.dhp.countrypropagation;
|
|
|
|
|
|
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.*;
|
|
|
|
import org.apache.commons.io.IOUtils;
|
|
|
|
import org.apache.hadoop.fs.FileSystem;
|
|
|
|
import org.apache.spark.SparkConf;
|
|
|
|
import org.apache.spark.api.java.JavaSparkContext;
|
|
|
|
import org.apache.spark.sql.Dataset;
|
|
|
|
import org.apache.spark.sql.Encoders;
|
|
|
|
import org.apache.spark.sql.SparkSession;
|
|
|
|
import org.apache.spark.sql.SaveMode;
|
|
|
|
|
|
|
|
import java.util.Arrays;
|
|
|
|
import java.util.List;
|
2020-04-08 18:08:41 +02:00
|
|
|
|
2020-04-08 13:04:04 +02:00
|
|
|
|
|
|
|
import static eu.dnetlib.dhp.PropagationConstant.createOutputDirs;
|
|
|
|
import static eu.dnetlib.dhp.PropagationConstant.getConstraintList;
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* For the association of the country to the datasource
|
|
|
|
* The association is computed only for datasource of specific type or having whitelisted ids
|
|
|
|
* The country is registered in the Organization associated to the Datasource, so the
|
|
|
|
* relation provides between Datasource and Organization is exploited to get the country for the datasource
|
|
|
|
*/
|
|
|
|
|
|
|
|
public class PrepareResultCountryAssociation {
|
|
|
|
private static final Logger log = LoggerFactory.getLogger(PrepareResultCountryAssociation.class);
|
|
|
|
|
|
|
|
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
|
|
|
|
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
|
|
|
|
|
|
String jsonConfiguration = IOUtils.toString(PrepareResultCountryAssociation.class
|
2020-04-08 18:08:41 +02:00
|
|
|
.getResourceAsStream("/eu/dnetlib/dhp/countrypropagation/input_prepare_dc_assoc.json"));
|
2020-04-08 13:04:04 +02:00
|
|
|
|
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
|
|
|
jsonConfiguration);
|
|
|
|
|
|
|
|
parser.parseArgument(args);
|
|
|
|
|
|
|
|
String inputPath = parser.get("sourcePath");
|
|
|
|
log.info("inputPath: {}", inputPath);
|
|
|
|
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
conf.set("hive.metastore.uris", parser.get("hive_metastore_uris"));
|
|
|
|
|
|
|
|
final SparkSession spark = SparkSession
|
|
|
|
.builder()
|
2020-04-08 18:08:41 +02:00
|
|
|
.appName(SparkCountryPropagationJob.class.getSimpleName())
|
2020-04-08 13:04:04 +02:00
|
|
|
.master(parser.get("master"))
|
|
|
|
.config(conf)
|
|
|
|
.enableHiveSupport()
|
|
|
|
.getOrCreate();
|
|
|
|
|
2020-04-08 18:08:41 +02:00
|
|
|
|
2020-04-08 13:04:04 +02:00
|
|
|
final String outputPath = "/tmp/provision/propagation/countrytoresultfrominstitutionalrepositories";
|
|
|
|
|
2020-04-08 18:08:41 +02:00
|
|
|
|
2020-04-08 13:04:04 +02:00
|
|
|
createOutputDirs(outputPath, FileSystem.get(spark.sparkContext().hadoopConfiguration()));
|
2020-04-08 18:08:41 +02:00
|
|
|
prepareDatasourceCountryAssociation(spark,
|
|
|
|
Arrays.asList(parser.get("whitelist").split(";")),
|
|
|
|
Arrays.asList(parser.get("allowedtypes").split(";")),
|
|
|
|
inputPath,
|
|
|
|
outputPath);
|
|
|
|
|
2020-04-08 13:04:04 +02:00
|
|
|
|
2020-04-08 18:08:41 +02:00
|
|
|
}
|
2020-04-08 13:04:04 +02:00
|
|
|
|
2020-04-08 18:08:41 +02:00
|
|
|
private static void prepareDatasourceCountryAssociation(SparkSession spark,
|
|
|
|
List<String> whitelist,
|
|
|
|
List<String> allowedtypes,
|
|
|
|
String inputPath,
|
|
|
|
String outputPath) {
|
2020-04-08 13:04:04 +02:00
|
|
|
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")
|
2020-04-08 18:08:41 +02:00
|
|
|
.map(item -> OBJECT_MAPPER.readValue(item, Datasource.class)).rdd(), Encoders.bean(Datasource.class));
|
2020-04-08 13:04:04 +02:00
|
|
|
|
|
|
|
Dataset<Relation> relation = spark.createDataset(sc.textFile(inputPath + "/relation")
|
2020-04-08 18:08:41 +02:00
|
|
|
.map(item -> OBJECT_MAPPER.readValue(item, Relation.class)).rdd(), Encoders.bean(Relation.class));
|
2020-04-08 13:04:04 +02:00
|
|
|
|
|
|
|
Dataset<Organization> organization = spark.createDataset(sc.textFile(inputPath + "/organization")
|
2020-04-08 18:08:41 +02:00
|
|
|
.map(item -> OBJECT_MAPPER.readValue(item, Organization.class)).rdd(), Encoders.bean(Organization.class));
|
2020-04-08 13:04:04 +02:00
|
|
|
|
|
|
|
datasource.createOrReplaceTempView("datasource");
|
|
|
|
relation.createOrReplaceTempView("relation");
|
|
|
|
organization.createOrReplaceTempView("organization");
|
|
|
|
|
|
|
|
String query = "SELECT source ds, country.classid country " +
|
|
|
|
"FROM ( SELECT id " +
|
2020-04-08 18:08:41 +02:00
|
|
|
" FROM datasource " +
|
|
|
|
" WHERE (datainfo.deletedbyinference = false " + whitelisted + ") " +
|
|
|
|
getConstraintList("datasourcetype.classid = '", allowedtypes) + ") d " +
|
2020-04-08 13:04:04 +02:00
|
|
|
"JOIN ( SELECT source, target " +
|
2020-04-08 18:08:41 +02:00
|
|
|
" FROM relation " +
|
|
|
|
" WHERE relclass = 'provides' " +
|
|
|
|
" AND datainfo.deletedbyinference = false ) rel " +
|
2020-04-08 13:04:04 +02:00
|
|
|
"ON d.id = rel.source " +
|
|
|
|
"JOIN (SELECT id, country " +
|
2020-04-08 18:08:41 +02:00
|
|
|
" FROM organization " +
|
|
|
|
" WHERE datainfo.deletedbyinference = false " +
|
|
|
|
" AND length(country.classid)>0) o " +
|
2020-04-08 13:04:04 +02:00
|
|
|
"ON o.id = rel.target";
|
|
|
|
|
|
|
|
spark.sql(query)
|
|
|
|
.as(Encoders.bean(DatasourceCountry.class))
|
|
|
|
.write()
|
|
|
|
.mode(SaveMode.Overwrite)
|
|
|
|
.parquet(outputPath + "/prepared_datasource_country");
|
|
|
|
|
|
|
|
|
|
|
|
}
|
2020-04-08 18:08:41 +02:00
|
|
|
|
|
|
|
|
2020-04-08 13:04:04 +02:00
|
|
|
}
|