D-Net
/
dnet-hadoop
15
4
Fork 10
Code Issues 14 Pull Requests 10 Projects 2 Releases 1 Wiki Activity

add. harvestingrange, respect IETFdraft rateLimit

This commit is contained in:
Andreas Czerniak 2021-09-01 22:37:19 +02:00
parent ebf3f47a02
commit c60be35b0a
4 changed files with 27 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dhp-common/src/main/java/eu/dnetlib/dhp/common/Constants.java
View File

@ -49,4 +49,9 @@ public class Constants {
public static final String CONTENT_INVALIDRECORDS = "InvalidRecords";
public static final String CONTENT_TRANSFORMEDRECORDS = "transformedItems";
// IETF Draft and used by Repositories like ZENODO , not included in APACHE HTTP java packages
// see https://ietf-wg-httpapi.github.io/ratelimit-headers/draft-ietf-httpapi-ratelimit-headers.html
public static final String HTTPHEADER_IETF_DRAFT_RATELIMIT_LIMIT = "X-RateLimit-Limit";
public static final String HTTPHEADER_IETF_DRAFT_RATELIMIT_REMAINING = "X-RateLimit-Remaining";
public static final String HTTPHEADER_IETF_DRAFT_RATELIMIT_RESET = "X-RateLimit-Reset";
}

13
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/HttpConnector2.java
View File

@ -20,7 +20,7 @@ import eu.dnetlib.dhp.aggregation.common.AggregatorReport;
/**
* Migrated from https://svn.driver.research-infrastructures.eu/driver/dnet45/modules/dnet-modular-collector-service/trunk/src/main/java/eu/dnetlib/data/collector/plugins/HttpConnector.java
*
* @author jochen, michele, andrea, alessia, claudio
* @author jochen, michele, andrea, alessia, claudio, andreas
*/
public class HttpConnector2 {
@ -112,6 +112,17 @@ public class HttpConnector2 {
}
int retryAfter = obtainRetryAfter(urlConn.getHeaderFields());
String rateLimit = urlConn.getHeaderField(eu.dnetlib.dhp.common.Constants.HTTPHEADER_IETF_DRAFT_RATELIMIT_LIMIT);
String rateRemaining = urlConn.getHeaderField(eu.dnetlib.dhp.common.Constants.HTTPHEADER_IETF_DRAFT_RATELIMIT_REMAINING);
if((rateLimit != null) && (rateRemaining != null) && ( Integer.parseInt(rateRemaining) < 2)) {
if (retryAfter > 0) {
backoffAndSleep(retryAfter);
} else {
backoffAndSleep(1000);
}
}
if (is2xx(urlConn.getResponseCode())) {
input = urlConn.getInputStream();
responseType = urlConn.getContentType();

10
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/oai/OaiCollectorPlugin.java
View File

@ -62,12 +62,14 @@ public class OaiCollectorPlugin implements CollectorPlugin {
throw new CollectorException("Param 'mdFormat' is null or empty");
}
if (fromDate != null && !fromDate.matches("\\d{4}-\\d{2}-\\d{2}") && !fromDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z")) {
throw new CollectorException("Invalid date (YYYY-MM-DD): " + fromDate);
if (fromDate != null && !fromDate.matches("\\d{4}-\\d{2}-\\d{2}")
&& !fromDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z")) {
throw new CollectorException("Invalid date (YYYY-MM-DD or YYYY-MM-DDT00:00:00Z): " + fromDate);
}
if (untilDate != null && !untilDate.matches("\\d{4}-\\d{2}-\\d{2}") && !untilDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z")) {
throw new CollectorException("Invalid date (YYYY-MM-DD): " + untilDate);
if (untilDate != null && !untilDate.matches("\\d{4}-\\d{2}-\\d{2}")
&& !untilDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z")) {
throw new CollectorException("Invalid date (YYYY-MM-DD or YYYY-MM-DDT00:00:00Z): " + untilDate);
}
final Iterator<Iterator<String>> iters = sets

6
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/oai/OaiIterator.java
View File

@ -107,10 +107,12 @@ public class OaiIterator implements Iterator<String> {
if (set != null && !set.isEmpty()) {
url += "&set=" + URLEncoder.encode(set, "UTF-8");
}
if (fromDate != null && (fromDate.matches("\\d{4}-\\d{2}-\\d{2}") || fromDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z"))) {
if (fromDate != null && (fromDate.matches("\\d{4}-\\d{2}-\\d{2}")
|| fromDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z"))) {
url += "&from=" + URLEncoder.encode(fromDate, "UTF-8");
}
if (untilDate != null && (untilDate.matches("\\d{4}-\\d{2}-\\d{2}") || untilDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z"))) {
if (untilDate != null && (untilDate.matches("\\d{4}-\\d{2}-\\d{2}")
|| untilDate.matches("\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}Z"))) {
url += "&until=" + URLEncoder.encode(untilDate, "UTF-8");
}
log.info("Start harvesting using url: " + url);