[aggregator graph] save invalid records aside for further inspection

This commit is contained in:
Claudio Atzori 2022-09-16 14:06:28 +02:00
parent 1e42d984e1
commit e370e940d8
6 changed files with 191 additions and 46 deletions

View File

@ -79,9 +79,6 @@ public class GenerateEntitiesApplication {
final String targetPath = parser.get("targetPath"); final String targetPath = parser.get("targetPath");
log.info("targetPath: {}", targetPath); log.info("targetPath: {}", targetPath);
final String invalidPath = parser.get("invalidPath");
log.info("invalidPath: {}", invalidPath);
final String isLookupUrl = parser.get("isLookupUrl"); final String isLookupUrl = parser.get("isLookupUrl");
log.info("isLookupUrl: {}", isLookupUrl); log.info("isLookupUrl: {}", isLookupUrl);
@ -103,8 +100,7 @@ public class GenerateEntitiesApplication {
final SparkConf conf = new SparkConf(); final SparkConf conf = new SparkConf();
runWithSparkSession(conf, isSparkSessionManaged, spark -> { runWithSparkSession(conf, isSparkSessionManaged, spark -> {
HdfsSupport.remove(targetPath, spark.sparkContext().hadoopConfiguration()); HdfsSupport.remove(targetPath, spark.sparkContext().hadoopConfiguration());
HdfsSupport.remove(invalidPath, spark.sparkContext().hadoopConfiguration()); generateEntities(spark, vocs, sourcePaths, targetPath, shouldHashId, mode);
generateEntities(spark, vocs, sourcePaths, targetPath, invalidPath, shouldHashId, mode);
}); });
} }
@ -113,7 +109,6 @@ public class GenerateEntitiesApplication {
final VocabularyGroup vocs, final VocabularyGroup vocs,
final String sourcePaths, final String sourcePaths,
final String targetPath, final String targetPath,
final String invalidPath,
final boolean shouldHashId, final boolean shouldHashId,
final Mode mode) { final Mode mode) {
@ -126,21 +121,6 @@ public class GenerateEntitiesApplication {
log.info("Generate entities from files:"); log.info("Generate entities from files:");
existingSourcePaths.forEach(log::info); existingSourcePaths.forEach(log::info);
for (final String sp : existingSourcePaths) {
RDD<String> invalidRecords = sc
.sequenceFile(sp, Text.class, Text.class)
.map(k -> new Tuple2<>(k._1().toString(), k._2().toString()))
.map(k -> tryApplyMapping(k._1(), k._2(), shouldHashId, vocs))
.filter(Objects::nonNull)
.rdd();
spark
.createDataset(invalidRecords, Encoders.STRING())
.write()
.mode(SaveMode.Append)
.option("compression", "gzip")
.text(invalidPath);
}
JavaRDD<Oaf> inputRdd = sc.emptyRDD(); JavaRDD<Oaf> inputRdd = sc.emptyRDD();
for (final String sp : existingSourcePaths) { for (final String sp : existingSourcePaths) {
@ -178,7 +158,7 @@ public class GenerateEntitiesApplication {
.saveAsTextFile(targetPath, GzipCodec.class); .saveAsTextFile(targetPath, GzipCodec.class);
} }
private static List<Oaf> convertToListOaf( public static List<Oaf> convertToListOaf(
final String id, final String id,
final String s, final String s,
final boolean shouldHashId, final boolean shouldHashId,
@ -219,19 +199,6 @@ public class GenerateEntitiesApplication {
} }
} }
private static String tryApplyMapping(
final String id,
final String s,
final boolean shouldHashId,
final VocabularyGroup vocs) {
final List<Oaf> oaf = convertToListOaf(id, s, shouldHashId, vocs);
if (Optional.ofNullable(oaf).map(List::isEmpty).orElse(false)) {
return s;
}
return null;
}
private static Oaf convertFromJson(final String s, final Class<? extends Oaf> clazz) { private static Oaf convertFromJson(final String s, final Class<? extends Oaf> clazz) {
try { try {
return OBJECT_MAPPER.readValue(s, clazz); return OBJECT_MAPPER.readValue(s, clazz);

