sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

comments

This commit is contained in:
Michele Artini 2024-03-12 14:53:34 +01:00
parent 841ca92246
commit d9b23a76c5
1 changed files with 8 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/base/BaseCollectorPlugin.java
View File

@ -45,15 +45,22 @@ public class BaseCollectorPlugin implements CollectorPlugin {
@Override
public Stream<String> collect(final ApiDescriptor api, final AggregatorReport report) throws CollectorException {
// get path to file
// the path of the dump file on HDFS
// http://oai.base-search.net/initial_load/base_oaipmh_dump-current.tar
// it could be downloaded from iis-cdh5-test-gw.ocean.icm.edu.pl and then copied on HDFS
final Path filePath = Optional
.ofNullable(api.getBaseUrl())
.map(Path::new)
.orElseThrow(() -> new CollectorException("missing baseUrl"));
// get the parameters for the connection to the OpenAIRE database.
// the database is used to obtain the list of the datasources that the plugin will collect
final String dbUrl = api.getParams().get("dbUrl");
final String dbUser = api.getParams().get("dbUser");
final String dbPassword = api.getParams().get("dbPassword");
// the types(comma separated, empty value for all) that the plugin will collect,
// the types should be expressed in the format of the normalized types of BASE (for example 1,121,...)
final String acceptedNormTypesString = api.getParams().get("acceptedNormTypes");
log.info("baseUrl: {}", filePath);