Merge pull request 'graph cleaning refactoring' (#282) from graph_cleaning_refactoring into beta

Reviewed-on: D-Net/dnet-hadoop#282
This commit is contained in:
Claudio Atzori 2023-05-02 10:40:02 +02:00
commit cdd33f7445
22 changed files with 1481 additions and 1769 deletions

View File

@ -16,6 +16,8 @@ import org.apache.commons.lang3.StringUtils;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Encoders;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.github.sisyphsu.dateparser.DateParserUtils;
import com.google.common.collect.Lists;
import com.google.common.collect.Sets;
@ -38,6 +40,127 @@ public class GraphCleaningFunctions extends CleaningFunctions {
public static final int TITLE_FILTER_RESIDUAL_LENGTH = 5;
public static <T extends Oaf> T cleanContext(T value, String contextId, String verifyParam) {
if (ModelSupport.isSubClass(value, Result.class)) {
final Result res = (Result) value;
if (shouldCleanContext(res, verifyParam)) {
res
.setContext(
res
.getContext()
.stream()
.filter(c -> !StringUtils.startsWith(c.getId().toLowerCase(), contextId))
.collect(Collectors.toList()));
}
return (T) res;
} else {
return value;
}
}
private static boolean shouldCleanContext(Result res, String verifyParam) {
boolean titleMatch = res
.getTitle()
.stream()
.filter(
t -> t
.getQualifier()
.getClassid()
.equalsIgnoreCase(ModelConstants.MAIN_TITLE_QUALIFIER.getClassid()))
.anyMatch(t -> t.getValue().toLowerCase().startsWith(verifyParam.toLowerCase()));
return titleMatch && Objects.nonNull(res.getContext());
}
public static <T extends Oaf> T cleanCountry(T value, String[] verifyParam, Set<String> hostedBy,
String collectedfrom, String country) {
if (ModelSupport.isSubClass(value, Result.class)) {
final Result res = (Result) value;
if (res.getInstance().stream().anyMatch(i -> hostedBy.contains(i.getHostedby().getKey())) ||
!res.getCollectedfrom().stream().anyMatch(cf -> cf.getValue().equals(collectedfrom))) {
return (T) res;
}
List<StructuredProperty> ids = getPidsAndAltIds(res).collect(Collectors.toList());
if (ids
.stream()
.anyMatch(
p -> p
.getQualifier()
.getClassid()
.equals(PidType.doi.toString()) && pidInParam(p.getValue(), verifyParam))) {
res
.setCountry(
res
.getCountry()
.stream()
.filter(
c -> toTakeCountry(c, country))
.collect(Collectors.toList()));
}
return (T) res;
} else {
return value;
}
}
private static <T extends Result> Stream<StructuredProperty> getPidsAndAltIds(T r) {
final Stream<StructuredProperty> resultPids = Optional
.ofNullable(r.getPid())
.map(Collection::stream)
.orElse(Stream.empty());
final Stream<StructuredProperty> instancePids = Optional
.ofNullable(r.getInstance())
.map(
instance -> instance
.stream()
.flatMap(
i -> Optional
.ofNullable(i.getPid())
.map(Collection::stream)
.orElse(Stream.empty())))
.orElse(Stream.empty());
final Stream<StructuredProperty> instanceAltIds = Optional
.ofNullable(r.getInstance())
.map(
instance -> instance
.stream()
.flatMap(
i -> Optional
.ofNullable(i.getAlternateIdentifier())
.map(Collection::stream)
.orElse(Stream.empty())))
.orElse(Stream.empty());
return Stream
.concat(
Stream.concat(resultPids, instancePids),
instanceAltIds);
}
private static boolean pidInParam(String value, String[] verifyParam) {
for (String s : verifyParam)
if (value.startsWith(s))
return true;
return false;
}
private static boolean toTakeCountry(Country c, String country) {
// If dataInfo is not set, or dataInfo.inferenceprovenance is not set or not present then it cannot be
// inserted via propagation
if (!Optional.ofNullable(c.getDataInfo()).isPresent())
return true;
if (!Optional.ofNullable(c.getDataInfo().getInferenceprovenance()).isPresent())
return true;
return !(c
.getClassid()
.equalsIgnoreCase(country) &&
c.getDataInfo().getInferenceprovenance().equals("propagation"));
}
public static <T extends Oaf> T fixVocabularyNames(T value) {
if (value instanceof Datasource) {
// nothing to clean here

View File

@ -1,122 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.io.Serializable;
import java.util.Optional;
import java.util.stream.Collectors;
import org.apache.commons.io.IOUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SaveMode;
import org.apache.spark.sql.SparkSession;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Result;
public class CleanContextSparkJob implements Serializable {
private static final Logger log = LoggerFactory.getLogger(CleanContextSparkJob.class);
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
public static void main(String[] args) throws Exception {
String jsonConfiguration = IOUtils
.toString(
CleanContextSparkJob.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/input_clean_context_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);
String inputPath = parser.get("inputPath");
log.info("inputPath: {}", inputPath);
String workingDir = parser.get("workingDir");
log.info("workingDir: {}", workingDir);
String contextId = parser.get("contextId");
log.info("contextId: {}", contextId);
String verifyParam = parser.get("verifyParam");
log.info("verifyParam: {}", verifyParam);
String graphTableClassName = parser.get("graphTableClassName");
log.info("graphTableClassName: {}", graphTableClassName);
Class<? extends Result> entityClazz = (Class<? extends Result>) Class.forName(graphTableClassName);
SparkConf conf = new SparkConf();
runWithSparkSession(
conf,
isSparkSessionManaged,
spark -> {
cleanContext(spark, contextId, verifyParam, inputPath, entityClazz, workingDir);
});
}
private static <T extends Result> void cleanContext(SparkSession spark, String contextId, String verifyParam,
String inputPath, Class<T> entityClazz, String workingDir) {
Dataset<T> res = spark
.read()
.textFile(inputPath)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, entityClazz),
Encoders.bean(entityClazz));
res.map((MapFunction<T, T>) r -> {
if (!r
.getTitle()
.stream()
.filter(
t -> t
.getQualifier()
.getClassid()
.equalsIgnoreCase(ModelConstants.MAIN_TITLE_QUALIFIER.getClassid()))
.anyMatch(t -> t.getValue().toLowerCase().startsWith(verifyParam.toLowerCase()))) {
return r;
}
r
.setContext(
r
.getContext()
.stream()
.filter(
c -> !c.getId().split("::")[0]
.equalsIgnoreCase(contextId))
.collect(Collectors.toList()));
return r;
}, Encoders.bean(entityClazz))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(workingDir);
spark
.read()
.textFile(workingDir)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, entityClazz),
Encoders.bean(entityClazz))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(inputPath);
}
}

