1
|
package eu.dnetlib.data.mapreduce.hbase.dataexport;
|
2
|
|
3
|
import eu.dnetlib.miscutils.functional.xml.DnetXsltFunctions;
|
4
|
import org.apache.commons.lang3.StringUtils;
|
5
|
import org.dom4j.Document;
|
6
|
import org.dom4j.tree.DefaultText;
|
7
|
|
8
|
import java.text.ParseException;
|
9
|
import java.util.Map;
|
10
|
import java.util.Objects;
|
11
|
|
12
|
public class RecordFilter {
|
13
|
|
14
|
private Map<String, String> criteria;
|
15
|
|
16
|
private String yearXpath;
|
17
|
private int fromYear;
|
18
|
private int toYear;
|
19
|
|
20
|
public RecordFilter(Map<String, String> criteria, String yearXpath, int fromYear, int toYear) {
|
21
|
this.criteria = criteria;
|
22
|
this.yearXpath = yearXpath;
|
23
|
this.fromYear = fromYear;
|
24
|
this.toYear = toYear;
|
25
|
}
|
26
|
|
27
|
public boolean matches(final Document record, final boolean strict) throws ParseException {
|
28
|
|
29
|
final String date = record.valueOf(yearXpath);
|
30
|
if (StringUtils.isBlank(date)) {
|
31
|
return false;
|
32
|
}
|
33
|
|
34
|
final String yyyy = DnetXsltFunctions.extractYear(date);
|
35
|
if (StringUtils.isBlank(yyyy)) {
|
36
|
return false;
|
37
|
}
|
38
|
final Integer year = Integer.valueOf(yyyy);
|
39
|
|
40
|
if (year < fromYear | year > toYear) {
|
41
|
return false;
|
42
|
}
|
43
|
|
44
|
|
45
|
boolean matched = false;
|
46
|
for(final Map.Entry<String, String> c : criteria.entrySet()) {
|
47
|
|
48
|
boolean matches = matched = record.selectNodes(c.getKey()).stream()
|
49
|
.filter(Objects::nonNull)
|
50
|
.map(o -> textOf(o))
|
51
|
.map(s -> ((String) s).toLowerCase())
|
52
|
.map(s -> ((String) s).trim())
|
53
|
.anyMatch(s -> {
|
54
|
return ((String) s).matches(c.getValue().toLowerCase());
|
55
|
});
|
56
|
|
57
|
if (matches && !strict) {
|
58
|
return true;
|
59
|
}
|
60
|
}
|
61
|
return matched;
|
62
|
}
|
63
|
|
64
|
private String textOf(final Object o) {
|
65
|
if (o instanceof org.dom4j.tree.DefaultText) {
|
66
|
return ((DefaultText) o).getText();
|
67
|
}
|
68
|
return o.toString();
|
69
|
}
|
70
|
|
71
|
}
|