1
|
package eu.dnetlib.xml;
|
2
|
|
3
|
import java.util.Arrays;
|
4
|
import java.util.stream.Collectors;
|
5
|
|
6
|
import org.apache.commons.lang3.StringUtils;
|
7
|
import org.apache.commons.logging.Log;
|
8
|
import org.apache.commons.logging.LogFactory;
|
9
|
|
10
|
import eu.dnetlib.data.mdstore.plugins.objects.Project;
|
11
|
|
12
|
/**
|
13
|
* Created by sandro on 12/6/16.
|
14
|
*/
|
15
|
public class ISTIUtilityFunction {
|
16
|
|
17
|
private static final Log log = LogFactory.getLog(ISTIUtilityFunction.class);
|
18
|
|
19
|
public static String cleanNames(final String s) {
|
20
|
if (StringUtils.isBlank(s)) { return ""; }
|
21
|
return cleanNames(s, StringUtils.countMatches(s, ",") > 1 ? "," : ";");
|
22
|
}
|
23
|
|
24
|
private static String cleanNames(final String s, final String sep) {
|
25
|
return Arrays.stream(s.split(sep))
|
26
|
.map(String::trim)
|
27
|
.map(ISTIUtilityFunction::clean)
|
28
|
.map(ISTIUtilityFunction::capitalize)
|
29
|
.filter(StringUtils::isNotBlank)
|
30
|
.collect(Collectors.joining("#"));
|
31
|
}
|
32
|
|
33
|
private static String clean(final String s) {
|
34
|
return s.replaceAll("\\(.*\\)", "")
|
35
|
.replaceAll("\\[.*\\]", "")
|
36
|
.replaceAll("(?i)^et\\.? al(\\.|\\s)*$", "")
|
37
|
.replaceAll("(?i)\\s*et\\.? al(\\.|\\s)*$", "")
|
38
|
.replaceAll("\\d|\\*", "")
|
39
|
.replaceAll("^(\\s|\\-|\\.)+", "")
|
40
|
.replaceAll("\\,", "")
|
41
|
.replaceAll("\\.", ". ")
|
42
|
.trim();
|
43
|
}
|
44
|
|
45
|
public static String cleanDoi(final String doi) {
|
46
|
final String x = doi.replaceAll("\\?", "")
|
47
|
.replaceAll("\\s*\\/\\s*", "/")
|
48
|
.replaceAll("\\s*\\-\\s*", "-")
|
49
|
.trim()
|
50
|
.replaceAll("\\s", "_");
|
51
|
|
52
|
if (!x.equals(doi)) {
|
53
|
log.info("Cleaning doi: " + doi + " -> " + x);
|
54
|
}
|
55
|
|
56
|
return x;
|
57
|
}
|
58
|
|
59
|
public static String capitalize(final String s) {
|
60
|
return Arrays.stream(s.split(" "))
|
61
|
.map(String::toLowerCase)
|
62
|
.map(StringUtils::capitalize)
|
63
|
.collect(Collectors.joining(" "));
|
64
|
}
|
65
|
|
66
|
public static boolean isValidProject(final String id) {
|
67
|
return Project.isValid(id);
|
68
|
}
|
69
|
|
70
|
// <xsl:for-each select="tokenize(istiFunction:cleanName(.), ';')">
|
71
|
// <xsl:choose>
|
72
|
// <xsl:when test="matches(normalize-space(.), ',(\s*[a-zA-Z]\.)+$')">
|
73
|
// <creator>
|
74
|
// <creatorName>
|
75
|
// <xsl:value-of select="normalize-space(translate(.,',',' '))"/>
|
76
|
// </creatorName>
|
77
|
// </creator>
|
78
|
// </xsl:when>
|
79
|
//
|
80
|
// <xsl:when test="matches(normalize-space(.), '(^[a-zA-Z\.]+,\s?[a-zA-Z\.\s]+$)|(^[a-zA-Z\.\s]+,\s?[a-zA-Z\.]+$)')">
|
81
|
// <creator>
|
82
|
// <creatorName>
|
83
|
// <xsl:value-of select="normalize-space(translate(.,',',' '))"/>
|
84
|
// </creatorName>
|
85
|
// </creator>
|
86
|
// </xsl:when>
|
87
|
//
|
88
|
// <xsl:otherwise>
|
89
|
// <xsl:for-each select="tokenize(., ',')">
|
90
|
// <creator>
|
91
|
// <creatorName>
|
92
|
// <xsl:value-of select="normalize-space(.)"/>
|
93
|
// </creatorName>
|
94
|
// </creator>
|
95
|
// </xsl:for-each>
|
96
|
// </xsl:otherwise>
|
97
|
// </xsl:choose>
|
98
|
// </xsl:for-each>
|
99
|
//
|
100
|
|
101
|
}
|