View File

@ -3,12 +3,16 @@ package eu.dnetlib.dhp.oa.graph.clean;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.Optional;
import java.util.*;
import java.util.stream.Stream;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.FilterFunction;
import org.apache.spark.api.java.function.FlatMapFunction;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.api.java.function.MapGroupsFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SaveMode;
@ -17,15 +21,22 @@ import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.common.collect.Sets;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.common.action.model.MasterDuplicate;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelSupport;
import eu.dnetlib.dhp.schema.oaf.KeyValue;
import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.schema.oaf.OafEntity;
import eu.dnetlib.dhp.schema.oaf.Result;
import eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions;
import eu.dnetlib.dhp.utils.ISLookupClientFactory;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import scala.Tuple2;
public class CleanGraphSparkJob {
@ -33,8 +44,13 @@ public class CleanGraphSparkJob {
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
public static void main(String[] args) throws Exception {
private ArgumentApplicationParser parser;
public CleanGraphSparkJob(ArgumentApplicationParser parser) {
this.parser = parser;
}
public static void main(String[] args) throws Exception {
String jsonConfiguration = IOUtils
.toString(
CleanGraphSparkJob.class
@ -49,30 +65,70 @@ public class CleanGraphSparkJob {
.orElse(Boolean.TRUE);
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
String isLookupUrl = parser.get("isLookupUrl");
log.info("isLookupUrl: {}", isLookupUrl);
ISLookUpService isLookup = ISLookupClientFactory.getLookUpService(isLookupUrl);
new CleanGraphSparkJob(parser).run(isSparkSessionManaged, isLookup);
}
public void run(Boolean isSparkSessionManaged, ISLookUpService isLookUpService)
throws ISLookUpException, ClassNotFoundException {
String inputPath = parser.get("inputPath");
log.info("inputPath: {}", inputPath);
String outputPath = parser.get("outputPath");
log.info("outputPath: {}", outputPath);
String isLookupUrl = parser.get("isLookupUrl");
log.info("isLookupUrl: {}", isLookupUrl);
String graphTableClassName = parser.get("graphTableClassName");
log.info("graphTableClassName: {}", graphTableClassName);
String contextId = parser.get("contextId");
log.info("contextId: {}", contextId);
String verifyParam = parser.get("verifyParam");
log.info("verifyParam: {}", verifyParam);
String datasourcePath = parser.get("hostedBy");
log.info("datasourcePath: {}", datasourcePath);
String country = parser.get("country");
log.info("country: {}", country);
String[] verifyCountryParam = Optional
.ofNullable(parser.get("verifyCountryParam"))
.map(s -> s.split(";"))
.orElse(new String[] {});
log.info("verifyCountryParam: {}", verifyCountryParam);
String collectedfrom = parser.get("collectedfrom");
log.info("collectedfrom: {}", collectedfrom);
String dsMasterDuplicatePath = parser.get("masterDuplicatePath");
log.info("masterDuplicatePath: {}", dsMasterDuplicatePath);
Boolean deepClean = Optional
.ofNullable(parser.get("deepClean"))
.map(Boolean::valueOf)
.orElse(Boolean.FALSE);
log.info("deepClean: {}", deepClean);
Class<? extends OafEntity> entityClazz = (Class<? extends OafEntity>) Class.forName(graphTableClassName);
final ISLookUpService isLookupService = ISLookupClientFactory.getLookUpService(isLookupUrl);
final VocabularyGroup vocs = VocabularyGroup.loadVocsFromIS(isLookupService);
final VocabularyGroup vocs = VocabularyGroup.loadVocsFromIS(isLookUpService);
SparkConf conf = new SparkConf();
conf.setAppName(CleanGraphSparkJob.class.getSimpleName() + "#" + entityClazz.getSimpleName());
runWithSparkSession(
conf,
isSparkSessionManaged,
spark -> {
HdfsSupport.remove(outputPath, spark.sparkContext().hadoopConfiguration());
cleanGraphTable(spark, vocs, inputPath, entityClazz, outputPath);
cleanGraphTable(
spark, vocs, inputPath, entityClazz, outputPath, contextId, verifyParam, datasourcePath, country,
verifyCountryParam, collectedfrom, dsMasterDuplicatePath, deepClean);
});
}
@ -81,21 +137,88 @@ public class CleanGraphSparkJob {
VocabularyGroup vocs,
String inputPath,
Class<T> clazz,
String outputPath) {
String outputPath, String contextId, String verifyParam, String datasourcePath, String country,
String[] verifyCountryParam, String collectedfrom, String dsMasterDuplicatePath,
Boolean deepClean) {
final CleaningRuleMap mapping = CleaningRuleMap.create(vocs);
readTableFromPath(spark, inputPath, clazz)
final Dataset<T> cleaned_basic = readTableFromPath(spark, inputPath, clazz)
.map((MapFunction<T, T>) GraphCleaningFunctions::fixVocabularyNames, Encoders.bean(clazz))
.map((MapFunction<T, T>) value -> OafCleaner.apply(value, mapping), Encoders.bean(clazz))
.map((MapFunction<T, T>) value -> GraphCleaningFunctions.cleanup(value, vocs), Encoders.bean(clazz))
.filter((FilterFunction<T>) GraphCleaningFunctions::filter)
.filter((FilterFunction<T>) GraphCleaningFunctions::filter);
// read the master-duplicate tuples
Dataset<MasterDuplicate> md = spark
.read()
.textFile(dsMasterDuplicatePath)
.map(as(MasterDuplicate.class), Encoders.bean(MasterDuplicate.class));
// prepare the resolved CF|HB references with the corresponding EMPTY master ID
Dataset<IdCfHbMapping> resolved = spark
.read()
.textFile(inputPath)
.map(as(clazz), Encoders.bean(clazz))
.flatMap(flattenCfHbFn(), Encoders.bean(IdCfHbMapping.class));
if (Boolean.FALSE.equals(deepClean)) {
if (Boolean.TRUE.equals(ModelSupport.isSubClass(clazz, Result.class))) {
save(fixCFHB(clazz, cleaned_basic, md, resolved), outputPath);
} else {
save(cleaned_basic, outputPath);
}
} else if (Boolean.TRUE.equals(ModelSupport.isSubClass(clazz, Result.class))) {
// load the hostedby mapping
Set<String> hostedBy = Sets
.newHashSet(
spark
.read()
.textFile(datasourcePath)
.collectAsList());
// perform the deep cleaning steps
final Dataset<T> cleaned_deep = fixCFHB(clazz, cleaned_basic, md, resolved)
.map(
(MapFunction<T, T>) value -> GraphCleaningFunctions.cleanContext(value, contextId, verifyParam),
Encoders.bean(clazz))
.map(
(MapFunction<T, T>) value -> GraphCleaningFunctions
.cleanCountry(value, verifyCountryParam, hostedBy, collectedfrom, country),
Encoders.bean(clazz));
save(cleaned_deep, outputPath);
} else {
save(cleaned_basic, outputPath);
}
}
private static <T extends Oaf> void save(final Dataset<T> dataset, final String outputPath) {
dataset
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(outputPath);
}
private static <T extends Oaf> Dataset<T> fixCFHB(Class<T> clazz, Dataset<T> results, Dataset<MasterDuplicate> md,
Dataset<IdCfHbMapping> resolved) {
// set the EMPTY master ID/NAME
Dataset<IdCfHbMapping> resolvedDs = resolved
.joinWith(md, resolved.col("cfhb").equalTo(md.col("duplicateId")))
.map(asIdCfHbMapping(), Encoders.bean(IdCfHbMapping.class))
.filter((FilterFunction<IdCfHbMapping>) m -> Objects.nonNull(m.getMasterId()));
return results
.joinWith(resolvedDs, results.col("id").equalTo(resolvedDs.col("resultId")), "left")
.groupByKey(
(MapFunction<Tuple2<T, IdCfHbMapping>, String>) t -> ((Result) t._1()).getId(), Encoders.STRING())
.mapGroups(getMapGroupsFunction(), Encoders.bean(clazz));
}
private static <T extends Oaf> Dataset<T> readTableFromPath(
SparkSession spark, String inputEntityPath, Class<T> clazz) {
@ -103,9 +226,104 @@ public class CleanGraphSparkJob {
return spark
.read()
.textFile(inputEntityPath)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, clazz),
Encoders.bean(clazz));
.map(as(clazz), Encoders.bean(clazz));
}
private static <R> MapFunction<String, R> as(Class<R> clazz) {
return s -> OBJECT_MAPPER.readValue(s, clazz);
}
private static <T extends Oaf> FlatMapFunction<T, IdCfHbMapping> flattenCfHbFn() {
return r -> Stream
.concat(
Optional
.ofNullable(r.getCollectedfrom())
.map(cf -> cf.stream().map(KeyValue::getKey))
.orElse(Stream.empty()),
Stream
.concat(
Optional
.ofNullable(((Result) r).getInstance())
.map(
instances -> instances
.stream()
.map(i -> Optional.ofNullable(i.getHostedby()).map(KeyValue::getKey).orElse("")))
.orElse(Stream.empty())
.filter(StringUtils::isNotBlank),
Optional
.ofNullable(((Result) r).getInstance())
.map(
instances -> instances
.stream()
.map(
i -> Optional
.ofNullable(i.getCollectedfrom())
.map(KeyValue::getKey)
.orElse("")))
.orElse(Stream.empty())
.filter(StringUtils::isNotBlank)))
.distinct()
.filter(StringUtils::isNotBlank)
.map(cfHb -> asIdCfHbMapping(((Result) r).getId(), cfHb))
.iterator();
}
private static MapFunction<Tuple2<IdCfHbMapping, MasterDuplicate>, IdCfHbMapping> asIdCfHbMapping() {
return t -> {
final IdCfHbMapping mapping = t._1();
Optional
.ofNullable(t._2())
.ifPresent(t2 -> {
mapping.setMasterId(t2.getMasterId());
mapping.setMasterName(t2.getMasterName());
});
return mapping;
};
}
private static IdCfHbMapping asIdCfHbMapping(String resultId, String cfHb) {
IdCfHbMapping m = new IdCfHbMapping(resultId);
m.setCfhb(cfHb);
return m;
}
private static <T extends Oaf> MapGroupsFunction<String, Tuple2<T, IdCfHbMapping>, T> getMapGroupsFunction() {
return new MapGroupsFunction<String, Tuple2<T, IdCfHbMapping>, T>() {
@Override
public T call(String key, Iterator<Tuple2<T, IdCfHbMapping>> values) {
final Tuple2<T, IdCfHbMapping> first = values.next();
final T res = first._1();
updateResult(res, first._2());
values.forEachRemaining(t -> updateResult(res, t._2()));
return res;
}
private void updateResult(T res, IdCfHbMapping m) {
if (Objects.nonNull(m)) {
filter(res.getCollectedfrom()).forEach(kv -> updateKeyValue(kv, m));
((Result) res).getInstance().forEach(i -> {
updateKeyValue(i.getHostedby(), m);
updateKeyValue(i.getCollectedfrom(), m);
});
}
}
private Stream<KeyValue> filter(List<KeyValue> kvs) {
return kvs
.stream()
.filter(kv -> StringUtils.isNotBlank(kv.getKey()) && StringUtils.isNotBlank(kv.getValue()));
}
private void updateKeyValue(final KeyValue kv, final IdCfHbMapping a) {
if (Objects.nonNull(kv) && Objects.nonNull(kv.getKey()) && kv.getKey().equals(a.getCfhb())) {
kv.setKey(a.getMasterId());
kv.setValue(a.getMasterName());
}
}
};
}
}

View File

@ -1,10 +1,9 @@
package eu.dnetlib.dhp.oa.graph.clean.country;
package eu.dnetlib.dhp.oa.graph.clean;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.io.Serializable;
import java.util.List;
import java.util.Optional;
import org.apache.commons.io.IOUtils;
@ -21,7 +20,6 @@ import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import scala.Tuple2;

View File

@ -1,5 +1,5 @@
package eu.dnetlib.dhp.oa.graph.clean.cfhb;
package eu.dnetlib.dhp.oa.graph.clean;
import java.io.Serializable;

