D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

merged from the osfPreprints_plugin branch

This commit is contained in:
Claudio Atzori 2024-09-30 11:24:17 +02:00
parent d175a9745f fa2532db30
commit 4e9f64e01a
7 changed files with 60 additions and 85 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/MergeEntitiesComparator.java
View File

@ -1,4 +1,3 @@
package eu.dnetlib.dhp.schema.oaf.utils;
import java.util.*;

9
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/CollectorWorker.java
View File

@ -69,12 +69,9 @@ public class CollectorWorker extends ReportingJob {
scheduleReport(counter);
try (SequenceFile.Writer writer = SequenceFile
.createWriter(
this.fileSystem.getConf(), SequenceFile.Writer.file(new Path(outputPath)), SequenceFile.Writer
.keyClass(IntWritable.class),
SequenceFile.Writer
.valueClass(Text.class),
SequenceFile.Writer.compression(SequenceFile.CompressionType.BLOCK, new DeflateCodec()))) {
.createWriter(this.fileSystem.getConf(), SequenceFile.Writer.file(new Path(outputPath)), SequenceFile.Writer
.keyClass(IntWritable.class), SequenceFile.Writer
.valueClass(Text.class), SequenceFile.Writer.compression(SequenceFile.CompressionType.BLOCK, new DeflateCodec()))) {
final IntWritable key = new IntWritable(counter.get());
final Text value = new Text();
plugin

4
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/osf/OsfPreprintsCollectorPlugin.java
View File

@ -36,9 +36,7 @@ public class OsfPreprintsCollectorPlugin implements CollectorPlugin {
.map(s -> NumberUtils.toInt(s, PAGE_SIZE_VALUE_DEFAULT))
.orElse(PAGE_SIZE_VALUE_DEFAULT);
if (StringUtils.isBlank(baseUrl)) {
throw new CollectorException("Param 'baseUrl' is null or empty");
}
if (StringUtils.isBlank(baseUrl)) { throw new CollectorException("Param 'baseUrl' is null or empty"); }
final OsfPreprintsIterator it = new OsfPreprintsIterator(baseUrl, pageSize, getClientParams());

11
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/osf/OsfPreprintsIterator.java
View File

@ -54,8 +54,7 @@ public class OsfPreprintsIterator implements Iterator<String> {
@Override
public boolean hasNext() {
synchronized (this.recordQueue) {
while (this.recordQueue.isEmpty() && StringUtils.isNotBlank(this.currentUrl)
&& this.currentUrl.startsWith("http")) {
while (this.recordQueue.isEmpty() && StringUtils.isNotBlank(this.currentUrl) && this.currentUrl.startsWith("http")) {
try {
this.currentUrl = downloadPage(this.currentUrl);
} catch (final CollectorException e) {
@ -64,9 +63,7 @@ public class OsfPreprintsIterator implements Iterator<String> {
}
}
if (!this.recordQueue.isEmpty()) {
return true;
}
if (!this.recordQueue.isEmpty()) { return true; }
return false;
}
@ -115,9 +112,7 @@ public class OsfPreprintsIterator implements Iterator<String> {
}
private Document downloadUrl(final String url, final int attempt) throws CollectorException {
if (attempt > MAX_ATTEMPTS) {
throw new CollectorException("Max Number of attempts reached, url:" + url);
}
if (attempt > MAX_ATTEMPTS) { throw new CollectorException("Max Number of attempts reached, url:" + url); }
if (attempt > 0) {
final int delay = (attempt * 5000);

12
dhp-workflows/dhp-aggregation/src/main/resources/eu/dnetlib/dhp/actionmanager/bipaffiliations/input_actionset_parameter.json
View File

@ -35,18 +35,6 @@
"paramRequired": true
}
,
{
"paramName": "wip",
"paramLongName": "webCrawlInputPath",
"paramDescription": "the path to get the input data from Web Crawl",
"paramRequired": true
},
{
"paramName": "pip",
"paramLongName": "publisherInputPath",
"paramDescription": "the path to get the input data from publishers",
"paramRequired": true
},
{
"paramName": "o",
"paramLongName": "outputPath",

6
dhp-workflows/dhp-aggregation/src/test/java/eu/dnetlib/dhp/collection/plugin/osf/OsfPreprintsCollectorPluginTest.java
View File

@ -50,8 +50,7 @@ public class OsfPreprintsCollectorPluginTest {
@Test
@Disabled
void test_one() throws CollectorException {
this.plugin
.collect(this.api, new AggregatorReport())
this.plugin.collect(this.api, new AggregatorReport())
.limit(1)
.forEach(log::info);
}
@ -96,8 +95,7 @@ public class OsfPreprintsCollectorPluginTest {
final HttpConnector2 connector = new HttpConnector2();
try {
final String res = connector
.getInputSource("https://api.osf.io/v2/preprints/ydtzx/contributors/?format=json");
final String res = connector.getInputSource("https://api.osf.io/v2/preprints/ydtzx/contributors/?format=json");
System.out.println(res);
fail();
} catch (final Throwable e) {

2
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/jpath/JsonPathTest.java
View File

@ -29,7 +29,7 @@ class JsonPathTest {
Assertions.assertNotNull(row);
Assertions.assertTrue(StringUtils.isNotBlank(row.getAs("identifier")));
System.out.println("row = " + row.getAs("country"));
System.out.println("row = " + row.getAs("countrytitle"));
}
@Test