View File

@ -0,0 +1,108 @@
package eu.dnetlib.dhp.oa.graph.raw;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.Arrays;
import java.util.List;
import java.util.Objects;
import java.util.Optional;
import java.util.stream.Collectors;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.io.Text;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.rdd.RDD;
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 eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.utils.ISLookupClientFactory;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import scala.Tuple2;
public class VerifyRecordsApplication {
private static final Logger log = LoggerFactory.getLogger(VerifyRecordsApplication.class);
public static void main(final String[] args) throws Exception {
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils
.toString(
VerifyRecordsApplication.class
.getResourceAsStream("/eu/dnetlib/dhp/oa/graph/verify_records_parameters.json")));
parser.parseArgument(args);
final Boolean isSparkSessionManaged = Optional
.ofNullable(parser.get("isSparkSessionManaged"))
.map(Boolean::valueOf)
.orElse(Boolean.TRUE);
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
final String sourcePaths = parser.get("sourcePaths");
log.info("sourcePaths: {}", sourcePaths);
final String invalidPath = parser.get("invalidPath");
log.info("invalidPath: {}", invalidPath);
final String isLookupUrl = parser.get("isLookupUrl");
log.info("isLookupUrl: {}", isLookupUrl);
final ISLookUpService isLookupService = ISLookupClientFactory.getLookUpService(isLookupUrl);
final VocabularyGroup vocs = VocabularyGroup.loadVocsFromIS(isLookupService);
final SparkConf conf = new SparkConf();
runWithSparkSession(conf, isSparkSessionManaged, spark -> {
HdfsSupport.remove(invalidPath, spark.sparkContext().hadoopConfiguration());
validateRecords(spark, sourcePaths, invalidPath, vocs);
});
}
private static void validateRecords(SparkSession spark, String sourcePaths, String invalidPath,
VocabularyGroup vocs) {
final JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
final List<String> existingSourcePaths = Arrays
.stream(sourcePaths.split(","))
.filter(p -> HdfsSupport.exists(p, sc.hadoopConfiguration()))
.collect(Collectors.toList());
log.info("Verify records in files:");
existingSourcePaths.forEach(log::info);
for (final String sp : existingSourcePaths) {
RDD<String> invalidRecords = sc
.sequenceFile(sp, Text.class, Text.class)
.map(k -> tryApplyMapping(k._1().toString(), k._2().toString(), true, vocs))
.filter(Objects::nonNull)
.rdd();
spark
.createDataset(invalidRecords, Encoders.STRING())
.write()
.mode(SaveMode.Append)
.option("compression", "gzip")
.text(invalidPath);
}
}
private static String tryApplyMapping(
final String id,
final String xmlRecord,
final boolean shouldHashId,
final VocabularyGroup vocs) {
final List<Oaf> oaf = GenerateEntitiesApplication.convertToListOaf(id, xmlRecord, shouldHashId, vocs);
if (Optional.ofNullable(oaf).map(List::isEmpty).orElse(false)) {
return xmlRecord;
}
return null;
}
}

View File

@ -4,6 +4,7 @@ package eu.dnetlib.dhp.oa.graph.raw.common;
import java.io.Closeable; import java.io.Closeable;
import java.io.IOException; import java.io.IOException;
import java.util.Arrays; import java.util.Arrays;
import java.util.List;
import java.util.Set; import java.util.Set;
import java.util.concurrent.atomic.AtomicInteger; import java.util.concurrent.atomic.AtomicInteger;
import java.util.stream.Collectors; import java.util.stream.Collectors;
@ -24,8 +25,11 @@ import org.apache.http.impl.client.HttpClients;
import com.fasterxml.jackson.core.JsonProcessingException; import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper; import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.oa.graph.raw.OafToOafMapper;
import eu.dnetlib.dhp.oa.graph.raw.OdfToOafMapper;
import eu.dnetlib.dhp.schema.mdstore.MDStoreWithInfo; import eu.dnetlib.dhp.schema.mdstore.MDStoreWithInfo;
import eu.dnetlib.dhp.schema.oaf.Oaf; import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.utils.DHPUtils; import eu.dnetlib.dhp.utils.DHPUtils;
public class AbstractMigrationApplication implements Closeable { public class AbstractMigrationApplication implements Closeable {

View File

@ -17,12 +17,6 @@
"paramDescription": "the path of the target file", "paramDescription": "the path of the target file",
"paramRequired": true "paramRequired": true
}, },
{
"paramName": "i",
"paramLongName": "invalidPath",
"paramDescription": "the path of the invalid records file",
"paramRequired": false
},
{ {
"paramName": "isu", "paramName": "isu",
"paramLongName": "isLookupUrl", "paramLongName": "isLookupUrl",

