forked from D-Net/dnet-hadoop
spark job to download orcid record modified after a fixed date
This commit is contained in:
parent
12756f9d41
commit
0b29bb7e3b
|
@ -27,8 +27,8 @@ import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|||
public class OrcidDownloader extends OrcidDSManager {
|
||||
|
||||
static final int REQ_LIMIT = 24;
|
||||
static final int REQ_MAX_TEST = 100;
|
||||
static final int RECORD_PARSED_COUNTER_LOG_INTERVAL = 10;
|
||||
// static final int REQ_MAX_TEST = 100;
|
||||
static final int RECORD_PARSED_COUNTER_LOG_INTERVAL = 10000;
|
||||
static final String DATE_FORMAT = "yyyy-MM-dd HH:mm:ss";
|
||||
static final String lastUpdate = "2019-09-30 00:00:00";
|
||||
private String lambdaFileName;
|
||||
|
@ -136,9 +136,9 @@ public class OrcidDownloader extends OrcidDSManager {
|
|||
}
|
||||
}
|
||||
|
||||
if (parsedRecordsCounter > REQ_MAX_TEST) {
|
||||
break;
|
||||
}
|
||||
// if (parsedRecordsCounter > REQ_MAX_TEST) {
|
||||
// break;
|
||||
// }
|
||||
if ((parsedRecordsCounter % RECORD_PARSED_COUNTER_LOG_INTERVAL) == 0) {
|
||||
Log
|
||||
.info(
|
||||
|
@ -148,9 +148,9 @@ public class OrcidDownloader extends OrcidDSManager {
|
|||
+ downloadedRecordsCounter
|
||||
+ " saved: "
|
||||
+ savedRecordsCounter);
|
||||
if (parsedRecordsCounter > REQ_MAX_TEST) {
|
||||
break;
|
||||
}
|
||||
// if (parsedRecordsCounter > REQ_MAX_TEST) {
|
||||
// break;
|
||||
// }
|
||||
}
|
||||
}
|
||||
long endDownload = System.currentTimeMillis();
|
||||
|
|
|
@ -3,53 +3,128 @@ package eu.dnetlib.doiboost.orcid;
|
|||
|
||||
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
||||
|
||||
import java.io.IOException;
|
||||
import java.text.SimpleDateFormat;
|
||||
import java.util.Date;
|
||||
import java.util.Optional;
|
||||
|
||||
import org.apache.commons.io.IOUtils;
|
||||
import org.apache.hadoop.io.Text;
|
||||
import org.apache.http.client.methods.CloseableHttpResponse;
|
||||
import org.apache.http.client.methods.HttpGet;
|
||||
import org.apache.http.impl.client.CloseableHttpClient;
|
||||
import org.apache.http.impl.client.HttpClients;
|
||||
import org.apache.spark.SparkConf;
|
||||
import org.apache.spark.sql.Dataset;
|
||||
import org.apache.spark.sql.Row;
|
||||
import org.apache.spark.api.java.JavaRDD;
|
||||
import org.apache.spark.api.java.JavaSparkContext;
|
||||
import org.apache.spark.api.java.function.Function;
|
||||
import org.apache.spark.sql.Encoders;
|
||||
import org.apache.spark.sql.SaveMode;
|
||||
import org.mortbay.log.Log;
|
||||
import org.slf4j.Logger;
|
||||
import org.slf4j.LoggerFactory;
|
||||
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
||||
import eu.dnetlib.doiboost.orcid.model.DownloadedRecordData;
|
||||
import scala.Tuple2;
|
||||
|
||||
public class SparkOrcidGenerateAuthors {
|
||||
|
||||
public static void main(String[] args) {
|
||||
static final String DATE_FORMAT = "yyyy-MM-dd HH:mm:ss";
|
||||
static final String lastUpdate = "2019-09-30 00:00:00";
|
||||
|
||||
public static void main(String[] args) throws IOException, Exception {
|
||||
Logger logger = LoggerFactory.getLogger(SparkOrcidGenerateAuthors.class);
|
||||
logger.info("[ SparkOrcidGenerateAuthors STARTED]");
|
||||
try {
|
||||
|
||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
||||
IOUtils
|
||||
.toString(
|
||||
SparkOrcidGenerateAuthors.class
|
||||
.getResourceAsStream(
|
||||
"/eu/dnetlib/dhp/doiboost/gen_orcid_authors_parameters.json")));
|
||||
parser.parseArgument(args);
|
||||
Boolean isSparkSessionManaged = Optional
|
||||
.ofNullable(parser.get("isSparkSessionManaged"))
