|
1
|
package eu.dnetlib.pace.clustering;
|
|
2
|
|
|
3
|
import java.util.Collection;
|
|
4
|
import java.util.List;
|
|
5
|
import java.util.Map;
|
|
6
|
|
|
7
|
import org.apache.commons.lang.RandomStringUtils;
|
|
8
|
import org.apache.commons.lang.StringUtils;
|
|
9
|
|
|
10
|
import com.google.common.collect.Lists;
|
|
11
|
|
|
12
|
public class SpaceTrimmingFieldValue extends AbstractClusteringFunction {
|
|
13
|
|
|
14
|
public SpaceTrimmingFieldValue(final Map<String, Integer> params) {
|
|
15
|
super(params);
|
|
16
|
}
|
|
17
|
|
|
18
|
@Override
|
|
19
|
protected Collection<String> doApply(final String s) {
|
|
20
|
final List<String> res = Lists.newArrayList();
|
|
21
|
|
|
22
|
res.add(StringUtils.isBlank(s) ? RandomStringUtils.random(getParams().get("randomLength")) : s.toLowerCase().replaceAll("\\s+", ""));
|
|
23
|
|
|
24
|
return res;
|
|
25
|
}
|
|
26
|
|
|
27
|
}
|