1
|
package eu.dnetlib.pace.clustering;
|
2
|
|
3
|
import java.util.Collection;
|
4
|
import java.util.List;
|
5
|
import java.util.Map;
|
6
|
import java.util.function.Predicate;
|
7
|
|
8
|
import com.google.common.collect.Sets;
|
9
|
|
10
|
import eu.dnetlib.pace.common.AbstractPaceFunctions;
|
11
|
import eu.dnetlib.pace.model.Field;
|
12
|
|
13
|
public abstract class AbstractClusteringFunction extends AbstractPaceFunctions implements ClusteringFunction {
|
14
|
|
15
|
protected Map<String, Integer> params;
|
16
|
|
17
|
public AbstractClusteringFunction(final Map<String, Integer> params) {
|
18
|
this.params = params;
|
19
|
}
|
20
|
|
21
|
protected abstract Collection<String> doApply(String s);
|
22
|
|
23
|
@Override
|
24
|
public Collection<String> apply(List<Field> fields) {
|
25
|
Collection<String> c = Sets.newLinkedHashSet();
|
26
|
fields.stream().filter(f -> !f.isEmpty()).forEach(f -> c.addAll(filterBlacklisted(doApply(filterStopWords(normalize(f.stringValue()), stopwords)), ngramBlacklist)));
|
27
|
return c;
|
28
|
}
|
29
|
|
30
|
public Map<String, Integer> getParams() {
|
31
|
return params;
|
32
|
}
|
33
|
|
34
|
protected Integer param(String name) {
|
35
|
return params.get(name);
|
36
|
}
|
37
|
}
|