|
||||
.map(Boolean::valueOf)
|
||||
.orElse(Boolean.TRUE);
|
||||
logger.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
||||
final String workingDirPath = parser.get("workingPath_orcid");
|
||||
logger.info("workingDirPath: ", workingDirPath);
|
||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
||||
IOUtils
|
||||
.toString(
|
||||
SparkOrcidGenerateAuthors.class
|
||||
.getResourceAsStream(
|
||||
"/eu/dnetlib/dhp/doiboost/gen_orcid_authors_parameters.json")));
|
||||
parser.parseArgument(args);
|
||||
Boolean isSparkSessionManaged = Optional
|
||||
.ofNullable(parser.get("isSparkSessionManaged"))
|
||||
.map(Boolean::valueOf)
|
||||
.orElse(Boolean.TRUE);
|
||||
logger.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
||||
final String workingPath = parser.get("workingPath");
|
||||
logger.info("workingPath: ", workingPath);
|
||||
final String outputAuthorsPath = parser.get("outputAuthorsPath");
|
||||
logger.info("outputAuthorsPath: ", outputAuthorsPath);
|
||||
final String token = parser.get("token");
|
||||
|
||||
SparkConf conf = new SparkConf();
|
||||
runWithSparkSession(
|
||||
conf,
|
||||
isSparkSessionManaged,
|
||||
spark -> {
|
||||
Dataset<Row> lambda = spark.read().load(workingDirPath + "last_modified.csv");
|
||||
logger.info("lambda file loaded.");
|
||||
});
|
||||
SparkConf conf = new SparkConf();
|
||||
runWithSparkSession(
|
||||
conf,
|
||||
isSparkSessionManaged,
|
||||
spark -> {
|
||||
JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
|
||||
JavaRDD<String> lamdaFileRDD = sc.textFile(workingPath + "last_modified.csv");
|
||||
Function<String, Boolean> isModifiedAfterFilter = line -> {
|
||||
String[] values = line.split(",");
|
||||
String orcidId = values[0];
|
||||
if (isModified(orcidId, values[3])) {
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
};
|
||||
Function<String, Tuple2<String, String>> downloadRecordFunction = line -> {
|
||||
String[] values = line.split(",");
|
||||
String orcidId = values[0];
|
||||
return downloadRecord(orcidId, token);
|
||||
};
|
||||
|
||||
} catch (Exception e) {
|
||||
|
||||
logger.info("****************************** " + e.getMessage());
|
||||
}
|
||||
lamdaFileRDD
|
||||
.filter(isModifiedAfterFilter)
|
||||
.map(downloadRecordFunction)
|
||||
.rdd()
|
||||
.saveAsTextFile(workingPath.concat(outputAuthorsPath));
|
||||
});
|
||||
|
||||
}
|
||||
|
||||
private static boolean isModified(String orcidId, String modifiedDate) {
|
||||
Date modifiedDateDt = null;
|
||||
Date lastUpdateDt = null;
|
||||
try {
|
||||
if (modifiedDate.length() != 19) {
|
||||
modifiedDate = modifiedDate.substring(0, 19);
|
||||
}
|
||||
modifiedDateDt = new SimpleDateFormat(DATE_FORMAT).parse(modifiedDate);
|
||||
lastUpdateDt = new SimpleDateFormat(DATE_FORMAT).parse(lastUpdate);
|
||||
} catch (Exception e) {
|
||||
Log.warn("[" + orcidId + "] Parsing date: ", e.getMessage());
|
||||
return true;
|
||||
}
|
||||
return modifiedDateDt.after(lastUpdateDt);
|
||||
}
|
||||
|
||||
private static Tuple2<String, String> downloadRecord(String orcidId, String token) {
|
||||
final DownloadedRecordData data = new DownloadedRecordData();
|
||||
data.setOrcidId(orcidId);
|
||||
try (CloseableHttpClient client = HttpClients.createDefault()) {
|
||||
HttpGet httpGet = new HttpGet("https://api.orcid.org/v3.0/" + orcidId + "/record");
|
||||
httpGet.addHeader("Accept", "application/vnd.orcid+xml");
|
||||
httpGet.addHeader("Authorization", String.format("Bearer %s", token));
|
||||
CloseableHttpResponse response = client.execute(httpGet);
|
||||
