1
|
package eu.dnetlib.data.mapreduce.hbase.dataexport;
|
2
|
|
3
|
import com.google.gson.Gson;
|
4
|
import eu.dnetlib.data.mapreduce.hbase.bulktag.ProtoMap;
|
5
|
import org.dom4j.Document;
|
6
|
import org.dom4j.DocumentException;
|
7
|
import org.dom4j.io.SAXReader;
|
8
|
import org.junit.Assert;
|
9
|
import org.junit.Test;
|
10
|
|
11
|
import java.text.ParseException;
|
12
|
import java.util.Map;
|
13
|
|
14
|
public class RecordFilterTest {
|
15
|
|
16
|
private static final String REGEX = ".*\\\\s(interdisciplinar.*|transdisciplinary.*)\\\\s.*";
|
17
|
|
18
|
private static final String DEFAULT_CRITERIA = "{ " +
|
19
|
"\"/*[local-name() ='record']/*[local-name() ='result']/*[local-name() ='metadata']/*[local-name() ='entity']/*[local-name() ='result']/*[local-name() ='datainfo']/*[local-name() ='deletedbyinference']/text()\" : \"false\", " +
|
20
|
"\"local-name(//*[local-name()='entity']/*)\" : \"result\" " +
|
21
|
"}";
|
22
|
|
23
|
private static final String CRITERIA = "{ " +
|
24
|
"\"/*[local-name() ='record']/*[local-name() ='result']/*[local-name() ='metadata']/*[local-name() ='entity']/*[local-name() ='result']/*[local-name() ='title']/text()\" : \""+REGEX+"\", " +
|
25
|
"\"/*[local-name() ='record']/*[local-name() ='result']/*[local-name() ='metadata']/*[local-name() ='entity']/*[local-name() ='result']/*[local-name() ='description']/text()\" : \""+REGEX+"\", " +
|
26
|
"\"/*[local-name() ='record']/*[local-name() ='result']/*[local-name() ='metadata']/*[local-name() ='entity']/*[local-name() ='result']/*[local-name() ='subject']/text()\" : \""+REGEX+"\" " +
|
27
|
"}";
|
28
|
|
29
|
private static final String YEAR_XPATH = "/*[local-name() ='record']/*[local-name() ='result']/*[local-name() ='metadata']/*[local-name() ='entity']/*[local-name() ='result']/*[local-name() ='dateofacceptance']";
|
30
|
|
31
|
@Test
|
32
|
public void recordNonMatchFilterTest() throws DocumentException, ParseException {
|
33
|
|
34
|
final Map<String, String> defaultCriteria = new Gson().fromJson(DEFAULT_CRITERIA, ProtoMap.class);
|
35
|
|
36
|
final RecordFilter defaultFilter = new RecordFilter(defaultCriteria, YEAR_XPATH, 1990, 2019);
|
37
|
|
38
|
final Document doc = new SAXReader().read(getClass().getResourceAsStream("non_match_record.xml"));
|
39
|
Assert.assertTrue(defaultFilter.matches(doc, true));
|
40
|
|
41
|
final RecordFilter filter = new RecordFilter(new Gson().fromJson(CRITERIA, ProtoMap.class), YEAR_XPATH, 1990, 2019);
|
42
|
Assert.assertFalse(filter.matches(doc, false));
|
43
|
}
|
44
|
|
45
|
@Test
|
46
|
public void recordMatchFilterTest() throws DocumentException, ParseException {
|
47
|
|
48
|
final Map<String, String> defaultCriteria = new Gson().fromJson(DEFAULT_CRITERIA, ProtoMap.class);
|
49
|
|
50
|
final RecordFilter defaultFilter = new RecordFilter(defaultCriteria, YEAR_XPATH, 1990, 2019);
|
51
|
|
52
|
final Document doc = new SAXReader().read(getClass().getResourceAsStream("match_record.xml"));
|
53
|
Assert.assertTrue(defaultFilter.matches(doc, true));
|
54
|
|
55
|
final RecordFilter filter = new RecordFilter(new Gson().fromJson(CRITERIA, ProtoMap.class), YEAR_XPATH, 1990, 2019);
|
56
|
Assert.assertTrue(filter.matches(doc, false));
|
57
|
}
|
58
|
|
59
|
}
|