View File

@ -1,227 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean.cfhb;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.Iterator;
import java.util.List;
import java.util.Objects;
import java.util.Optional;
import java.util.stream.Collectors;
import java.util.stream.Stream;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.FilterFunction;
import org.apache.spark.api.java.function.FlatMapFunction;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.api.java.function.MapGroupsFunction;
import org.apache.spark.sql.*;
import org.apache.spark.sql.expressions.Aggregator;
import org.jetbrains.annotations.NotNull;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.common.action.model.MasterDuplicate;
import eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob;
import eu.dnetlib.dhp.schema.oaf.Instance;
import eu.dnetlib.dhp.schema.oaf.KeyValue;
import eu.dnetlib.dhp.schema.oaf.Result;
import eu.dnetlib.dhp.utils.DHPUtils;
import scala.Tuple2;
public class CleanCfHbSparkJob {
private static final Logger log = LoggerFactory.getLogger(CleanCfHbSparkJob.class);
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
public static void main(String[] args) throws Exception {
String jsonConfiguration = IOUtils
.toString(
CleanCountrySparkJob.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/input_clean_cfhb_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);
String inputPath = parser.get("inputPath");
log.info("inputPath: {}", inputPath);
String resolvedPath = parser.get("resolvedPath");
log.info("resolvedPath: {}", resolvedPath);
String outputPath = parser.get("outputPath");
log.info("outputPath: {}", outputPath);
String dsMasterDuplicatePath = parser.get("masterDuplicatePath");
log.info("masterDuplicatePath: {}", dsMasterDuplicatePath);
String graphTableClassName = parser.get("graphTableClassName");
log.info("graphTableClassName: {}", graphTableClassName);
Class<? extends Result> entityClazz = (Class<? extends Result>) Class.forName(graphTableClassName);
SparkConf conf = new SparkConf();
runWithSparkSession(
conf,
isSparkSessionManaged,
spark -> {
HdfsSupport.remove(outputPath, spark.sparkContext().hadoopConfiguration());
HdfsSupport.remove(resolvedPath, spark.sparkContext().hadoopConfiguration());
cleanCfHb(
spark, inputPath, entityClazz, resolvedPath, dsMasterDuplicatePath, outputPath);
});
}
private static <T extends Result> void cleanCfHb(SparkSession spark, String inputPath, Class<T> entityClazz,
String resolvedPath, String masterDuplicatePath, String outputPath) {
// read the master-duplicate tuples
Dataset<MasterDuplicate> md = spark
.read()
.textFile(masterDuplicatePath)
.map(as(MasterDuplicate.class), Encoders.bean(MasterDuplicate.class));
// prepare the resolved CF|HB references with the corresponding EMPTY master ID
Dataset<IdCfHbMapping> resolved = spark
.read()
.textFile(inputPath)
.map(as(entityClazz), Encoders.bean(entityClazz))
.flatMap(flattenCfHbFn(), Encoders.bean(IdCfHbMapping.class));
// set the EMPTY master ID/NAME and save it
resolved
.joinWith(md, resolved.col("cfhb").equalTo(md.col("duplicateId")))
.map(asIdCfHbMapping(), Encoders.bean(IdCfHbMapping.class))
.filter((FilterFunction<IdCfHbMapping>) m -> Objects.nonNull(m.getMasterId()))
.write()
.mode(SaveMode.Overwrite)
.json(resolvedPath);
// read again the resolved CF|HB mapping
Dataset<IdCfHbMapping> resolvedDS = spark
.read()
.textFile(resolvedPath)
.map(as(IdCfHbMapping.class), Encoders.bean(IdCfHbMapping.class));
// read the result table
Dataset<T> res = spark
.read()
.textFile(inputPath)
.map(as(entityClazz), Encoders.bean(entityClazz));
// Join the results with the resolved CF|HB mapping, apply the mapping and save it
res
.joinWith(resolvedDS, res.col("id").equalTo(resolvedDS.col("resultId")), "left")
.groupByKey((MapFunction<Tuple2<T, IdCfHbMapping>, String>) t -> t._1().getId(), Encoders.STRING())
.mapGroups(getMapGroupsFunction(), Encoders.bean(entityClazz))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(outputPath);
}
private static MapFunction<Tuple2<IdCfHbMapping, MasterDuplicate>, IdCfHbMapping> asIdCfHbMapping() {
return t -> {
final IdCfHbMapping mapping = t._1();
Optional
.ofNullable(t._2())
.ifPresent(t2 -> {
mapping.setMasterId(t2.getMasterId());
mapping.setMasterName(t2.getMasterName());
});
return mapping;
};
}
private static <T extends Result> FlatMapFunction<T, IdCfHbMapping> flattenCfHbFn() {
return r -> Stream
.concat(
Optional
.ofNullable(r.getCollectedfrom())
.map(cf -> cf.stream().map(KeyValue::getKey))
.orElse(Stream.empty()),
Stream
.concat(
Optional
.ofNullable(r.getInstance())
.map(
instances -> instances
.stream()
.map(i -> Optional.ofNullable(i.getHostedby()).map(KeyValue::getKey).orElse("")))
.orElse(Stream.empty())
.filter(StringUtils::isNotBlank),
Optional
.ofNullable(r.getInstance())
.map(
instances -> instances
.stream()
.map(
i -> Optional
.ofNullable(i.getCollectedfrom())
.map(KeyValue::getKey)
.orElse("")))
.orElse(Stream.empty())
.filter(StringUtils::isNotBlank)))
.distinct()
.filter(StringUtils::isNotBlank)
.map(cfHb -> asIdCfHbMapping(r.getId(), cfHb))
.iterator();
}
private static <T extends Result> MapGroupsFunction<String, Tuple2<T, IdCfHbMapping>, T> getMapGroupsFunction() {
return new MapGroupsFunction<String, Tuple2<T, IdCfHbMapping>, T>() {
@Override
public T call(String key, Iterator<Tuple2<T, IdCfHbMapping>> values) {
final Tuple2<T, IdCfHbMapping> first = values.next();
final T res = first._1();
updateResult(res, first._2());
values.forEachRemaining(t -> updateResult(res, t._2()));
return res;
}
private void updateResult(T res, IdCfHbMapping m) {
if (Objects.nonNull(m)) {
res.getCollectedfrom().forEach(kv -> updateKeyValue(kv, m));
res.getInstance().forEach(i -> {
updateKeyValue(i.getHostedby(), m);
updateKeyValue(i.getCollectedfrom(), m);
});
}
}
private void updateKeyValue(final KeyValue kv, final IdCfHbMapping a) {
if (kv.getKey().equals(a.getCfhb())) {
kv.setKey(a.getMasterId());
kv.setValue(a.getMasterName());
}
}
};
}
private static IdCfHbMapping asIdCfHbMapping(String resultId, String cfHb) {
IdCfHbMapping m = new IdCfHbMapping(resultId);
m.setCfhb(cfHb);
return m;
}
private static <R> MapFunction<String, R> as(Class<R> clazz) {
return s -> OBJECT_MAPPER.readValue(s, clazz);
}
}

View File

@ -1,211 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean.country;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.io.Serializable;
import java.util.Collection;
import java.util.List;
import java.util.Optional;
import java.util.function.Function;
import java.util.stream.Collectors;
import java.util.stream.Stream;
import javax.swing.text.html.Option;
import org.apache.commons.io.IOUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.FilterFunction;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SaveMode;
import org.apache.spark.sql.SparkSession;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* @author miriam.baglioni
* @Date 20/07/22
*/
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob;
import eu.dnetlib.dhp.schema.oaf.Country;
import eu.dnetlib.dhp.schema.oaf.Instance;
import eu.dnetlib.dhp.schema.oaf.Result;
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
import eu.dnetlib.dhp.schema.oaf.utils.PidType;
public class CleanCountrySparkJob implements Serializable {
private static final Logger log = LoggerFactory.getLogger(CleanCountrySparkJob.class);
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
public static void main(String[] args) throws Exception {
String jsonConfiguration = IOUtils
.toString(
CleanCountrySparkJob.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/input_clean_country_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);
String inputPath = parser.get("inputPath");
log.info("inputPath: {}", inputPath);
String workingDir = parser.get("workingDir");
log.info("workingDir: {}", workingDir);
String datasourcePath = parser.get("hostedBy");
log.info("datasourcePath: {}", datasourcePath);
String country = parser.get("country");
log.info("country: {}", country);
String[] verifyParam = parser.get("verifyParam").split(";");
log.info("verifyParam: {}", verifyParam);
String collectedfrom = parser.get("collectedfrom");
log.info("collectedfrom: {}", collectedfrom);
String graphTableClassName = parser.get("graphTableClassName");
log.info("graphTableClassName: {}", graphTableClassName);
Class<? extends Result> entityClazz = (Class<? extends Result>) Class.forName(graphTableClassName);
SparkConf conf = new SparkConf();
runWithSparkSession(
conf,
isSparkSessionManaged,
spark -> {
cleanCountry(
spark, country, verifyParam, inputPath, entityClazz, workingDir, collectedfrom, datasourcePath);
});
}
private static <T extends Result> void cleanCountry(SparkSession spark, String country, String[] verifyParam,
String inputPath, Class<T> entityClazz, String workingDir, String collectedfrom, String datasourcePath) {
List<String> hostedBy = spark
.read()
.textFile(datasourcePath)
.collectAsList();
Dataset<T> res = spark
.read()
.textFile(inputPath)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, entityClazz),
Encoders.bean(entityClazz));
res.map((MapFunction<T, T>) r -> {
if (r.getInstance().stream().anyMatch(i -> hostedBy.contains(i.getHostedby().getKey())) ||
!r.getCollectedfrom().stream().anyMatch(cf -> cf.getValue().equals(collectedfrom))) {
return r;
}
List<StructuredProperty> ids = getPidsAndAltIds(r).collect(Collectors.toList());
if (ids
.stream()
.anyMatch(
p -> p
.getQualifier()
.getClassid()
.equals(PidType.doi.toString()) && pidInParam(p.getValue(), verifyParam))) {
r
.setCountry(
r
.getCountry()
.stream()
.filter(
c -> toTakeCountry(c, country))
.collect(Collectors.toList()));
}
return r;
}, Encoders.bean(entityClazz))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(workingDir);
spark
.read()
.textFile(workingDir)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, entityClazz),
Encoders.bean(entityClazz))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(inputPath);
}
private static <T extends Result> Stream<StructuredProperty> getPidsAndAltIds(T r) {
final Stream<StructuredProperty> resultPids = Optional
.ofNullable(r.getPid())
.map(Collection::stream)
.orElse(Stream.empty());
final Stream<StructuredProperty> instancePids = Optional
.ofNullable(r.getInstance())
.map(
instance -> instance
.stream()
.flatMap(
i -> Optional
.ofNullable(i.getPid())
.map(Collection::stream)
.orElse(Stream.empty())))
.orElse(Stream.empty());
final Stream<StructuredProperty> instanceAltIds = Optional
.ofNullable(r.getInstance())
.map(
instance -> instance
.stream()
.flatMap(
i -> Optional
.ofNullable(i.getAlternateIdentifier())
.map(Collection::stream)
.orElse(Stream.empty())))
.orElse(Stream.empty());
return Stream
.concat(
Stream.concat(resultPids, instancePids),
instanceAltIds);
}
private static boolean pidInParam(String value, String[] verifyParam) {
for (String s : verifyParam)
if (value.startsWith(s))
return true;
return false;
}
private static boolean toTakeCountry(Country c, String country) {
// If dataInfo is not set, or dataInfo.inferenceprovenance is not set or not present then it cannot be
// inserted via propagation
if (!Optional.ofNullable(c.getDataInfo()).isPresent())
return true;
if (!Optional.ofNullable(c.getDataInfo().getInferenceprovenance()).isPresent())
return true;
return !(c
.getClassid()
.equalsIgnoreCase(country) &&
c.getDataInfo().getInferenceprovenance().equals("propagation"));
}
}

