forked from D-Net/dnet-hadoop
changed to query for community map just once and save the result for remaining executions
This commit is contained in:
parent
872d7783fc
commit
c892c7dfa7
|
@ -0,0 +1,87 @@
|
||||||
|
|
||||||
|
package eu.dnetlib.dhp.oa.graph.dump.community;
|
||||||
|
|
||||||
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
||||||
|
|
||||||
|
import java.io.Serializable;
|
||||||
|
import java.util.Optional;
|
||||||
|
import java.util.Set;
|
||||||
|
import java.util.stream.Collectors;
|
||||||
|
|
||||||
|
import org.apache.spark.SparkConf;
|
||||||
|
import org.apache.spark.SparkContext;
|
||||||
|
import org.apache.spark.sql.Dataset;
|
||||||
|
import org.apache.spark.sql.SaveMode;
|
||||||
|
import org.apache.spark.sql.SparkSession;
|
||||||
|
|
||||||
|
import eu.dnetlib.dhp.oa.graph.dump.Utils;
|
||||||
|
import eu.dnetlib.dhp.schema.dump.oaf.community.CommunityResult;
|
||||||
|
import org.slf4j.Logger;
|
||||||
|
import org.slf4j.LoggerFactory;
|
||||||
|
|
||||||
|
public class CommunitySplitS3 implements Serializable {
|
||||||
|
private static final Logger log = LoggerFactory.getLogger(CommunitySplitS3.class);
|
||||||
|
public void run(Boolean isSparkSessionManaged, String inputPath, String outputPath, String communityMapPath) {
|
||||||
|
// public void run(Boolean isSparkSessionManaged, String inputPath, String outputPath, CommunityMap
|
||||||
|
// communityMap) {
|
||||||
|
SparkConf conf = new SparkConf();
|
||||||
|
runWithSparkSession(
|
||||||
|
conf,
|
||||||
|
isSparkSessionManaged,
|
||||||
|
spark -> {
|
||||||
|
SparkContext sc = spark.sparkContext();
|
||||||
|
sc.hadoopConfiguration().set("fs.s3.impl", "org.apache.hadoop.fs.s3native.NativeS3FileSystem");
|
||||||
|
sc.hadoopConfiguration().set("fs.s3.awsAccessKeyId", "AK0MM6C2BYA0K1PNJYYX");
|
||||||
|
sc.hadoopConfiguration().set("fs.s3.awsSecretAccessKey", "fpeiqUUpKAUOtO6JWMWLTxxlSxJ+yGYwHozm3jHK");
|
||||||
|
execSplit(spark, inputPath, outputPath, communityMapPath); // communityMap.keySet());// ,
|
||||||
|
// inputClazz);
|
||||||
|
// execSplit(spark, inputPath, outputPath, communityMap.keySet());
|
||||||
|
});
|
||||||
|
}
|
||||||
|
|
||||||
|
private static void execSplit(SparkSession spark, String inputPath, String outputPath,
|
||||||
|
String communityMapPath) {
|
||||||
|
// Set<String> communities) {
|
||||||
|
|
||||||
|
Set<String> communities = Utils.getCommunityMap(spark, communityMapPath).keySet();
|
||||||
|
|
||||||
|
Dataset<CommunityResult> result = Utils
|
||||||
|
.readPath(spark, inputPath + "/publication", CommunityResult.class)
|
||||||
|
.union(Utils.readPath(spark, inputPath + "/dataset", CommunityResult.class))
|
||||||
|
.union(Utils.readPath(spark, inputPath + "/orp", CommunityResult.class))
|
||||||
|
.union(Utils.readPath(spark, inputPath + "/software", CommunityResult.class));
|
||||||
|
|
||||||
|
communities
|
||||||
|
.stream()
|
||||||
|
.forEach(c -> printResult(c, result, outputPath));
|
||||||
|
|
||||||
|
}
|
||||||
|
|
||||||
|
private static void printResult(String c, Dataset<CommunityResult> result, String outputPath) {
|
||||||
|
Dataset<CommunityResult> community_products = result
|
||||||
|
.filter(r -> containsCommunity(r, c));
|
||||||
|
|
||||||
|
if (community_products.count() > 0) {
|
||||||
|
log.info("Writing dump for community: {} ", c);
|
||||||
|
community_products
|
||||||
|
.repartition(1)
|
||||||
|
.write()
|
||||||
|
.option("compression", "gzip")
|
||||||
|
.mode(SaveMode.Overwrite)
|
||||||
|
.json(outputPath + "/" + c);
|
||||||
|
}
|
||||||
|
|
||||||
|
}
|
||||||
|
|
||||||
|
private static boolean containsCommunity(CommunityResult r, String c) {