int statusCode = response.getStatusLine().getStatusCode();
|
||||
data.setStatusCode(statusCode);
|
||||
if (statusCode != 200) {
|
||||
Log
|
||||
.warn(
|
||||
"Downloading " + orcidId + " status code: " + response.getStatusLine().getStatusCode());
|
||||
return data.toTuple2();
|
||||
}
|
||||
data
|
||||
.setCompressedData(
|
||||
ArgumentApplicationParser.compressArgument(IOUtils.toString(response.getEntity().getContent())));
|
||||
} catch (Throwable e) {
|
||||
Log.warn("Downloading " + orcidId, e.getMessage());
|
||||
data.setErrorMessage(e.getMessage());
|
||||
return data.toTuple2();
|
||||
}
|
||||
return data.toTuple2();
|
||||
}
|
||||
}
|
||||
|
|
|
@ -0,0 +1,63 @@
|
|||
|
||||
package eu.dnetlib.doiboost.orcid.model;
|
||||
|
||||
import java.io.Serializable;
|
||||
|
||||
import org.apache.hadoop.io.Text;
|
||||
|
||||
import com.google.gson.JsonObject;
|
||||
|
||||
import scala.Tuple2;
|
||||
|
||||
public class DownloadedRecordData implements Serializable {
|
||||
|
||||
private String orcidId;
|
||||
private String statusCode;
|
||||
private String compressedData;
|
||||
private String errorMessage;
|
||||
|
||||
public Tuple2<String, String> toTuple2() {
|
||||
JsonObject data = new JsonObject();
|
||||
data.addProperty("statusCode", getStatusCode());
|
||||
if (getCompressedData() != null) {
|
||||
data.addProperty("compressedData", getCompressedData());
|
||||
}
|
||||
if (getErrorMessage() != null) {
|
||||
data.addProperty("errorMessage", getErrorMessage());
|
||||
}
|
||||
return new Tuple2<>(orcidId, data.toString());
|
||||
}
|
||||
|
||||
public String getErrorMessage() {
|
||||
return errorMessage;
|
||||
}
|
||||
|
||||
public void setErrorMessage(String errorMessage) {
|
||||
this.errorMessage = errorMessage;
|
||||
}
|
||||
|
||||
public String getOrcidId() {
|
||||
return orcidId;
|
||||
}
|
||||
|
||||
public void setOrcidId(String orcidId) {
|
||||
this.orcidId = orcidId;
|
||||
}
|
||||
|
||||
public int getStatusCode() {
|
||||
return Integer.parseInt(statusCode);
|
||||
}
|
||||
|
||||
public void setStatusCode(int statusCode) {
|
||||
this.statusCode = Integer.toString(statusCode);
|
||||
}
|
||||
|
||||
public String getCompressedData() {
|
||||
return compressedData;
|
||||
}
|
||||
|
||||
public void setCompressedData(String compressedData) {
|
||||
this.compressedData = compressedData;
|
||||
}
|
||||
|
||||
}
|
|
@ -1,4 +1,4 @@
|
|||
[{"paramName": "mt","paramLongName": "master","paramDescription": "should be local or yarn","paramRequired": true},
|
||||
{"paramName":"d", "paramLongName":"workingPath_orcid", "paramDescription": "the default work path", "paramRequired": true},
|
||||
{"paramName":"o", "paramLongName":"outputAuthorsPath", "paramDescription": "the relative folder of the sequencial file to write", "paramRequired": true}
|
||||
[{"paramName":"w", "paramLongName":"workingPath", "paramDescription": "the working path", "paramRequired": true},
|
||||
{"paramName":"t", "paramLongName":"token", "paramDescription": "token to grant access", "paramRequired": true},
|
||||
{"paramName":"o", "paramLongName":"outputAuthorsPath", "paramDescription": "the relative folder of the sequencial file to write the authors data", "paramRequired": true}
|
||||
]
|
|
@ -1,9 +1,13 @@
|
|||
<workflow-app name="Gen Orcid Authors" xmlns="uri:oozie:workflow:0.5">
|
||||
<parameters>
|
||||
<property>
|
||||
<name>workingPath_activities</name>
|
||||
<name>workingPath</name>
|
||||
<description>the working dir base path</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>token</name>
|
||||