View File

@ -83,12 +83,57 @@
</property>
</parameters>
<start to="fork_clean_graph"/>
<start to="prepare_info"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<fork name="prepare_info">
<path start="select_datasourceId_from_country"/>
<path start="get_ds_master_duplicate"/>
</fork>
<action name="select_datasourceId_from_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Select datasource ID from country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.GetDatasourceFromCountry</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=10000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--country</arg><arg>${country}</arg>
</spark>
<ok to="wait_prepare"/>
<error to="Kill"/>
</action>
<action name="get_ds_master_duplicate">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.clean.MasterDuplicateAction</main-class>
<arg>--postgresUrl</arg><arg>${postgresURL}</arg>
<arg>--postgresUser</arg><arg>${postgresUser}</arg>
<arg>--postgresPassword</arg><arg>${postgresPassword}</arg>
<arg>--hdfsPath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
</java>
<ok to="wait_prepare"/>
<error to="Kill"/>
</action>
<join name="wait_prepare" to="fork_clean_graph"/>
<fork name="fork_clean_graph">
<path start="clean_publication"/>
<path start="clean_dataset"/>
@ -115,12 +160,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=15000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/publication</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/publication</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -141,12 +194,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=8000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/dataset</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/dataset</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -167,12 +228,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=5000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/otherresearchproduct</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -193,12 +262,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=2000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/software</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/software</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -219,12 +296,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=1000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/datasource</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/datasource</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Datasource</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -245,12 +330,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=1000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/organization</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/organization</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Organization</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -271,12 +364,20 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=2000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/project</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/project</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Project</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
@ -297,486 +398,26 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
--conf spark.sql.shuffle.partitions=20000
</spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/relation</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/relation</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Relation</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyCountryParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--deepClean</arg><arg>${shouldClean}</arg>
</spark>
<ok to="wait_clean"/>
<error to="Kill"/>
</action>
<join name="wait_clean" to="clean_context"/>
<decision name="clean_context">
<switch>
<case to="fork_clean_context">${wf:conf('shouldClean') eq true}</case>
<default to="End"/>
</switch>
</decision>
<fork name="fork_clean_context">
<path start="clean_publication_context"/>
<path start="clean_dataset_context"/>
<path start="clean_otherresearchproduct_context"/>
<path start="clean_software_context"/>
</fork>
<action name="clean_publication_context">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean publications context</name>
<class>eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/publication</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/publication</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
</spark>
<ok to="wait_clean_context"/>
<error to="Kill"/>
</action>
<action name="clean_dataset_context">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean datasets Context</name>
<class>eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/dataset</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/dataset</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
</spark>
<ok to="wait_clean_context"/>
<error to="Kill"/>
</action>
<action name="clean_otherresearchproduct_context">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean otherresearchproducts context</name>
<class>eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/otherresearchproduct</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
</spark>
<ok to="wait_clean_context"/>
<error to="Kill"/>
</action>
<action name="clean_software_context">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean softwares context</name>
<class>eu.dnetlib.dhp.oa.graph.clean.CleanContextSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/software</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/software</arg>
<arg>--contextId</arg><arg>${contextId}</arg>
<arg>--verifyParam</arg><arg>${verifyParam}</arg>
</spark>
<ok to="wait_clean_context"/>
<error to="Kill"/>
</action>
<join name="wait_clean_context" to="select_datasourceId_from_country"/>
<action name="select_datasourceId_from_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Select datasource ID from country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.country.GetDatasourceFromCountry</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--country</arg><arg>${country}</arg>
</spark>
<ok to="fork_clean_country"/>
<error to="Kill"/>
</action>
<fork name="fork_clean_country">
<path start="clean_publication_country"/>
<path start="clean_dataset_country"/>
<path start="clean_otherresearchproduct_country"/>
<path start="clean_software_country"/>
</fork>
<action name="clean_publication_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean publication country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/publication</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/publication</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
</spark>
<ok to="wait_clean_country"/>
<error to="Kill"/>
</action>
<action name="clean_dataset_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean dataset country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/dataset</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/dataset</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
</spark>
<ok to="wait_clean_country"/>
<error to="Kill"/>
</action>
<action name="clean_otherresearchproduct_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean otherresearchproduct country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/otherresearchproduct</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
</spark>
<ok to="wait_clean_country"/>
<error to="Kill"/>
</action>
<action name="clean_software_country">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Clean software country</name>
<class>eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/software</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/software</arg>
<arg>--country</arg><arg>${country}</arg>
<arg>--verifyParam</arg><arg>${verifyCountryParam}</arg>
<arg>--hostedBy</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
</spark>
<ok to="wait_clean_country"/>
<error to="Kill"/>
</action>
<join name="wait_clean_country" to="should_patch_datasource_ids"/>
<decision name="should_patch_datasource_ids">
<switch>
<case to="get_ds_master_duplicate">${wf:conf('shouldClean') eq true}</case>
<default to="End"/>
</switch>
</decision>
<action name="get_ds_master_duplicate">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.clean.MasterDuplicateAction</main-class>
<arg>--postgresUrl</arg><arg>${postgresURL}</arg>
<arg>--postgresUser</arg><arg>${postgresUser}</arg>
<arg>--postgresPassword</arg><arg>${postgresPassword}</arg>
<arg>--hdfsPath</arg><arg>${workingDir}/masterduplicate</arg>
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
</java>
<ok to="fork_patch_cfhb"/>
<error to="Kill"/>
</action>
<fork name="fork_patch_cfhb">
<path start="patch_publication_cfhb"/>
<path start="patch_dataset_cfhb"/>
<path start="patch_otherresearchproduct_cfhb"/>
<path start="patch_software_cfhb"/>
</fork>
<action name="patch_publication_cfhb">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>patch publication cfhb</name>
<class>eu.dnetlib.dhp.oa.graph.clean.cfhb.CleanCfHbSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/publication</arg>
<arg>--resolvedPath</arg><arg>${workingDir}/cfHbResolved/publication</arg>
<arg>--outputPath</arg><arg>${workingDir}/cfHbPatched/publication</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
</spark>
<ok to="wait_clean_cfhb"/>
<error to="Kill"/>
</action>
<action name="patch_dataset_cfhb">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>patch dataset cfhb</name>
<class>eu.dnetlib.dhp.oa.graph.clean.cfhb.CleanCfHbSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/dataset</arg>
<arg>--resolvedPath</arg><arg>${workingDir}/cfHbResolved/dataset</arg>
<arg>--outputPath</arg><arg>${workingDir}/cfHbPatched/dataset</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
</spark>
<ok to="wait_clean_cfhb"/>
<error to="Kill"/>
</action>
<action name="patch_otherresearchproduct_cfhb">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>patch otherresearchproduct cfhb</name>
<class>eu.dnetlib.dhp.oa.graph.clean.cfhb.CleanCfHbSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
<arg>--resolvedPath</arg><arg>${workingDir}/cfHbResolved/otherresearchproduct</arg>
<arg>--outputPath</arg><arg>${workingDir}/cfHbPatched/otherresearchproduct</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
</spark>
<ok to="wait_clean_cfhb"/>
<error to="Kill"/>
</action>
<action name="patch_software_cfhb">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>patch software cfhb</name>
<class>eu.dnetlib.dhp.oa.graph.clean.cfhb.CleanCfHbSparkJob</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-cores=${sparkExecutorCores}
--executor-memory=${sparkExecutorMemory}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680
</spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}/software</arg>
<arg>--resolvedPath</arg><arg>${workingDir}/cfHbResolved/software</arg>
<arg>--outputPath</arg><arg>${workingDir}/cfHbPatched/software</arg>
<arg>--graphTableClassName</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--masterDuplicatePath</arg><arg>${workingDir}/masterduplicate</arg>
</spark>
<ok to="wait_clean_cfhb"/>
<error to="Kill"/>
</action>
<join name="wait_clean_cfhb" to="fork_copy_cfhb_patched_results"/>
<fork name="fork_copy_cfhb_patched_results">
<path start="copy_cfhb_patched_publication"/>
<path start="copy_cfhb_patched_dataset"/>
<path start="copy_cfhb_patched_otherresearchproduct"/>
<path start="copy_cfhb_patched_software"/>
</fork>
<action name="copy_cfhb_patched_publication">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<prepare>
<delete path="${graphOutputPath}/publication"/>
</prepare>
<arg>${workingDir}/cfHbPatched/publication</arg>
<arg>${graphOutputPath}/publication</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<action name="copy_cfhb_patched_dataset">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<prepare>
<delete path="${graphOutputPath}/dataset"/>
</prepare>
<arg>${workingDir}/cfHbPatched/dataset</arg>
<arg>${graphOutputPath}/dataset</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<action name="copy_cfhb_patched_otherresearchproduct">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<prepare>
<delete path="${graphOutputPath}/otherresearchproduct"/>
</prepare>
<arg>${workingDir}/cfHbPatched/otherresearchproduct</arg>
<arg>${graphOutputPath}/otherresearchproduct</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<action name="copy_cfhb_patched_software">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<prepare>
<delete path="${graphOutputPath}/software"/>
</prepare>
<arg>${workingDir}/cfHbPatched/software</arg>
<arg>${graphOutputPath}/software</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<join name="copy_wait" to="End"/>
<join name="wait_clean" to="End"/>
<end name="End"/>

View File

@ -28,5 +28,53 @@
"paramLongName": "graphTableClassName",
"paramDescription": "class name moelling the graph table",
"paramRequired": true
},
{
"paramName": "ci",
"paramLongName": "contextId",
"paramDescription": "the id of the context to be removed",
"paramRequired": false
},
{
"paramName": "vf",
"paramLongName": "verifyParam",
"paramDescription": "the parameter to be verified to remove the context",
"paramRequired": false
},
{
"paramName": "c",
"paramLongName": "country",
"paramDescription": "the id of the context to be removed",
"paramRequired": false
},
{
"paramName": "vfc",
"paramLongName": "verifyCountryParam",
"paramDescription": "the parameter to be verified to remove the country",
"paramRequired": false
},
{
"paramName": "cf",
"paramLongName": "collectedfrom",
"paramDescription": "the collectedfrom value for which we should apply the cleaning",
"paramRequired": false
},
{
"paramName": "hb",
"paramLongName": "hostedBy",
"paramDescription": "the set of datasources having the specified country in the graph searched for in the hostedby of the results",
"paramRequired": false
},
{
"paramName": "md",
"paramLongName": "masterDuplicatePath",
"paramDescription": "path to the file on HDFS holding the datasource id tuples [master, duplicate]",
"paramRequired": false
},
{
"paramName": "dc",
"paramLongName": "deepClean",
"paramDescription": "flag to activate further cleaning steps",
"paramRequired": true
}
]