|
||||||
|
if (Optional.ofNullable(r.getContext()).isPresent()) {
|
||||||
|
return r
|
||||||
|
.getContext()
|
||||||
|
.stream()
|
||||||
|
.filter(con -> con.getCode().equals(c))
|
||||||
|
.collect(Collectors.toList())
|
||||||
|
.size() > 0;
|
||||||
|
}
|
||||||
|
return false;
|
||||||
|
}
|
||||||
|
}
|
|
@ -0,0 +1,122 @@
|
||||||
|
|
||||||
|
package eu.dnetlib.dhp.oa.graph.dump.community;
|
||||||
|
|
||||||
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
||||||
|
|
||||||
|
import java.io.BufferedWriter;
|
||||||
|
import java.io.IOException;
|
||||||
|
import java.io.OutputStreamWriter;
|
||||||
|
import java.io.Serializable;
|
||||||
|
import java.nio.charset.StandardCharsets;
|
||||||
|
import java.util.Optional;
|
||||||
|
|
||||||
|
import org.apache.commons.io.IOUtils;
|
||||||
|
import org.apache.hadoop.conf.Configuration;
|
||||||
|
import org.apache.hadoop.fs.FSDataOutputStream;
|
||||||
|
import org.apache.hadoop.fs.FileSystem;
|
||||||
|
import org.apache.hadoop.fs.Path;
|
||||||
|
import org.apache.spark.SparkConf;
|
||||||
|
import org.slf4j.Logger;
|
||||||
|
import org.slf4j.LoggerFactory;
|
||||||
|
|
||||||
|
import com.fasterxml.jackson.core.JsonProcessingException;
|
||||||
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
||||||
|
|
||||||
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
||||||
|
import eu.dnetlib.dhp.oa.graph.dump.QueryInformationSystem;
|
||||||
|
import eu.dnetlib.dhp.oa.graph.dump.Utils;
|
||||||
|
import eu.dnetlib.dhp.oa.graph.dump.graph.CreateContextEntities;
|
||||||
|
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
|
||||||
|
|
||||||
|
public class SaveCommunityMap implements Serializable {
|
||||||
|
|
||||||
|
private static final Logger log = LoggerFactory.getLogger(SaveCommunityMap.class);
|
||||||
|
private final QueryInformationSystem queryInformationSystem;
|
||||||
|
|
||||||
|
private final Configuration conf;
|
||||||
|
private final BufferedWriter writer;
|
||||||
|
|
||||||
|
public SaveCommunityMap(String hdfsPath, String hdfsNameNode, String isLookUpUrl) throws IOException {
|
||||||
|
conf = new Configuration();
|
||||||
|
conf.set("fs.defaultFS", hdfsNameNode);
|
||||||
|
FileSystem fileSystem = FileSystem.get(conf);
|
||||||
|
Path hdfsWritePath = new Path(hdfsPath);
|
||||||
|
FSDataOutputStream fsDataOutputStream = null;
|
||||||
|
if (fileSystem.exists(hdfsWritePath)) {
|
||||||
|
fsDataOutputStream = fileSystem.append(hdfsWritePath);
|
||||||
|
} else {
|
||||||
|
fsDataOutputStream = fileSystem.create(hdfsWritePath);
|
||||||
|
}
|
||||||
|
|
||||||
|
queryInformationSystem = new QueryInformationSystem();
|
||||||
|
queryInformationSystem.setIsLookUp(Utils.getIsLookUpService(isLookUpUrl));
|
||||||
|
|
||||||
|
writer = new BufferedWriter(new OutputStreamWriter(fsDataOutputStream, StandardCharsets.UTF_8));
|
||||||
|
|
||||||
|
}
|
||||||
|
|
||||||
|
public static void main(String[] args) throws Exception {
|
||||||
|
String jsonConfiguration = IOUtils
|
||||||
|
.toString(
|
||||||
|
SaveCommunityMap.class
|
||||||
|
.getResourceAsStream(
|
||||||
|
"/eu/dnetlib/dhp/oa/graph/dump/input_cm_parameters.json"));
|
||||||
|
|
||||||
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
|
||||||
|
parser.parseArgument(args);
|
||||||
|
|
||||||
|
final String nameNode = parser.get("nameNode");
|
||||||
|
log.info("nameNode: {}", nameNode);
|
||||||
|
|
||||||
|
final String outputPath = parser.get("outputPath");
|
||||||
|
log.info("outputPath: {}", outputPath);
|
||||||
|
|
||||||
|
final String isLookUpUrl = parser.get("isLookUpUrl");
|
||||||
|
