andreas.czerniak
/
BrBETA_dnet-hadoop
forked from D-Net/dnet-hadoop
1
1
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

implementation of the size threshold on authors list match

This commit is contained in:
miconis 2022-03-08 16:49:28 +01:00
parent 8f07f0c537
commit 699612dd17
1 changed files with 5 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dnet-pace-core/src/main/java/eu/dnetlib/pace/tree/AuthorsMatch.java
View File

@ -25,6 +25,7 @@ public class AuthorsMatch extends AbstractComparator {
private double NAME_THRESHOLD;
private double FULLNAME_THRESHOLD;
private String MODE; //full or surname
private int SIZE_THRESHOLD;
private int common;
public AuthorsMatch(Map<String, String> params){
@ -35,6 +36,7 @@ public class AuthorsMatch extends AbstractComparator {
SURNAME_THRESHOLD = Double.parseDouble(params.getOrDefault("surname_th", "0.95"));
NAME_THRESHOLD = Double.parseDouble(params.getOrDefault("name_th", "0.95"));
FULLNAME_THRESHOLD = Double.parseDouble(params.getOrDefault("fullname_th", "0.9"));
SIZE_THRESHOLD = Integer.parseInt(params.getOrDefault("size_th", "20"));
common = 0;
}
@ -48,6 +50,9 @@ public class AuthorsMatch extends AbstractComparator {
if (a.isEmpty() || b.isEmpty())
return -1;
if (((FieldList) a).size() > SIZE_THRESHOLD || ((FieldList) a).size() > SIZE_THRESHOLD)
return 1.0;
List<Person> aList = ((FieldList) a).stringList().stream().map(author -> new Person(author, false)).collect(Collectors.toList());
List<Person> bList = ((FieldList) b).stringList().stream().map(author -> new Person(author, false)).collect(Collectors.toList());