View File

@ -1,289 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.util.List;
import org.apache.commons.io.FileUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SparkSession;
import org.junit.jupiter.api.AfterAll;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.BeforeAll;
import org.junit.jupiter.api.Test;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.schema.oaf.Publication;
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
public class CleanContextTest {
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static SparkSession spark;
private static Path workingDir;
private static final Logger log = LoggerFactory.getLogger(CleanContextTest.class);
@BeforeAll
public static void beforeAll() throws IOException {
workingDir = Files.createTempDirectory(CleanContextTest.class.getSimpleName());
log.info("using work dir {}", workingDir);
SparkConf conf = new SparkConf();
conf.setAppName(CleanContextTest.class.getSimpleName());
conf.setMaster("local[*]");
conf.set("spark.driver.host", "localhost");
conf.set("hive.metastore.local", "true");
conf.set("spark.ui.enabled", "false");
conf.set("spark.sql.warehouse.dir", workingDir.toString());
conf.set("hive.metastore.warehouse.dir", workingDir.resolve("warehouse").toString());
spark = SparkSession
.builder()
.appName(CleanContextTest.class.getSimpleName())
.config(conf)
.getOrCreate();
}
@AfterAll
public static void afterAll() throws IOException {
FileUtils.deleteDirectory(workingDir.toFile());
spark.stop();
}
@Test
public void testResultClean() throws Exception {
final String sourcePath = getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/publication_clean_context.json")
.getPath();
final String prefix = "gcube ";
spark
.read()
.textFile(sourcePath)
.map(
(MapFunction<String, Publication>) r -> OBJECT_MAPPER.readValue(r, Publication.class),
Encoders.bean(Publication.class))
.write()
.json(workingDir.toString() + "/publication");
CleanContextSparkJob.main(new String[] {
"--isSparkSessionManaged", Boolean.FALSE.toString(),
"--inputPath", workingDir.toString() + "/publication",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--workingDir", workingDir.toString() + "/working",
"--contextId", "sobigdata",
"--verifyParam", "gCube "
});
final JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
JavaRDD<Publication> tmp = sc
.textFile(workingDir.toString() + "/publication")
.map(item -> OBJECT_MAPPER.readValue(item, Publication.class));
Assertions.assertEquals(7, tmp.count());
// original result with sobigdata context and gcube as starting string in the main title for the publication
Assertions
.assertEquals(
0,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::0224aae28af558f21768dbc6439c7a95"))
.collect()
.get(0)
.getContext()
.size());
// original result with sobigdata context without gcube as starting string in the main title for the publication
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1b67df2d9"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"sobigdata::projects::2",
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1b67df2d9"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
// original result with sobigdata context with gcube as starting string in the subtitle
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6fa6af"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"sobigdata::projects::2",
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6fa6af"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
List<StructuredProperty> titles = tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6fa6af"))
.collect()
.get(0)
.getTitle();
Assertions.assertEquals(1, titles.size());
Assertions.assertTrue(titles.get(0).getValue().toLowerCase().startsWith(prefix));
Assertions.assertEquals("subtitle", titles.get(0).getQualifier().getClassid());
// original result with sobigdata context with gcube not as starting string in the main title
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9a97f"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"sobigdata::projects::1",
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9a97f"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
titles = tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9a97f"))
.collect()
.get(0)
.getTitle();
Assertions.assertEquals(1, titles.size());
Assertions.assertFalse(titles.get(0).getValue().toLowerCase().startsWith(prefix));
Assertions.assertTrue(titles.get(0).getValue().toLowerCase().contains(prefix.trim()));
Assertions.assertEquals("main title", titles.get(0).getQualifier().getClassid());
// original result with sobigdata in context and also other contexts with gcube as starting string for the main
// title
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4669a378a73661417182c208e6fdab53"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"dh-ch",
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4669a378a73661417182c208e6fdab53"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
titles = tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4669a378a73661417182c208e6fdab53"))
.collect()
.get(0)
.getTitle();
Assertions.assertEquals(1, titles.size());
Assertions.assertTrue(titles.get(0).getValue().toLowerCase().startsWith(prefix));
Assertions.assertEquals("main title", titles.get(0).getQualifier().getClassid());
// original result with multiple main title one of which whith gcube as starting string and with 2 contextes
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4a9152e80f860eab99072e921d74a0ff"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"dh-ch",
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4a9152e80f860eab99072e921d74a0ff"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
titles = tmp
.filter(p -> p.getId().equals("50|DansKnawCris::4a9152e80f860eab99072e921d74a0ff"))
.collect()
.get(0)
.getTitle();
Assertions.assertEquals(2, titles.size());
Assertions
.assertTrue(
titles
.stream()
.anyMatch(
t -> t.getQualifier().getClassid().equals("main title")
&& t.getValue().toLowerCase().startsWith(prefix)));
// original result without sobigdata in context with gcube as starting string for the main title
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6f6e8"))
.collect()
.get(0)
.getContext()
.size());
Assertions
.assertEquals(
"dh-ch",
tmp
.filter(p -> p.getId().equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6f6e8"))
.collect()
.get(0)
.getContext()
.get(0)
.getId());
titles = tmp
.filter(p -> p.getId().equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6f6e8"))
.collect()
.get(0)
.getTitle();
Assertions.assertEquals(2, titles.size());
Assertions
.assertTrue(
titles
.stream()
.anyMatch(
t -> t.getQualifier().getClassid().equals("main title")
&& t.getValue().toLowerCase().startsWith(prefix)));
}
}

View File

@ -1,190 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import org.apache.commons.io.FileUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SparkSession;
import org.junit.jupiter.api.AfterAll;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.BeforeAll;
import org.junit.jupiter.api.Test;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* @author miriam.baglioni
* @Date 20/07/22
*/
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.oa.graph.clean.country.CleanCountrySparkJob;
import eu.dnetlib.dhp.schema.oaf.Dataset;
import eu.dnetlib.dhp.schema.oaf.Publication;
public class CleanCountryTest {
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static SparkSession spark;
private static Path workingDir;
private static final Logger log = LoggerFactory.getLogger(CleanContextTest.class);
@BeforeAll
public static void beforeAll() throws IOException {
workingDir = Files.createTempDirectory(CleanCountryTest.class.getSimpleName());
log.info("using work dir {}", workingDir);
SparkConf conf = new SparkConf();
conf.setAppName(CleanCountryTest.class.getSimpleName());
conf.setMaster("local[*]");
conf.set("spark.driver.host", "localhost");
conf.set("hive.metastore.local", "true");
conf.set("spark.ui.enabled", "false");
conf.set("spark.sql.warehouse.dir", workingDir.toString());
conf.set("hive.metastore.warehouse.dir", workingDir.resolve("warehouse").toString());
spark = SparkSession
.builder()
.appName(CleanCountryTest.class.getSimpleName())
.config(conf)
.getOrCreate();
}
@AfterAll
public static void afterAll() throws IOException {
FileUtils.deleteDirectory(workingDir.toFile());
spark.stop();
}
@Test
public void testResultClean() throws Exception {
final String sourcePath = getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/publication_clean_country.json")
.getPath();
spark
.read()
.textFile(sourcePath)
.map(
(MapFunction<String, Publication>) r -> OBJECT_MAPPER.readValue(r, Publication.class),
Encoders.bean(Publication.class))
.write()
.json(workingDir.toString() + "/publication");
CleanCountrySparkJob.main(new String[] {
"--isSparkSessionManaged", Boolean.FALSE.toString(),
"--inputPath", workingDir.toString() + "/publication",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--workingDir", workingDir.toString() + "/working",
"--country", "NL",
"--verifyParam", "10.17632",
"--collectedfrom", "NARCIS",
"--hostedBy", getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/hostedBy")
.getPath()
});
final JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
JavaRDD<Publication> tmp = sc
.textFile(workingDir.toString() + "/publication")
.map(item -> OBJECT_MAPPER.readValue(item, Publication.class));
Assertions.assertEquals(8, tmp.count());
// original result with NL country and doi starting with Mendely prefix, but not collectedfrom NARCIS
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::0224aae28af558f21768dbc6439c7a95"))
.collect()
.get(0)
.getCountry()
.size());
// original result with NL country and pid not starting with Mendely prefix
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1b67df2d9"))
.collect()
.get(0)
.getCountry()
.size());
// original result with NL country and doi starting with Mendely prefix and collectedfrom NARCIS but not
// inserted with propagation
Assertions
.assertEquals(
1,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6fa6af"))
.collect()
.get(0)
.getCountry()
.size());
// original result with NL country and doi starting with Mendely prefix and collectedfrom NARCIS inserted with
// propagation
Assertions
.assertEquals(
0,
tmp
.filter(p -> p.getId().equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6fa6ag"))
.collect()
.get(0)
.getCountry()
.size());
}
@Test
public void testDatasetClean() throws Exception {
final String sourcePath = getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/dataset_clean_country.json")
.getPath();
spark
.read()
.textFile(sourcePath)
.map(
(MapFunction<String, Dataset>) r -> OBJECT_MAPPER.readValue(r, Dataset.class),
Encoders.bean(Dataset.class))
.write()
.json(workingDir.toString() + "/dataset");
CleanCountrySparkJob.main(new String[] {
"--isSparkSessionManaged", Boolean.FALSE.toString(),
"--inputPath", workingDir.toString() + "/dataset",
"-graphTableClassName", Dataset.class.getCanonicalName(),
"-workingDir", workingDir.toString() + "/working",
"-country", "NL",
"-verifyParam", "10.17632",
"-collectedfrom", "NARCIS",
"-hostedBy", getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/hostedBy")
.getPath()
});
final JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
JavaRDD<Dataset> tmp = sc
.textFile(workingDir.toString() + "/dataset")
.map(item -> OBJECT_MAPPER.readValue(item, Dataset.class));
Assertions.assertEquals(1, tmp.count());
Assertions.assertEquals(0, tmp.first().getCountry().size());
}
}