log.info("isLookUpUrl: {}", isLookUpUrl);
|
||||||
|
|
||||||
|
// Boolean isSparkSessionManaged = Optional
|
||||||
|
// .ofNullable(parser.get("isSparkSessionManaged"))
|
||||||
|
// .map(Boolean::valueOf)
|
||||||
|
// .orElse(Boolean.TRUE);
|
||||||
|
// log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
||||||
|
|
||||||
|
final SaveCommunityMap scm = new SaveCommunityMap(outputPath, nameNode, isLookUpUrl);
|
||||||
|
|
||||||
|
scm.saveCommunityMap();
|
||||||
|
|
||||||
|
// CommunityMap communityMap = queryInformationSystem.getCommunityMap();
|
||||||
|
|
||||||
|
// SparkConf conf = new SparkConf();
|
||||||
|
//
|
||||||
|
// runWithSparkSession(
|
||||||
|
// conf,
|
||||||
|
// isSparkSessionManaged,
|
||||||
|
// spark -> {
|
||||||
|
// Utils.removeOutputDir(spark, outputPath);
|
||||||
|
//
|
||||||
|
//// execDump(spark, inputPath, outputPath, communityMap, inputClazz, outputClazz, graph);// ,
|
||||||
|
// // dumpClazz);
|
||||||
|
// });
|
||||||
|
|
||||||
|
// Configuration conf = new Configuration();
|
||||||
|
// conf.set("fs.defaultFS", nameNode);
|
||||||
|
// FileSystem fileSystem = FileSystem.get(conf);
|
||||||
|
// Path hdfsWritePath = new Path(outputPath);
|
||||||
|
// FSDataOutputStream fsDataOutputStream = null;
|
||||||
|
// if (fileSystem.exists(hdfsWritePath)) {
|
||||||
|
// fsDataOutputStream = fileSystem.append(hdfsWritePath);
|
||||||
|
// } else {
|
||||||
|
// fsDataOutputStream = fileSystem.create(hdfsWritePath);
|
||||||
|
// }
|
||||||
|
//
|
||||||
|
// BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(fsDataOutputStream, StandardCharsets.UTF_8));
|
||||||
|
//
|
||||||
|
// writer.write(OBJECT_MAPPER.writeValueAsString(communityMap));
|
||||||
|
// writer.close();
|
||||||
|
}
|
||||||
|
|
||||||
|
private void saveCommunityMap() throws ISLookUpException, IOException {
|
||||||
|
writer.write(Utils.OBJECT_MAPPER.writeValueAsString(queryInformationSystem.getCommunityMap()));
|
||||||
|
writer.close();
|
||||||
|
}
|
||||||
|
}
|
|
@ -45,6 +45,8 @@ public class SparkDumpCommunityProducts implements Serializable {
|
||||||
final String resultClassName = parser.get("resultTableName");
|
final String resultClassName = parser.get("resultTableName");
|
||||||
log.info("resultTableName: {}", resultClassName);
|
log.info("resultTableName: {}", resultClassName);
|
||||||
|
|
||||||
|
String communityMapPath = parser.get("communityMapPath");
|
||||||
|
|
||||||
final String isLookUpUrl = parser.get("isLookUpUrl");
|
final String isLookUpUrl = parser.get("isLookUpUrl");
|
||||||
log.info("isLookUpUrl: {}", isLookUpUrl);
|
log.info("isLookUpUrl: {}", isLookUpUrl);
|
||||||
|
|
||||||
|
@ -56,7 +58,15 @@ public class SparkDumpCommunityProducts implements Serializable {
|
||||||
|
|
||||||
DumpProducts dump = new DumpProducts();
|
DumpProducts dump = new DumpProducts();
|
||||||
|
|
||||||
dump.run(isSparkSessionManaged, inputPath, outputPath, communityMap, inputClazz, CommunityResult.class, false);
|
dump
|
||||||
|
.run(
|
||||||
|
isSparkSessionManaged, inputPath, outputPath, communityMapPath, inputClazz, CommunityResult.class,
|
||||||
|
false);
|
||||||
|
|
||||||
|
// dump
|
||||||
|
// .run(
|
||||||
|
// isSparkSessionManaged, inputPath, outputPath, communityMap, inputClazz, CommunityResult.class,
|
||||||
|
// false);
|
||||||
|
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
|
@ -0,0 +1,64 @@
|
||||||
|
|
||||||
|
package eu.dnetlib.dhp.oa.graph.dump.community;
|
||||||
|
|
||||||
|
import java.io.Serializable;
|
||||||
|
import java.util.Optional;
|
||||||
|
|
||||||
|
import org.apache.commons.io.IOUtils;
|
||||||
|
import org.slf4j.Logger;
|
||||||