<description>access token</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>sparkDriverMemory</name>
|
||||
<description>memory for driver process</description>
|
||||
|
@ -28,7 +32,6 @@
|
|||
<action name="ResetWorkingPath">
|
||||
<fs>
|
||||
<delete path='${workingPath_activities}/authors'/>
|
||||
<mkdir path='${workingPath_activities}/authors'/>
|
||||
</fs>
|
||||
<ok to="Gen_Orcid_Authors"/>
|
||||
<error to="Kill"/>
|
||||
|
@ -43,10 +46,11 @@
|
|||
<name>Gen_Orcid_Authors</name>
|
||||
<class>eu.dnetlib.doiboost.orcid.SparkOrcidGenerateAuthors</class>
|
||||
<jar>dhp-doiboost-1.2.1-SNAPSHOT.jar</jar>
|
||||
<spark-opts>--num-executors 50 --conf spark.yarn.jars="hdfs://hadoop-rm1.garr-pa1.d4science.org:8020/user/oozie/share/lib/lib_20180405103059/spark2" --executor-memory=${sparkExecutorMemory} --executor-cores=${sparkExecutorCores} --driver-memory=${sparkDriverMemory}
|
||||
<spark-opts>--num-executors 24 --conf spark.yarn.jars="hdfs://hadoop-rm1.garr-pa1.d4science.org:8020/user/oozie/share/lib/lib_20180405103059/spark2" --executor-memory=${sparkExecutorMemory} --executor-cores=${sparkExecutorCores} --driver-memory=${sparkDriverMemory}
|
||||
</spark-opts>
|
||||
<arg>-mt</arg> <arg>yarn</arg>
|
||||
<arg>--workingPath_orcid</arg><arg>${workingPath_activities}/</arg>
|
||||
<arg>-w</arg><arg>${workingPath}/</arg>
|
||||
<arg>-o</arg><arg>authors/</arg>
|
||||
<arg>-t</arg><arg>${token}</arg>
|
||||
</spark>
|
||||
<ok to="End"/>
|
||||
<error to="Kill"/>
|
||||
|
|
File diff suppressed because one or more lines are too long
|
@ -0,0 +1,136 @@
|
|||
|
||||
package eu.dnetlib.doiboost.orcid;
|
||||
|
||||
import static org.junit.jupiter.api.Assertions.assertTrue;
|
||||
|
||||
import java.io.BufferedReader;
|
||||
import java.io.IOException;
|
||||
import java.io.InputStreamReader;
|
||||
import java.text.ParseException;
|
||||
import java.text.SimpleDateFormat;
|
||||
import java.util.Arrays;
|
||||
import java.util.Date;
|
||||
import java.util.List;
|
||||
|
||||
import org.apache.commons.io.IOUtils;
|
||||
import org.apache.http.client.methods.CloseableHttpResponse;
|
||||
import org.apache.http.client.methods.HttpGet;
|
||||
import org.apache.http.impl.client.CloseableHttpClient;
|
||||
import org.apache.http.impl.client.HttpClients;
|
||||
import org.junit.jupiter.api.Test;
|
||||
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
||||
|
||||
public class OrcidClientTest {
|
||||
final String orcidId = "0000-0001-7291-3210";
|
||||
final int REQ_LIMIT = 24;
|
||||
final int REQ_MAX_TEST = 100;
|
||||
final int RECORD_DOWNLOADED_COUNTER_LOG_INTERVAL = 10;
|
||||
final String DATE_FORMAT = "yyyy-MM-dd HH:mm:ss";
|
||||
final String toRetrieveDate = "2020-05-06 23:59:46.031145";
|
||||
String toNotRetrieveDate = "2019-09-29 23:59:59.000000";
|
||||
String lastUpdate = "2019-09-30 00:00:00";
|
||||
String shortDate = "2020-05-06 16:06:11";
|
||||
|
||||
// curl -i -H "Accept: application/vnd.orcid+xml"
|
||||
// -H 'Authorization: Bearer 78fdb232-7105-4086-8570-e153f4198e3d'
|
||||
// 'https://api.orcid.org/v3.0/0000-0001-7291-3210/record'
|
||||
|
||||
public String testDownloadRecord(String orcidId) throws Exception {
|
||||
try (CloseableHttpClient client = HttpClients.createDefault()) {
|
||||
HttpGet httpGet = new HttpGet("https://api.orcid.org/v3.0/" + orcidId + "/record");
|
||||
httpGet.addHeader("Accept", "application/vnd.orcid+xml");
|
||||
httpGet.addHeader("Authorization", "Bearer 78fdb232-7105-4086-8570-e153f4198e3d");
|
||||
CloseableHttpResponse response = client.execute(httpGet);
|
||||
if (response.getStatusLine().getStatusCode() != 200) {