View File

@ -0,0 +1,924 @@
package eu.dnetlib.dhp.oa.graph.clean;
import static org.junit.jupiter.api.Assertions.*;
import static org.mockito.Mockito.lenient;
import java.io.File;
import java.io.IOException;
import java.net.URISyntaxException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.util.*;
import java.util.stream.Collectors;
import org.apache.commons.cli.ParseException;
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.IOUtils;
import org.apache.commons.io.filefilter.FalseFileFilter;
import org.apache.commons.io.filefilter.TrueFileFilter;
import org.apache.commons.lang3.StringUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.FilterFunction;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SparkSession;
import org.junit.jupiter.api.AfterAll;
import org.junit.jupiter.api.BeforeAll;
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.DeserializationFeature;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
@ExtendWith(MockitoExtension.class)
public class CleanGraphSparkJobTest {
private static final Logger log = LoggerFactory.getLogger(CleanGraphSparkJobTest.class);
public static final ObjectMapper MAPPER = new ObjectMapper()
.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
@Mock
private ISLookUpService isLookUpService;
private VocabularyGroup vocabularies;
private CleaningRuleMap mapping;
private static SparkSession spark;
private static Path testBaseTmpPath;
private static String graphInputPath;
private static String graphOutputPath;
private static String dsMasterDuplicatePath;
@BeforeAll
public static void beforeAll() throws IOException, URISyntaxException {
testBaseTmpPath = Files.createTempDirectory(CleanGraphSparkJobTest.class.getSimpleName());
log.info("using test base path {}", testBaseTmpPath);
File basePath = Paths
.get(
Objects
.requireNonNull(
CleanGraphSparkJobTest.class.getResource("/eu/dnetlib/dhp/oa/graph/clean/graph"))
.toURI())
.toFile();
List<File> paths = FileUtils
.listFilesAndDirs(basePath, FalseFileFilter.FALSE, TrueFileFilter.TRUE)
.stream()
.filter(f -> !f.getAbsolutePath().endsWith("/graph"))
.collect(Collectors.toList());
for (File path : paths) {
String type = StringUtils.substringAfterLast(path.getAbsolutePath(), "/");
FileUtils
.copyDirectory(
path,
testBaseTmpPath.resolve("input").resolve("graph").resolve(type).toFile());
}
FileUtils
.copyFileToDirectory(
Paths
.get(
CleanGraphSparkJobTest.class
.getResource("/eu/dnetlib/dhp/oa/graph/clean/cfhb/masterduplicate.json")
.toURI())
.toFile(),
testBaseTmpPath.resolve("workingDir").resolve("masterduplicate").toFile());
graphInputPath = testBaseTmpPath.resolve("input").resolve("graph").toString();
graphOutputPath = testBaseTmpPath.resolve("output").resolve("graph").toString();
dsMasterDuplicatePath = testBaseTmpPath.resolve("workingDir").resolve("masterduplicate").toString();
SparkConf conf = new SparkConf();
conf.setAppName(CleanGraphSparkJobTest.class.getSimpleName());
conf.setMaster("local[*]");
conf.set("spark.driver.host", "localhost");
conf.set("hive.metastore.local", "true");
conf.set("spark.ui.enabled", "false");
conf.set("spark.sql.warehouse.dir", testBaseTmpPath.toString());
conf.set("hive.metastore.warehouse.dir", testBaseTmpPath.resolve("warehouse").toString());
spark = SparkSession
.builder()
.config(conf)
.getOrCreate();
}
@BeforeEach
public void setUp() throws ISLookUpException, IOException {
lenient().when(isLookUpService.quickSearchProfile(VocabularyGroup.VOCABULARIES_XQUERY)).thenReturn(vocs());
lenient()
.when(isLookUpService.quickSearchProfile(VocabularyGroup.VOCABULARY_SYNONYMS_XQUERY))
.thenReturn(synonyms());
vocabularies = VocabularyGroup.loadVocsFromIS(isLookUpService);
mapping = CleaningRuleMap.create(vocabularies);
}
@AfterAll
public static void afterAll() throws IOException {
FileUtils.deleteDirectory(testBaseTmpPath.toFile());
spark.stop();
}
@Test
void testCleanRelations() throws Exception {
spark
.read()
.textFile(graphInputPath.toString() + "/relation")
.map(as(Relation.class), Encoders.bean(Relation.class))
.collectAsList()
.forEach(
r -> assertFalse(
vocabularies.getTerms(ModelConstants.DNET_RELATION_RELCLASS).contains(r.getRelClass())));
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/relation",
"--outputPath", graphOutputPath + "/relation",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Relation.class.getCanonicalName(),
"--deepClean", "false",
"--masterDuplicatePath", dsMasterDuplicatePath,
})).run(false, isLookUpService);
spark
.read()
.textFile(graphOutputPath.toString() + "/relation")
.map(as(Relation.class), Encoders.bean(Relation.class))
.collectAsList()
.forEach(r -> {
assertTrue(vocabularies.getTerms(ModelConstants.DNET_RELATION_RELCLASS).contains(r.getRelClass()));
assertTrue(vocabularies.getTerms(ModelConstants.DNET_RELATION_SUBRELTYPE).contains(r.getSubRelType()));
assertEquals("iis", r.getDataInfo().getProvenanceaction().getClassid());
assertEquals("Inferred by OpenAIRE", r.getDataInfo().getProvenanceaction().getClassname());
});
}
@Test
void testFilter_invisible_true() throws Exception {
assertNotNull(vocabularies);
assertNotNull(mapping);
String json = IOUtils
.toString(
Objects
.requireNonNull(
getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/result_invisible.json")));
Publication p_in = MAPPER.readValue(json, Publication.class);
assertTrue(p_in instanceof Result);
assertTrue(p_in instanceof Publication);
assertEquals(true, GraphCleaningFunctions.filter(p_in));
}
@Test
void testFilter_true_nothing_to_filter() throws Exception {
assertNotNull(vocabularies);
assertNotNull(mapping);
String json = IOUtils
.toString(
Objects
.requireNonNull(
getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/result.json")));
Publication p_in = MAPPER.readValue(json, Publication.class);
assertTrue(p_in instanceof Result);
assertTrue(p_in instanceof Publication);
assertEquals(true, GraphCleaningFunctions.filter(p_in));
}
@Test
void testFilter_missing_invisible() throws Exception {
assertNotNull(vocabularies);
assertNotNull(mapping);
String json = IOUtils
.toString(
Objects
.requireNonNull(
getClass()
.getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/result_missing_invisible.json")));
Publication p_in = MAPPER.readValue(json, Publication.class);
assertTrue(p_in instanceof Result);
assertTrue(p_in instanceof Publication);
assertEquals(true, GraphCleaningFunctions.filter(p_in));
}
@Test
void testCleaning_publication() throws Exception {
final String id = "50|CSC_________::2250a70c903c6ac6e4c01438259e9375";
Publication p_in = read(spark, graphInputPath + "/publication", Publication.class)
.filter(String.format("id = '%s'", id))
.first();
assertNull(p_in.getBestaccessright());
assertTrue(p_in instanceof Result);
assertTrue(p_in instanceof Publication);
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/publication",
"--outputPath", graphOutputPath + "/publication",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--deepClean", "false",
"--masterDuplicatePath", dsMasterDuplicatePath,
})).run(false, isLookUpService);
Publication p = read(spark, graphOutputPath + "/publication", Publication.class)
.filter(String.format("id = '%s'", id))
.first();
assertNull(p.getPublisher());
assertEquals("und", p.getLanguage().getClassid());
assertEquals("Undetermined", p.getLanguage().getClassname());
assertEquals("DE", p.getCountry().get(0).getClassid());
assertEquals("Germany", p.getCountry().get(0).getClassname());
assertEquals("0018", p.getInstance().get(0).getInstancetype().getClassid());
assertEquals("Annotation", p.getInstance().get(0).getInstancetype().getClassname());
assertEquals("0027", p.getInstance().get(1).getInstancetype().getClassid());
assertEquals("Model", p.getInstance().get(1).getInstancetype().getClassname());
assertEquals("0038", p.getInstance().get(2).getInstancetype().getClassid());
assertEquals("Other literature type", p.getInstance().get(2).getInstancetype().getClassname());
assertEquals("CLOSED", p.getInstance().get(0).getAccessright().getClassid());
assertEquals("Closed Access", p.getInstance().get(0).getAccessright().getClassname());
Set<String> pidTerms = vocabularies.getTerms(ModelConstants.DNET_PID_TYPES);
assertTrue(
p
.getPid()
.stream()
.map(StructuredProperty::getQualifier)
.allMatch(q -> pidTerms.contains(q.getClassid())));
List<Instance> poi = p.getInstance();
assertNotNull(poi);
assertEquals(3, poi.size());
final Instance poii = poi.get(0);
assertNotNull(poii);
assertNotNull(poii.getPid());
assertEquals(2, poii.getPid().size());
assertTrue(
poii.getPid().stream().anyMatch(s -> s.getValue().equals("10.1007/s109090161569x")));
assertTrue(poii.getPid().stream().anyMatch(s -> s.getValue().equals("10.1008/abcd")));
assertNotNull(poii.getAlternateIdentifier());
assertEquals(1, poii.getAlternateIdentifier().size());
assertTrue(
poii
.getAlternateIdentifier()
.stream()
.anyMatch(s -> s.getValue().equals("10.1009/qwerty")));
assertEquals(3, p.getTitle().size());
List<String> titles = p
.getTitle()
.stream()
.map(StructuredProperty::getValue)
.collect(Collectors.toList());
assertTrue(titles.contains("omic"));
assertTrue(
titles.contains("Optical response of strained- and unstrained-silicon cold-electron bolometers test"));
assertTrue(titles.contains("「マキャベリ的知性と心の理論の進化論」 リチャード・バーン, アンドリュー・ホワイトゥン 編/藤田和生, 山下博志, 友永雅巳 監訳"));
assertEquals("CLOSED", p.getBestaccessright().getClassid());
assertNull(p.getPublisher());
assertEquals("1970-10-07", p.getDateofacceptance().getValue());
assertEquals("0038", p.getInstance().get(2).getInstancetype().getClassid());
assertEquals("Other literature type", p.getInstance().get(2).getInstancetype().getClassname());
final List<Instance> pci = p.getInstance();
assertNotNull(pci);
assertEquals(3, pci.size());
final Instance pcii = pci.get(0);
assertNotNull(pcii);
assertNotNull(pcii.getPid());
assertEquals(2, pcii.getPid().size());
assertTrue(
pcii.getPid().stream().anyMatch(s -> s.getValue().equals("10.1007/s109090161569x")));
assertTrue(pcii.getPid().stream().anyMatch(s -> s.getValue().equals("10.1008/abcd")));
assertNotNull(pcii.getAlternateIdentifier());
assertEquals(1, pcii.getAlternateIdentifier().size());
assertTrue(
pcii
.getAlternateIdentifier()
.stream()
.anyMatch(s -> s.getValue().equals("10.1009/qwerty")));
assertNotNull(p.getSubject());
List<Subject> fos_subjects = p
.getSubject()
.stream()
.filter(s -> ModelConstants.DNET_SUBJECT_FOS_CLASSID.equals(s.getQualifier().getClassid()))
.collect(Collectors.toList());
assertNotNull(fos_subjects);
assertEquals(2, fos_subjects.size());
assertTrue(
fos_subjects
.stream()
.anyMatch(
s -> "0101 mathematics".equals(s.getValue()) &
ModelConstants.DNET_SUBJECT_FOS_CLASSID.equals(s.getQualifier().getClassid()) &
"sysimport:crosswalk:datasetarchive"
.equals(s.getDataInfo().getProvenanceaction().getClassid())));
assertTrue(
fos_subjects
.stream()
.anyMatch(
s -> "0102 computer and information sciences".equals(s.getValue()) &
ModelConstants.DNET_SUBJECT_FOS_CLASSID.equals(s.getQualifier().getClassid())));
verify_keyword(p, "In Situ Hybridization");
verify_keyword(p, "Avicennia");
}
@Test
void testCleanDoiBoost() throws IOException, ParseException, ISLookUpException, ClassNotFoundException {
verifyFiltering(1, "50|doi_________::b0baa0eb88a5788f0b8815560d2a32f2");
}
@Test
void testCleanDoiBoost2() throws IOException, ParseException, ISLookUpException, ClassNotFoundException {
verifyFiltering(1, "50|doi_________::4972b0ca81b96b225aed8038bb965656");
}
private void verifyFiltering(int expectedCount, String id)
throws ISLookUpException, ClassNotFoundException, IOException, ParseException {
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/publication",
"--outputPath", graphOutputPath + "/publication",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--deepClean", "false",
"--masterDuplicatePath", dsMasterDuplicatePath,
})).run(false, isLookUpService);
Dataset<Publication> p = read(spark, graphOutputPath + "/publication", Publication.class)
.filter(String.format("id = '%s'", id));
assertEquals(expectedCount, p.count());
}
@Test
void testCleanContext() throws Exception {
final String prefix = "gcube ";
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/publication",
"--outputPath", graphOutputPath + "/publication",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--deepClean", "true",
"--contextId", "sobigdata",
"--verifyParam", "gCube ",
"--masterDuplicatePath", dsMasterDuplicatePath,
"--country", "NL",
"--verifyCountryParam", "10.17632",
"--collectedfrom", "NARCIS",
"--hostedBy", Objects
.requireNonNull(
getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/hostedBy"))
