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

fixed a xpath

This commit is contained in:
Michele Artini 2024-03-08 08:23:07 +01:00
parent 6648d710a3
commit 3b5163d8e3
1 changed files with 17 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/collection/plugin/base/BaseCollectorPlugin.java
View File

@ -59,9 +59,7 @@ public class BaseCollectorPlugin implements CollectorPlugin {
log.info("dbPassword: {}", "***");
try {
if (!this.fs.exists(filePath)) {
throw new CollectorException("path does not exist: " + filePath);
}
if (!this.fs.exists(filePath)) { throw new CollectorException("path does not exist: " + filePath); }
} catch (final Throwable e) {
throw new CollectorException(e);
}
@ -82,28 +80,33 @@ public class BaseCollectorPlugin implements CollectorPlugin {
try (final DbClient dbClient = new DbClient(dbUrl, dbUser, dbPassword)) {
final String sql = IOUtils
.toString(
BaseAnalyzerJob.class
.toString(BaseAnalyzerJob.class
.getResourceAsStream("/eu/dnetlib/dhp/collection/plugin/base/sql/opendoar-accepted.sql"));
dbClient.processResults(sql, row -> {
try {
accepted.add(row.getString("id"));
final String dsId = row.getString("id");
log.info("Accepted Datasource: " + dsId);
accepted.add(dsId);
} catch (final SQLException e) {
log.error("Error in SQL", e);
throw new RuntimeException("Error in SQL", e);
}
});
} catch (final IOException e) {
log.error("Error accessong SQL", e);
throw new CollectorException("Error accessong SQL", e);
}
log.info("Accepted Datasources (TOTAL): " + accepted.size());
return accepted;
}
private boolean filterXml(final String xml, final Set<String> acceptedOpendoarIds, final AggregatorReport report) {
try {
final String id = DocumentHelper.parseText(xml).valueOf("//*[local-name='collection']/@opendoar_id").trim();
final String id = DocumentHelper.parseText(xml).valueOf("//*[local-name()='collection']/@opendoar_id").trim();
return (StringUtils.isNotBlank(id) && acceptedOpendoarIds.contains("opendoar____::" + id.trim()));
} catch (final DocumentException e) {
log.error("Error parsing document", e);