|
||||
System.out
|
||||
.println("Downloading " + orcidId + " status code: " + response.getStatusLine().getStatusCode());
|
||||
}
|
||||
return IOUtils.toString(response.getEntity().getContent());
|
||||
} catch (Throwable e) {
|
||||
e.printStackTrace();
|
||||
}
|
||||
return new String("");
|
||||
}
|
||||
|
||||
// @Test
|
||||
public void testLambdaFileParser() throws Exception {
|
||||
try (BufferedReader br = new BufferedReader(
|
||||
new InputStreamReader(this.getClass().getResourceAsStream("last_modified.csv")))) {
|
||||
String line;
|
||||
int counter = 0;
|
||||
int nReqTmp = 0;
|
||||
long startDownload = System.currentTimeMillis();
|
||||
long startReqTmp = System.currentTimeMillis();
|
||||
while ((line = br.readLine()) != null) {
|
||||
counter++;
|
||||
// skip headers line
|
||||
if (counter == 1) {
|
||||
continue;
|
||||
}
|
||||
String[] values = line.split(",");
|
||||
List<String> recordInfo = Arrays.asList(values);
|
||||
testDownloadRecord(recordInfo.get(0));
|
||||
long endReq = System.currentTimeMillis();
|
||||
nReqTmp++;
|
||||
if (nReqTmp == REQ_LIMIT) {
|
||||
long reqSessionDuration = endReq - startReqTmp;
|
||||
if (reqSessionDuration <= 1000) {
|
||||
System.out
|
||||
.println(
|
||||
"\nreqSessionDuration: " + reqSessionDuration + " nReqTmp: " + nReqTmp + " wait ....");
|
||||
Thread.sleep(1000 - reqSessionDuration);
|
||||
} else {
|
||||
nReqTmp = 0;
|
||||
startReqTmp = System.currentTimeMillis();
|
||||
}
|
||||
}
|
||||
|
||||
if (counter > REQ_MAX_TEST) {
|
||||
break;
|
||||
}
|
||||
if ((counter % RECORD_DOWNLOADED_COUNTER_LOG_INTERVAL) == 0) {
|
||||
System.out.println("Current record downloaded: " + counter);
|
||||
}
|
||||
}
|
||||
long endDownload = System.currentTimeMillis();
|
||||
long downloadTime = endDownload - startDownload;
|
||||
System.out.println("Download time: " + ((downloadTime / 1000) / 60) + " minutes");
|
||||
}
|
||||
}
|
||||
|
||||
// @Test
|
||||
public void getRecordDatestamp() throws ParseException {
|
||||
Date toRetrieveDateDt = new SimpleDateFormat(DATE_FORMAT).parse(toRetrieveDate);
|
||||
Date toNotRetrieveDateDt = new SimpleDateFormat(DATE_FORMAT).parse(toNotRetrieveDate);
|
||||
Date lastUpdateDt = new SimpleDateFormat(DATE_FORMAT).parse(lastUpdate);
|
||||
assertTrue(toRetrieveDateDt.after(lastUpdateDt));
|
||||
assertTrue(!toNotRetrieveDateDt.after(lastUpdateDt));
|
||||
}
|
||||
|
||||
public void testDate(String value) throws ParseException {
|
||||
System.out.println(value.toString());
|
||||
if (value.length() != 19) {
|
||||
value = value.substring(0, 19);
|
||||
}
|
||||
Date valueDt = new SimpleDateFormat(DATE_FORMAT).parse(value);
|
||||
System.out.println(valueDt.toString());
|
||||
}
|
||||
|
||||
// @Test
|
||||
public void testModifiedDate() throws ParseException {
|
||||
testDate(toRetrieveDate);
|
||||
testDate(toNotRetrieveDate);
|
||||
testDate(shortDate);
|
||||
}
|
||||
|
||||
// @Test
|
||||
public void testReadBase64CompressedRecord() throws Exception {
|
||||
final String base64CompressedRecord = IOUtils
|
||||
.toString(getClass().getResourceAsStream("0000-0001-6645-509X.compressed.base64"));
|
||||
final String recordFromSeqFile = ArgumentApplicationParser.decompressValue(base64CompressedRecord);
|
||||
System.out.println(recordFromSeqFile);
|
||||
final String downloadedRecord = testDownloadRecord("0000-0001-6645-509X");
|
||||
assertTrue(recordFromSeqFile.equals(downloadedRecord));
|
||||
}
|
||||
}
|
|
@ -0,0 +1 @@
|
|||
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
|
Loading…
Reference in New Issue