.getPath()
})).run(false, isLookUpService);
Dataset<Publication> pubs = read(spark, graphOutputPath + "/publication", Publication.class)
.filter((FilterFunction<Publication>) p1 -> StringUtils.endsWith(p1.getId(), "_ctx"));
assertEquals(7, pubs.count());
// original result with sobigdata context and gcube as starting string in the main title for the publication
assertEquals(
0,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::0224aae28af558f21768dbc6439a_ctx"))
.first()
.getContext()
.size());
// original result with sobigdata context without gcube as starting string in the main title for the publication
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1b67d_ctx"))
.first()
.getContext()
.size());
assertEquals(
"sobigdata::projects::2",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1b67d_ctx"))
.first()
.getContext()
.get(0)
.getId());
// original result with sobigdata context with gcube as starting string in the subtitle
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6f_ctx"))
.first()
.getContext()
.size());
assertEquals(
"sobigdata::projects::2",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6f_ctx"))
.first()
.getContext()
.get(0)
.getId());
List<StructuredProperty> titles = pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c81248c335f0aa07e06817ece6f_ctx"))
.first()
.getTitle();
assertEquals(1, titles.size());
assertTrue(titles.get(0).getValue().toLowerCase().startsWith(prefix));
assertEquals("subtitle", titles.get(0).getQualifier().getClassid());
// original result with sobigdata context with gcube not as starting string in the main title
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9_ctx"))
.first()
.getContext()
.size());
assertEquals(
"sobigdata::projects::1",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9_ctx"))
.first()
.getContext()
.get(0)
.getId());
titles = pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c9f068ddc930360bec6925488a9_ctx"))
.first()
.getTitle();
assertEquals(1, titles.size());
assertFalse(titles.get(0).getValue().toLowerCase().startsWith(prefix));
assertTrue(titles.get(0).getValue().toLowerCase().contains(prefix.trim()));
assertEquals("main title", titles.get(0).getQualifier().getClassid());
// original result with sobigdata in context and also other contexts with gcube as starting string for the main
// title
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4669a378a73661417182c208e6fd_ctx"))
.first()
.getContext()
.size());
assertEquals(
"dh-ch",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4669a378a73661417182c208e6fd_ctx"))
.first()
.getContext()
.get(0)
.getId());
titles = pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4669a378a73661417182c208e6fd_ctx"))
.first()
.getTitle();
assertEquals(1, titles.size());
assertTrue(titles.get(0).getValue().toLowerCase().startsWith(prefix));
assertEquals("main title", titles.get(0).getQualifier().getClassid());
// original result with multiple main title one of which whith gcube as starting string and with 2 contextes
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4a9152e80f860eab99072e921d74_ctx"))
.first()
.getContext()
.size());
assertEquals(
"dh-ch",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4a9152e80f860eab99072e921d74_ctx"))
.first()
.getContext()
.get(0)
.getId());
titles = pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::4a9152e80f860eab99072e921d74_ctx"))
.first()
.getTitle();
assertEquals(2, titles.size());
assertTrue(
titles
.stream()
.anyMatch(
t -> t.getQualifier().getClassid().equals("main title")
&& t.getValue().toLowerCase().startsWith(prefix)));
// original result without sobigdata in context with gcube as starting string for the main title
assertEquals(
1,
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6_ctx"))
.first()
.getContext()
.size());
assertEquals(
"dh-ch",
pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6_ctx"))
.first()
.getContext()
.get(0)
.getId());
titles = pubs
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|dedup_wf_001::01e6a28565ca01376b7548e530c6_ctx"))
.first()
.getTitle();
assertEquals(2, titles.size());
assertTrue(
titles
.stream()
.anyMatch(
t -> t.getQualifier().getClassid().equals("main title")
&& t.getValue().toLowerCase().startsWith(prefix)));
}
@Test
void testCleanCfHbSparkJob() throws Exception {
final Dataset<Publication> pubs_in = read(spark, graphInputPath + "/publication", Publication.class);
final Publication p1_in = pubs_in
.filter("id = '50|doi_________::09821844208a5cd6300b2bfb13b_cfhb'")
.first();
assertEquals("10|re3data_____::4c4416659cb74c2e0e891a883a047cbc", p1_in.getCollectedfrom().get(0).getKey());
assertEquals("Bacterial Protein Interaction Database - DUP", p1_in.getCollectedfrom().get(0).getValue());
assertEquals(
"10|re3data_____::4c4416659cb74c2e0e891a883a047cbc",
p1_in.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"Bacterial Protein Interaction Database - DUP", p1_in.getInstance().get(0).getCollectedfrom().getValue());
final Publication p2_in = pubs_in
.filter("id = '50|DansKnawCris::0dd644304b7116e8e58da3a5e3a_cfhb'")
.first();
assertEquals("10|opendoar____::788b4ac1e172d8e520c2b9461c0a3d35", p2_in.getCollectedfrom().get(0).getKey());
assertEquals("FILUR DATA - DUP", p2_in.getCollectedfrom().get(0).getValue());
assertEquals(
"10|opendoar____::788b4ac1e172d8e520c2b9461c0a3d35",
p2_in.getInstance().get(0).getCollectedfrom().getKey());
assertEquals("FILUR DATA - DUP", p2_in.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|re3data_____::6ffd7bc058f762912dc494cd9c175341", p2_in.getInstance().get(0).getHostedby().getKey());
assertEquals("depositar - DUP", p2_in.getInstance().get(0).getHostedby().getValue());
final Publication p3_in = pubs_in
.filter("id = '50|DansKnawCris::203a27996ddc0fd1948258e5b7d_cfhb'")
.first();
assertEquals("10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", p3_in.getCollectedfrom().get(0).getKey());
assertEquals("DANS (Data Archiving and Networked Services)", p3_in.getCollectedfrom().get(0).getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f",
p3_in.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", p3_in.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", p3_in.getInstance().get(0).getHostedby().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", p3_in.getInstance().get(0).getHostedby().getValue());
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/publication",
"--outputPath", graphOutputPath + "/publication",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--deepClean", "true",
"--contextId", "sobigdata",
"--verifyParam", "gCube ",
"--masterDuplicatePath", dsMasterDuplicatePath,
"--country", "NL",
"--verifyCountryParam", "10.17632",
"--collectedfrom", "NARCIS",
"--hostedBy", Objects
.requireNonNull(
getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/hostedBy"))
.getPath()
})).run(false, isLookUpService);
assertTrue(Files.exists(Paths.get(graphOutputPath, "publication")));
final Dataset<Publication> pubs_out = read(spark, graphOutputPath + "/publication", Publication.class)
.filter((FilterFunction<Publication>) p -> StringUtils.endsWith(p.getId(), "_cfhb"));
assertEquals(3, pubs_out.count());
final Publication p1_out = pubs_out
.filter("id = '50|doi_________::09821844208a5cd6300b2bfb13b_cfhb'")
.first();
assertEquals("10|fairsharing_::a29d1598024f9e87beab4b98411d48ce", p1_out.getCollectedfrom().get(0).getKey());
assertEquals("Bacterial Protein Interaction Database", p1_out.getCollectedfrom().get(0).getValue());
assertEquals(
"10|fairsharing_::a29d1598024f9e87beab4b98411d48ce",
p1_out.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"Bacterial Protein Interaction Database", p1_out.getInstance().get(0).getCollectedfrom().getValue());
final Publication p2_out = pubs_out
.filter("id = '50|DansKnawCris::0dd644304b7116e8e58da3a5e3a_cfhb'")
.first();
assertEquals("10|re3data_____::fc1db64b3964826913b1e9eafe830490", p2_out.getCollectedfrom().get(0).getKey());
assertEquals("FULIR Data", p2_out.getCollectedfrom().get(0).getValue());
assertEquals(
"10|re3data_____::fc1db64b3964826913b1e9eafe830490",
p2_out.getInstance().get(0).getCollectedfrom().getKey());
assertEquals("FULIR Data", p2_out.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|fairsharing_::3f647cadf56541fb9513cb63ec370187", p2_out.getInstance().get(0).getHostedby().getKey());
assertEquals("depositar", p2_out.getInstance().get(0).getHostedby().getValue());
final Publication p3_out = pubs_out
.filter("id = '50|DansKnawCris::203a27996ddc0fd1948258e5b7d_cfhb'")
.first();
assertEquals("10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", p3_out.getCollectedfrom().get(0).getKey());
assertEquals("DANS (Data Archiving and Networked Services)", p3_out.getCollectedfrom().get(0).getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f",
p3_out.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", p3_out.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", p3_out.getInstance().get(0).getHostedby().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", p3_out.getInstance().get(0).getHostedby().getValue());
}
@Test
void testCleanCountry() throws Exception {
new CleanGraphSparkJob(
args(
"/eu/dnetlib/dhp/oa/graph/input_clean_graph_parameters.json",
new String[] {
"--inputPath", graphInputPath + "/publication",
"--outputPath", graphOutputPath + "/publication",
"--isLookupUrl", "lookupurl",
"--graphTableClassName", Publication.class.getCanonicalName(),
"--deepClean", "true",
"--contextId", "sobigdata",
"--verifyParam", "gCube ",
"--masterDuplicatePath", dsMasterDuplicatePath,
"--country", "NL",
"--verifyCountryParam", "10.17632",
"--collectedfrom", "NARCIS",
"--hostedBy", Objects
.requireNonNull(
getClass()
.getResource("/eu/dnetlib/dhp/oa/graph/clean/hostedBy"))
.getPath()
})).run(false, isLookUpService);
final Dataset<Publication> pubs_out = read(spark, graphOutputPath + "/publication", Publication.class)
.filter((FilterFunction<Publication>) p -> StringUtils.endsWith(p.getId(), "_country"));
assertEquals(8, pubs_out.count());
// original result with NL country and doi starting with Mendely prefix, but not collectedfrom NARCIS
assertEquals(
1,
pubs_out
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::0224aae28af558f21768dbc6_country"))
.first()
.getCountry()
.size());
// original result with NL country and pid not starting with Mendely prefix
assertEquals(
1,
pubs_out
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::20c414a3b1c742d5dd3851f1_country"))
.first()
.getCountry()
.size());
// original result with NL country and doi starting with Mendely prefix and collectedfrom NARCIS but not
// inserted with propagation
assertEquals(
1,
pubs_out
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c81248c335f0aa07e06817e_country"))
.first()
.getCountry()
.size());
// original result with NL country and doi starting with Mendely prefix and collectedfrom NARCIS inserted with
// propagation
assertEquals(
0,
pubs_out
.filter(
(FilterFunction<Publication>) p -> p
.getId()
.equals("50|DansKnawCris::3c81248c335f0aa07e06817d_country"))
.first()
.getCountry()
.size());
}
private List<String> vocs() throws IOException {
return IOUtils
.readLines(
Objects
.requireNonNull(
getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/terms.txt")));
}
private List<String> synonyms() throws IOException {
return IOUtils
.readLines(
Objects
.requireNonNull(
getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/synonyms.txt")));
}
private <R> org.apache.spark.sql.Dataset<R> read(SparkSession spark, String path, Class<R> clazz) {
return spark
.read()
.textFile(path)
.map(as(clazz), Encoders.bean(clazz));
}
private static <R> MapFunction<String, R> as(Class<R> clazz) {
return s -> MAPPER.readValue(s, clazz);
}
private static String classPathResourceAsString(String path) throws IOException {
return IOUtils
.toString(
Objects
.requireNonNull(
CleanGraphSparkJobTest.class.getResourceAsStream(path)));
}
private ArgumentApplicationParser args(String paramSpecs, String[] args) throws IOException, ParseException {
ArgumentApplicationParser parser = new ArgumentApplicationParser(classPathResourceAsString(paramSpecs));
parser.parseArgument(args);
return parser;
}
private static void verify_keyword(Publication p_cleaned, String subject) {
Optional<Subject> s1 = p_cleaned
.getSubject()
.stream()
.filter(s -> s.getValue().equals(subject))
.findFirst();
assertTrue(s1.isPresent());
assertEquals(ModelConstants.DNET_SUBJECT_KEYWORD, s1.get().getQualifier().getClassid());
assertEquals(ModelConstants.DNET_SUBJECT_KEYWORD, s1.get().getQualifier().getClassname());
}
}