View File

@ -446,10 +446,34 @@
<join name="wait_import" to="fork_generate_entities"/> <join name="wait_import" to="fork_generate_entities"/>
<fork name="fork_generate_entities"> <fork name="fork_generate_entities">
<path start="GenerateEntities_claim"/> <path start="VerifyRecords_claim"/>
<path start="GenerateEntities"/> <path start="VerifyRecords"/>
</fork> </fork>
<action name="VerifyRecords_claim">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>VerifyRecords_claim</name>
<class>eu.dnetlib.dhp.oa.graph.raw.VerifyRecordsApplication</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory ${sparkExecutorMemory}
--executor-cores ${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePaths</arg><arg>${contentPath}/db_claims,${contentPath}/oaf_claims,${contentPath}/odf_claims</arg>
<arg>--invalidPath</arg><arg>${workingDir}/invalid_records_claim</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
</spark>
<ok to="GenerateEntities_claim"/>
<error to="Kill"/>
</action>
<action name="GenerateEntities_claim"> <action name="GenerateEntities_claim">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
@ -468,7 +492,6 @@
</spark-opts> </spark-opts>
<arg>--sourcePaths</arg><arg>${contentPath}/db_claims,${contentPath}/oaf_claims,${contentPath}/odf_claims</arg> <arg>--sourcePaths</arg><arg>${contentPath}/db_claims,${contentPath}/oaf_claims,${contentPath}/odf_claims</arg>
<arg>--targetPath</arg><arg>${workingDir}/entities_claim</arg> <arg>--targetPath</arg><arg>${workingDir}/entities_claim</arg>
<arg>--invalidPath</arg><arg>${workingDir}/invalid_records_claim</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg> <arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--shouldHashId</arg><arg>${shouldHashId}</arg> <arg>--shouldHashId</arg><arg>${shouldHashId}</arg>
<arg>--mode</arg><arg>claim</arg> <arg>--mode</arg><arg>claim</arg>
@ -500,6 +523,30 @@
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<action name="VerifyRecords">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>VerifyRecords</name>
<class>eu.dnetlib.dhp.oa.graph.raw.VerifyRecordsApplication</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory ${sparkExecutorMemory}
--executor-cores ${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePaths</arg><arg>${contentPath}/db_openaire,${contentPath}/db_openorgs,${contentPath}/oaf_records,${contentPath}/odf_records,${contentPath}/oaf_records_hdfs,${contentPath}/odf_records_hdfs,${contentPath}/oaf_records_invisible</arg>
<arg>--invalidPath</arg><arg>${workingDir}/invalid_records</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
</spark>
<ok to="GenerateEntities"/>
<error to="Kill"/>
</action>
<action name="GenerateEntities"> <action name="GenerateEntities">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master> <master>yarn</master>
@ -518,7 +565,6 @@
</spark-opts> </spark-opts>
<arg>--sourcePaths</arg><arg>${contentPath}/db_openaire,${contentPath}/db_openorgs,${contentPath}/oaf_records,${contentPath}/odf_records,${contentPath}/oaf_records_hdfs,${contentPath}/odf_records_hdfs,${contentPath}/oaf_records_invisible</arg> <arg>--sourcePaths</arg><arg>${contentPath}/db_openaire,${contentPath}/db_openorgs,${contentPath}/oaf_records,${contentPath}/odf_records,${contentPath}/oaf_records_hdfs,${contentPath}/odf_records_hdfs,${contentPath}/oaf_records_invisible</arg>
<arg>--targetPath</arg><arg>${workingDir}/entities</arg> <arg>--targetPath</arg><arg>${workingDir}/entities</arg>
<arg>--invalidPath</arg><arg>${workingDir}/invalid_records</arg>
<arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg> <arg>--isLookupUrl</arg><arg>${isLookupUrl}</arg>
<arg>--shouldHashId</arg><arg>${shouldHashId}</arg> <arg>--shouldHashId</arg><arg>${shouldHashId}</arg>
</spark> </spark>

View File

@ -0,0 +1,26 @@
[
{
"paramName": "issm",
"paramLongName": "isSparkSessionManaged",
"paramDescription": "when true will stop SparkSession after job execution",
"paramRequired": false
},
{
"paramName": "s",
"paramLongName": "sourcePaths",
"paramDescription": "the HDFS source paths which contains the sequential file (comma separated)",
"paramRequired": true
},
{
"paramName": "i",
"paramLongName": "invalidPath",
"paramDescription": "the path of the invalid records file",
"paramRequired": false
},
{
"paramName": "isu",
"paramLongName": "isLookupUrl",
"paramDescription": "the url of the ISLookupService",
"paramRequired": true
}
]