|
1
|
package eu.dnetlib.resolver;
|
|
2
|
|
|
3
|
import java.util.ArrayList;
|
|
4
|
import java.util.Arrays;
|
|
5
|
import java.util.List;
|
|
6
|
|
|
7
|
import com.ximpleware.AutoPilot;
|
|
8
|
import com.ximpleware.VTDGen;
|
|
9
|
import com.ximpleware.VTDNav;
|
|
10
|
import eu.dnetlib.resolver.model.ResolvedObject;
|
|
11
|
import eu.dnetlib.resolver.model.SubjectType;
|
|
12
|
import eu.dnetlib.resolver.parser.AbstractResolverParser;
|
|
13
|
import eu.dnetlib.resolver.parser.UtilityParser;
|
|
14
|
import eu.dnetlib.resolver.parser.UtilityParser.Node;
|
|
15
|
import org.apache.commons.logging.Log;
|
|
16
|
import org.apache.commons.logging.LogFactory;
|
|
17
|
|
|
18
|
/**
|
|
19
|
* Created by sandro on 10/3/16.
|
|
20
|
*/
|
|
21
|
public class OpenAireParser extends AbstractResolverParser {
|
|
22
|
|
|
23
|
private static final Log log = LogFactory.getLog(OpenAireParser.class);
|
|
24
|
|
|
25
|
@Override
|
|
26
|
public ResolvedObject parseObject(final String record) {
|
|
27
|
try {
|
|
28
|
final ResolvedObject currentObject = new ResolvedObject();
|
|
29
|
final VTDGen vg = new VTDGen();
|
|
30
|
vg.setDoc(record.getBytes());
|
|
31
|
vg.parse(true);
|
|
32
|
|
|
33
|
final VTDNav vn = vg.getNav();
|
|
34
|
final AutoPilot ap = new AutoPilot(vn);
|
|
35
|
|
|
36
|
ap.declareXPathNameSpace("oaf", "http://namespace.openaire.eu/oaf");
|
|
37
|
|
|
38
|
final List<Node> subjectNodes =
|
|
39
|
UtilityParser.getTextValuesWithAttributes(ap, vn, "//oaf:result//subject", Arrays.asList("classname", "schemename"));
|
|
40
|
|
|
41
|
//Setting subjects
|
|
42
|
if (subjectNodes != null && subjectNodes.size() > 0) {
|
|
43
|
final List<SubjectType> currentSubjects = new ArrayList<>();
|
|
44
|
|
|
45
|
subjectNodes.forEach(it ->
|
|
46
|
currentSubjects.add(new SubjectType(it.getAttributes().get("schemename"), it.getTextValue()))
|
|
47
|
);
|
|
48
|
currentObject.setSubjects(currentSubjects);
|
|
49
|
}
|
|
50
|
|
|
51
|
//Setting Titles
|
|
52
|
|
|
53
|
final List<String> titles = UtilityParser.getTextValue(ap, vn, "//oaf:result/title");
|
|
54
|
|
|
55
|
if (titles != null && titles.size() > 0) {
|
|
56
|
|
|
57
|
currentObject.setTitles(titles);
|
|
58
|
}
|
|
59
|
|
|
60
|
//Setting authors
|
|
61
|
final List<String> authorNodes = UtilityParser.getTextValue(ap, vn, "//oaf:result//fullname");
|
|
62
|
|
|
63
|
if (authorNodes != null && authorNodes.size() > 0) {
|
|
64
|
currentObject.setAuthors(authorNodes);
|
|
65
|
}
|
|
66
|
|
|
67
|
//Setting descriptions
|
|
68
|
final List<String> descriptions = UtilityParser.getTextValue(ap, vn, "//oaf:result//description");
|
|
69
|
|
|
70
|
if (descriptions != null && descriptions.size() > 0) {
|
|
71
|
currentObject.setDescription(descriptions.get(0));
|
|
72
|
}
|
|
73
|
|
|
74
|
//resulttype classid
|
|
75
|
final List<Node> resutlTypes = UtilityParser.getTextValuesWithAttributes(ap, vn, "//oaf:result/resulttype", Arrays.asList("classid"));
|
|
76
|
|
|
77
|
if (resutlTypes != null && resutlTypes.size() > 0) {
|
|
78
|
final String type = resutlTypes.get(0).getAttributes().get("classid");
|
|
79
|
setType(currentObject, type);
|
|
80
|
}
|
|
81
|
|
|
82
|
return currentObject;
|
|
83
|
} catch (Throwable e) {
|
|
84
|
log.error("Error on parsing object ", e);
|
|
85
|
return null;
|
|
86
|
}
|
|
87
|
|
|
88
|
}
|
|
89
|
|
|
90
|
}
|