View File

@ -13,7 +13,6 @@ import java.util.stream.Collectors;
import java.util.stream.Stream;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.hdfs.server.datanode.fsdataset.impl.MappableBlock;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
@ -59,7 +58,7 @@ public class GraphCleaningFunctionsTest {
void testCleanRelations() throws Exception {
List<String> lines = IOUtils
.readLines(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/relation.json"));
.readLines(getClass().getResourceAsStream("/eu/dnetlib/dhp/oa/graph/clean/graph/relation/relation.json"));
for (String json : lines) {
Relation r_in = MAPPER.readValue(json, Relation.class);
assertNotNull(r_in);

View File

@ -1,213 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean.cfhb;
import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertTrue;
import java.io.File;
import java.io.IOException;
import java.net.URISyntaxException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import org.apache.commons.io.FileUtils;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Encoders;
import org.apache.spark.sql.SparkSession;
import org.junit.jupiter.api.*;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.schema.oaf.Dataset;
import eu.dnetlib.dhp.schema.oaf.Publication;
public class CleanCfHbSparkJobTest {
private static final Logger log = LoggerFactory.getLogger(CleanCfHbSparkJobTest.class);
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static SparkSession spark;
private static Path testBaseTmpPath;
private static String resolvedPath;
private static String graphInputPath;
private static String graphOutputPath;
private static String dsMasterDuplicatePath;
@BeforeAll
public static void beforeAll() throws IOException, URISyntaxException {
testBaseTmpPath = Files.createTempDirectory(CleanCfHbSparkJobTest.class.getSimpleName());
log.info("using test base path {}", testBaseTmpPath);
final File entitiesSources = Paths
.get(CleanCfHbSparkJobTest.class.getResource("/eu/dnetlib/dhp/oa/graph/clean/cfhb/entities").toURI())
.toFile();
FileUtils
.copyDirectory(
entitiesSources,
testBaseTmpPath.resolve("input").resolve("entities").toFile());
FileUtils
.copyFileToDirectory(
Paths
.get(
CleanCfHbSparkJobTest.class
.getResource("/eu/dnetlib/dhp/oa/graph/clean/cfhb/masterduplicate.json")
.toURI())
.toFile(),
testBaseTmpPath.resolve("workingDir").resolve("masterduplicate").toFile());
graphInputPath = testBaseTmpPath.resolve("input").resolve("entities").toString();
resolvedPath = testBaseTmpPath.resolve("workingDir").resolve("cfHbResolved").toString();
graphOutputPath = testBaseTmpPath.resolve("workingDir").resolve("cfHbPatched").toString();
dsMasterDuplicatePath = testBaseTmpPath.resolve("workingDir").resolve("masterduplicate").toString();
SparkConf conf = new SparkConf();
conf.setAppName(CleanCfHbSparkJobTest.class.getSimpleName());
conf.setMaster("local[*]");
conf.set("spark.driver.host", "localhost");
conf.set("spark.ui.enabled", "false");
spark = SparkSession
.builder()
.appName(CleanCfHbSparkJobTest.class.getSimpleName())
.config(conf)
.getOrCreate();
}
@AfterAll
public static void afterAll() throws IOException {
FileUtils.deleteDirectory(testBaseTmpPath.toFile());
spark.stop();
}
@Test
void testCleanCfHbSparkJob() throws Exception {
final String outputPath = graphOutputPath + "/dataset";
final String inputPath = graphInputPath + "/dataset";
org.apache.spark.sql.Dataset<Dataset> records = read(spark, inputPath, Dataset.class);
Dataset d = records
.filter("id = '50|doi_________::09821844208a5cd6300b2bfb13bca1b9'")
.first();
assertEquals("10|re3data_____::4c4416659cb74c2e0e891a883a047cbc", d.getCollectedfrom().get(0).getKey());
assertEquals("Bacterial Protein Interaction Database - DUP", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|re3data_____::4c4416659cb74c2e0e891a883a047cbc", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"Bacterial Protein Interaction Database - DUP", d.getInstance().get(0).getCollectedfrom().getValue());
d = records
.filter("id = '50|DansKnawCris::0dd644304b7116e8e58da3a5e3adc37a'")
.first();
assertEquals("10|opendoar____::788b4ac1e172d8e520c2b9461c0a3d35", d.getCollectedfrom().get(0).getKey());
assertEquals("FILUR DATA - DUP", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|opendoar____::788b4ac1e172d8e520c2b9461c0a3d35", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals("FILUR DATA - DUP", d.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|re3data_____::6ffd7bc058f762912dc494cd9c175341", d.getInstance().get(0).getHostedby().getKey());
assertEquals("depositar - DUP", d.getInstance().get(0).getHostedby().getValue());
d = records
.filter("id = '50|DansKnawCris::203a27996ddc0fd1948258e5b7dec61c'")
.first();
assertEquals("10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getCollectedfrom().get(0).getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getHostedby().getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getHostedby().getValue());
CleanCfHbSparkJob
.main(
new String[] {
"--isSparkSessionManaged", Boolean.FALSE.toString(),
"--inputPath", inputPath,
"--outputPath", outputPath,
"--resolvedPath", resolvedPath + "/dataset",
"--graphTableClassName", Dataset.class.getCanonicalName(),
"--masterDuplicatePath", dsMasterDuplicatePath
});
assertTrue(Files.exists(Paths.get(graphOutputPath, "dataset")));
records = read(spark, outputPath, Dataset.class);
assertEquals(3, records.count());
d = records
.filter("id = '50|doi_________::09821844208a5cd6300b2bfb13bca1b9'")
.first();
assertEquals("10|fairsharing_::a29d1598024f9e87beab4b98411d48ce", d.getCollectedfrom().get(0).getKey());
assertEquals("Bacterial Protein Interaction Database", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|fairsharing_::a29d1598024f9e87beab4b98411d48ce", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals("Bacterial Protein Interaction Database", d.getInstance().get(0).getCollectedfrom().getValue());
d = records
.filter("id = '50|DansKnawCris::0dd644304b7116e8e58da3a5e3adc37a'")
.first();
assertEquals("10|re3data_____::fc1db64b3964826913b1e9eafe830490", d.getCollectedfrom().get(0).getKey());
assertEquals("FULIR Data", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|re3data_____::fc1db64b3964826913b1e9eafe830490", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals("FULIR Data", d.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|fairsharing_::3f647cadf56541fb9513cb63ec370187", d.getInstance().get(0).getHostedby().getKey());
assertEquals("depositar", d.getInstance().get(0).getHostedby().getValue());
d = records
.filter("id = '50|DansKnawCris::203a27996ddc0fd1948258e5b7dec61c'")
.first();
assertEquals("10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getCollectedfrom().get(0).getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getHostedby().getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getHostedby().getValue());
d = records
.filter("id = '50|DansKnawCris::203a27996ddc0fd1948258e5b7dec61c'")
.first();
assertEquals("10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getCollectedfrom().get(0).getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getCollectedfrom().get(0).getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getCollectedfrom().getKey());
assertEquals(
"DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getCollectedfrom().getValue());
assertEquals(
"10|openaire____::c6df70599aa984f16ee52b4b86d2e89f", d.getInstance().get(0).getHostedby().getKey());
assertEquals("DANS (Data Archiving and Networked Services)", d.getInstance().get(0).getHostedby().getValue());
}
private <R> org.apache.spark.sql.Dataset<R> read(SparkSession spark, String path, Class<R> clazz) {
return spark
.read()
.textFile(path)
.map(as(clazz), Encoders.bean(clazz));
}
private static <R> MapFunction<String, R> as(Class<R> clazz) {
return s -> OBJECT_MAPPER.readValue(s, clazz);
}
}

View File

@ -0,0 +1,11 @@
<configuration>
<appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
<encoder>
<pattern>%d{HH:mm:ss.SSS} %-5level %logger{36} - %msg%n</pattern>
</encoder>
</appender>
<root level="info">
<appender-ref ref="STDOUT" />
</root>
</configuration>