1
|
package eu.dnetlib.dli.resolver;
|
2
|
|
3
|
import com.google.common.collect.Lists;
|
4
|
import com.ximpleware.AutoPilot;
|
5
|
import com.ximpleware.VTDGen;
|
6
|
import com.ximpleware.VTDNav;
|
7
|
import eu.dnetlib.data.transform.VtdUtilityParser;
|
8
|
import eu.dnetlib.dli.resolver.model.*;
|
9
|
import eu.dnetlib.pid.resolver.model.ObjectType;
|
10
|
import eu.dnetlib.pid.resolver.model.SubjectType;
|
11
|
|
12
|
import java.util.Arrays;
|
13
|
import java.util.List;
|
14
|
|
15
|
/**
|
16
|
* Created by sandro on 3/6/17.
|
17
|
*/
|
18
|
public class ENAParser {
|
19
|
private static String ENA_NS_PREFIX = "r3d100010527";
|
20
|
|
21
|
public DLIResolvedObject parser(final String record, final String pid, final String pidType) throws Exception {
|
22
|
final VTDGen vg = new VTDGen();
|
23
|
vg.setDoc(record.getBytes());
|
24
|
vg.parse(true);
|
25
|
final VTDNav vn = vg.getNav();
|
26
|
final AutoPilot ap = new AutoPilot(vn);
|
27
|
|
28
|
final DLIResolvedObject result = new DLIResolvedObject();
|
29
|
List<VtdUtilityParser.Node> textValuesWithAttributes = VtdUtilityParser.getTextValuesWithAttributes(ap, vn, "//ROOT/entry", Arrays.asList("taxonomicDivision", "moleculeType", "firstPublic"));
|
30
|
|
31
|
|
32
|
if (textValuesWithAttributes != null && !textValuesWithAttributes.isEmpty()) {
|
33
|
result.setType(ObjectType.dataset);
|
34
|
final VtdUtilityParser.Node firstNode = textValuesWithAttributes.get(0);
|
35
|
result.setDate(firstNode.getAttributes().get("firstPublic"));
|
36
|
|
37
|
final String moleculeType = firstNode.getAttributes().get("moleculeType");
|
38
|
final String taxonomicDivision = firstNode.getAttributes().get("taxonomicDivision");
|
39
|
result.setSubjects(Arrays.asList(new SubjectType("moleculeType", moleculeType), new SubjectType("taxonomicDivision", taxonomicDivision)));
|
40
|
|
41
|
|
42
|
} else return null;
|
43
|
result.setTitles(Arrays.asList(VtdUtilityParser.getSingleValue(ap, vn, "//ROOT/entry/description")));
|
44
|
|
45
|
result.setPid(pid);
|
46
|
result.setPidType(pidType);
|
47
|
|
48
|
result.setCompletionStatus(CompletionStatus.complete.toString());
|
49
|
final DLIObjectProvenance provenance = new DLIObjectProvenance();
|
50
|
provenance.setCompletionStatus(CompletionStatus.complete.toString());
|
51
|
provenance.setProvisionMode(ObjectProvisionMode.resolved.toString());
|
52
|
DLIPIDResolver.setDatasourceProvenance(provenance, ENA_NS_PREFIX);
|
53
|
result.setDatasourceProvenance(Lists.newArrayList(provenance));
|
54
|
|
55
|
return result;
|
56
|
}
|
57
|
|
58
|
}
|