1
|
package eu.dnetlib.data.utils;
|
2
|
|
3
|
import java.io.IOException;
|
4
|
import java.io.StringReader;
|
5
|
import java.io.StringWriter;
|
6
|
import java.io.Writer;
|
7
|
import java.util.List;
|
8
|
import java.util.stream.Collectors;
|
9
|
|
10
|
import org.apache.commons.lang3.StringUtils;
|
11
|
import org.apache.commons.logging.Log;
|
12
|
import org.apache.commons.logging.LogFactory;
|
13
|
import org.dom4j.Document;
|
14
|
import org.dom4j.DocumentException;
|
15
|
import org.dom4j.Node;
|
16
|
import org.dom4j.io.SAXReader;
|
17
|
import org.jbibtex.BibTeXDatabase;
|
18
|
import org.jbibtex.BibTeXEntry;
|
19
|
import org.jbibtex.BibTeXFormatter;
|
20
|
import org.jbibtex.Key;
|
21
|
import org.jbibtex.StringValue;
|
22
|
import org.jbibtex.StringValue.Style;
|
23
|
|
24
|
public class BibTexConverter {
|
25
|
|
26
|
private static final Log log = LogFactory.getLog(BibTexConverter.class);
|
27
|
|
28
|
public static String asBibTex(final String xml) {
|
29
|
try {
|
30
|
final Document doc = new SAXReader().read(new StringReader(xml));
|
31
|
|
32
|
final BibTeXEntry dbEntry = new BibTeXEntry(
|
33
|
findBibTexType(doc.valueOf("//*[local-name() = 'resourceType']")),
|
34
|
new Key(doc.valueOf("//*[local-name() = 'identifier']")));
|
35
|
|
36
|
addSimpleStringValue(dbEntry, BibTeXEntry.KEY_TITLE, doc, "//*[local-name() = 'title']");
|
37
|
addMultipleStringValue(dbEntry, BibTeXEntry.KEY_AUTHOR, doc, "//*[local-name() = 'creator']");
|
38
|
addSimpleStringValue(dbEntry, BibTeXEntry.KEY_YEAR, doc, "//*[local-name() = 'date']");
|
39
|
|
40
|
final BibTeXFormatter formatter = new BibTeXFormatter();
|
41
|
final Writer res = new StringWriter();
|
42
|
final BibTeXDatabase db = new BibTeXDatabase();
|
43
|
db.addObject(dbEntry);
|
44
|
formatter.format(db, res);
|
45
|
return res.toString();
|
46
|
} catch (final DocumentException | IOException e) {
|
47
|
log.warn("Error generating bibtex", e);
|
48
|
return "";
|
49
|
}
|
50
|
}
|
51
|
|
52
|
private static void addSimpleStringValue(final BibTeXEntry dbEntry, final Key key, final Document doc, final String xpath) {
|
53
|
final String val = doc.valueOf(xpath).trim();
|
54
|
if (StringUtils.isNotBlank(val)) {
|
55
|
dbEntry.addField(key, new StringValue(val, Style.BRACED));
|
56
|
}
|
57
|
}
|
58
|
|
59
|
@SuppressWarnings("unchecked")
|
60
|
private static void addMultipleStringValue(final BibTeXEntry dbEntry, final Key key, final Document doc, final String xpath) {
|
61
|
final String val = ((List<Node>) doc.selectNodes(xpath))
|
62
|
.stream()
|
63
|
.map(Node::getText)
|
64
|
.map(String::trim)
|
65
|
.filter(StringUtils::isNotEmpty)
|
66
|
.collect(Collectors.joining(" and "));
|
67
|
|
68
|
if (StringUtils.isNotBlank(val)) {
|
69
|
dbEntry.addField(key, new StringValue(val, Style.BRACED));
|
70
|
}
|
71
|
}
|
72
|
|
73
|
private static Key findBibTexType(final String resourceType) {
|
74
|
switch (resourceType) {
|
75
|
case "Conference object":
|
76
|
return BibTeXEntry.TYPE_INPROCEEDINGS;
|
77
|
case "Report":
|
78
|
return BibTeXEntry.TYPE_TECHREPORT;
|
79
|
case "Article":
|
80
|
return BibTeXEntry.TYPE_ARTICLE;
|
81
|
case "Other":
|
82
|
return BibTeXEntry.TYPE_MISC;
|
83
|
case "Part of book or chapter of book":
|
84
|
return BibTeXEntry.TYPE_INBOOK;
|
85
|
case "Book":
|
86
|
return BibTeXEntry.TYPE_BOOK;
|
87
|
case "Bachelor thesis":
|
88
|
return BibTeXEntry.TYPE_MASTERSTHESIS;
|
89
|
case "Doctoral thesis":
|
90
|
return BibTeXEntry.TYPE_PHDTHESIS;
|
91
|
case "Patent":
|
92
|
return BibTeXEntry.TYPE_MISC;
|
93
|
case "Master thesis":
|
94
|
return BibTeXEntry.TYPE_MASTERSTHESIS;
|
95
|
default:
|
96
|
return BibTeXEntry.TYPE_MISC;
|
97
|
}
|
98
|
}
|
99
|
}
|