|
import org.slf4j.LoggerFactory;
|
||||||
|
|
||||||
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
||||||
|
import eu.dnetlib.dhp.oa.graph.dump.QueryInformationSystem;
|
||||||
|
import eu.dnetlib.dhp.utils.ISLookupClientFactory;
|
||||||
|
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
|
||||||
|
|
||||||
|
public class SparkSplitForCommunityS3 implements Serializable {
|
||||||
|
|
||||||
|
private static final Logger log = LoggerFactory.getLogger(SparkSplitForCommunityS3.class);
|
||||||
|
|
||||||
|
public static void main(String[] args) throws Exception {
|
||||||
|
String jsonConfiguration = IOUtils
|
||||||
|
.toString(
|
||||||
|
SparkSplitForCommunityS3.class
|
||||||
|
.getResourceAsStream(
|
||||||
|
"/eu/dnetlib/dhp/oa/graph/dump/split_parameters.json"));
|
||||||
|
|
||||||
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
|
||||||
|
parser.parseArgument(args);
|
||||||
|
|
||||||
|
Boolean isSparkSessionManaged = Optional
|
||||||
|
.ofNullable(parser.get("isSparkSessionManaged"))
|
||||||
|
.map(Boolean::valueOf)
|
||||||
|
.orElse(Boolean.TRUE);
|
||||||
|
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
||||||
|
|
||||||
|
final String inputPath = parser.get("sourcePath");
|
||||||
|
log.info("inputPath: {}", inputPath);
|
||||||
|
|
||||||
|
final String outputPath = parser.get("outputPath");
|
||||||
|
log.info("outputPath: {}", outputPath);
|
||||||
|
|
||||||
|
final String communityMapPath = parser.get("communityMapPath");
|
||||||
|
|
||||||
|
final String isLookUpUrl = parser.get("isLookUpUrl");
|
||||||
|
log.info("isLookUpUrl: {}", isLookUpUrl);
|
||||||
|
|
||||||
|
CommunitySplitS3 split = new CommunitySplitS3();
|
||||||
|
|
||||||
|
// CommunityMap communityMap;
|
||||||
|
|
||||||
|
// QueryInformationSystem queryInformationSystem = new QueryInformationSystem();
|
||||||
|
// queryInformationSystem.setIsLookUp(getIsLookUpService(isLookUpUrl));
|
||||||
|
// communityMap = queryInformationSystem.getCommunityMap();
|
||||||
|
|
||||||
|
split.run(isSparkSessionManaged, inputPath, outputPath, communityMapPath);
|
||||||
|
// split.run(isSparkSessionManaged, inputPath, outputPath, communityMap);
|
||||||
|
|
||||||
|
}
|
||||||
|
|
||||||
|
public static ISLookUpService getIsLookUpService(String isLookUpUrl) {
|
||||||
|
return ISLookupClientFactory.getLookUpService(isLookUpUrl);
|
||||||
|
}
|
||||||
|
|
||||||
|
}
|
|
@ -0,0 +1,36 @@
|
||||||
|
|
||||||
|
[
|
||||||
|
{
|
||||||
|
"paramName":"cmp",
|
||||||
|
"paramLongName":"communityMapPath",
|
||||||
|
"paramDescription": "the path to the serialization of the community map",
|
||||||
|
"paramRequired": false
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"paramName":"is",
|
||||||
|
"paramLongName":"isLookUpUrl",
|
||||||
|
"paramDescription": "URL of the isLookUp Service",
|
||||||
|
"paramRequired": true
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"paramName":"s",
|
||||||
|
"paramLongName":"sourcePath",
|
||||||
|
"paramDescription": "the path of the sequencial file to read",
|
||||||
|
"paramRequired": true
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"paramName": "out",
|
||||||
|
"paramLongName": "outputPath",
|
||||||
|
"paramDescription": "the path used to store temporary output files",
|
||||||
|
"paramRequired": true
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"paramName": "ssm",
|
||||||
|
"paramLongName": "isSparkSessionManaged",
|
||||||
|
"paramDescription": "true if the spark session is managed, false otherwise",
|
||||||
|
"paramRequired": false
|
||||||
|
}
|
||||||
|
]
|
||||||
|
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue