|
1
|
package eu.dnetlib.data.transform.xml;
|
|
2
|
|
|
3
|
import com.google.common.collect.Lists;
|
|
4
|
import com.google.protobuf.Descriptors;
|
|
5
|
import eu.dnetlib.data.proto.*;
|
|
6
|
import eu.dnetlib.data.proto.WdsPublicationProtos;
|
|
7
|
import org.apache.commons.lang3.StringUtils;
|
|
8
|
import org.apache.commons.lang3.exception.ExceptionUtils;
|
|
9
|
import org.w3c.dom.Node;
|
|
10
|
import org.w3c.dom.NodeList;
|
|
11
|
|
|
12
|
import java.util.List;
|
|
13
|
|
|
14
|
public class WDSPmfToHbaseXsltFunctions extends OafToHbaseXsltFunctions {
|
|
15
|
|
|
16
|
|
|
17
|
public static String pmfPublication(
|
|
18
|
final String resultId,
|
|
19
|
final String provenance,
|
|
20
|
final String trust,
|
|
21
|
final NodeList about,
|
|
22
|
final String originalId,
|
|
23
|
final String dateOfCollection,
|
|
24
|
final String dateOfTransformation,
|
|
25
|
final NodeList metadataNodes) {
|
|
26
|
try {
|
|
27
|
final ValueMap values = ValueMap.parseNodeList(metadataNodes);
|
|
28
|
final Descriptors.Descriptor mDesc = PublicationProtos.Publication.Metadata.getDescriptor();
|
|
29
|
final DNGFProtos.DNGFEntity.Builder entity = DNGFProtos.DNGFEntity.newBuilder();
|
|
30
|
|
|
31
|
|
|
32
|
final PublicationProtos.Publication.Metadata.Builder metadata = buildMetadata(values, mDesc);
|
|
33
|
addField(metadata, mDesc.findFieldByName("dateofacceptance"), values.get("date").listValues());
|
|
34
|
final PublicationProtos.Publication.Builder publication = buildPublication(metadata, values, mDesc, "", "");
|
|
35
|
parseProject(publication.getMetadataBuilder(), metadataNodes);
|
|
36
|
final List<FieldTypeProtos.StructuredProperty> pids = Lists.newArrayList();
|
|
37
|
pids.addAll(parsePids(metadataNodes));
|
|
38
|
entity.setType(TypeProtos.Type.publication).setId(resultId);
|
|
39
|
entity.setDateoftransformation(StringUtils.isBlank(dateOfTransformation) ? "" : dateOfTransformation);
|
|
40
|
entity.setDateofcollection(StringUtils.isBlank(dateOfCollection) ? "" : dateOfCollection);
|
|
41
|
entity.addAllPid(pids);
|
|
42
|
entity.setDateofcollection(dateOfCollection)
|
|
43
|
.setDateoftransformation(dateOfTransformation).setOaiprovenance(getOAIProvenance(about));
|
|
44
|
entity.setPublication(publication);
|
|
45
|
|
|
46
|
final DNGFProtos.DNGF oaf = getOaf(entity, getDataInfo(about, provenance, trust, false, false));
|
|
47
|
return base64(oaf.toByteArray());
|
|
48
|
} catch (final Throwable e) {
|
|
49
|
System.err.println(ExceptionUtils.getStackTrace(e));
|
|
50
|
throw new IllegalArgumentException(e);
|
|
51
|
}
|
|
52
|
}
|
|
53
|
|
|
54
|
private static void parseProject(final PublicationProtos.Publication.Metadata.Builder metadata, final NodeList metadataNodes ) {
|
|
55
|
for (int i = 0; i < metadataNodes.getLength(); i++) {
|
|
56
|
final Node node = metadataNodes.item(i);
|
|
57
|
String acronym = null;
|
|
58
|
String funder = null;
|
|
59
|
String fundingStream = null;
|
|
60
|
String grantId = null;
|
|
61
|
String name = null;
|
|
62
|
if (node.getNodeType() == Node.ELEMENT_NODE) {
|
|
63
|
if (node.getLocalName().equalsIgnoreCase("project")) {
|
|
64
|
acronym = node.getAttributes().getNamedItem("acronym").getTextContent();
|
|
65
|
funder = node.getAttributes().getNamedItem("funder").getTextContent();
|
|
66
|
fundingStream = node.getAttributes().getNamedItem("fundingStream").getTextContent();
|
|
67
|
grantId = node.getAttributes().getNamedItem("grantId").getTextContent();
|
|
68
|
name = node.getTextContent();
|
|
69
|
metadata.addExtension(WdsPublicationProtos.WdsPublication.projects,
|
|
70
|
FieldTypeProtos.ProjectRelation.newBuilder()
|
|
71
|
.setAcronym(acronym)
|
|
72
|
.setFunder(funder)
|
|
73
|
.setFundingStream(fundingStream)
|
|
74
|
.setGrantId(grantId)
|
|
75
|
.setName(name)
|
|
76
|
.build());
|
|
77
|
|
|
78
|
|
|
79
|
|
|
80
|
|
|
81
|
}
|
|
82
|
|
|
83
|
}
|
|
84
|
}
|
|
85
|
|
|
86
|
}
|
|
87
|
|
|
88
|
|
|
89
|
protected static List<FieldTypeProtos.StructuredProperty> parsePids(final NodeList nodelist) {
|
|
90
|
|
|
91
|
final List<FieldTypeProtos.StructuredProperty> pids = Lists.newArrayList();
|
|
92
|
|
|
93
|
for (int i = 0; i < nodelist.getLength(); i++) {
|
|
94
|
final Node node = nodelist.item(i);
|
|
95
|
String pidType = null;
|
|
96
|
String pidValue = null;
|
|
97
|
if (node.getNodeType() == Node.ELEMENT_NODE) {
|
|
98
|
if (node.getLocalName().equalsIgnoreCase("pid")) {
|
|
99
|
pidType = node.getAttributes().getNamedItem("type").getTextContent();
|
|
100
|
pidValue = node.getTextContent();
|
|
101
|
if (StringUtils.isNotEmpty(pidType) && StringUtils.isNotEmpty(pidValue))
|
|
102
|
{
|
|
103
|
pids.add(getStructuredProperty(pidValue, pidType, pidType, "dnet:pid_types", "dnet:pid_types"));
|
|
104
|
}
|
|
105
|
}
|
|
106
|
//this is to handle dataset pids
|
|
107
|
}
|
|
108
|
}
|
|
109
|
return pids;
|
|
110
|
}
|
|
111
|
}
|