|
1
|
package eu.dnetlib.rmi.data;
|
|
2
|
|
|
3
|
import java.time.LocalDate;
|
|
4
|
import java.time.ZoneId;
|
|
5
|
import java.time.temporal.ChronoUnit;
|
|
6
|
import java.util.List;
|
|
7
|
|
|
8
|
import com.google.gson.Gson;
|
|
9
|
import eu.dnetlib.rmi.data.DownloadItem.OpenAccessValues;
|
|
10
|
|
|
11
|
// TODO: Auto-generated Javadoc
|
|
12
|
|
|
13
|
/**
|
|
14
|
* The Class AbstractDownloadPlugin.
|
|
15
|
*/
|
|
16
|
public abstract class AbstractDownloadPlugin {
|
|
17
|
|
|
18
|
public static final int DEFAULT_TIMEOUT = 5000;
|
|
19
|
|
|
20
|
/**
|
|
21
|
* The regular expression.
|
|
22
|
*/
|
|
23
|
protected List<String> regularExpression;
|
|
24
|
|
|
25
|
/**
|
|
26
|
* Check open access.
|
|
27
|
*
|
|
28
|
* @param input the input
|
|
29
|
* @return the download item
|
|
30
|
*/
|
|
31
|
public DownloadItem checkOpenAccess(final DownloadItem input) {
|
|
32
|
if (input != null) {
|
|
33
|
OpenAccessValues openAccess = OpenAccessValues.valueOf(input.getOpenAccess());
|
|
34
|
switch (openAccess) {
|
|
35
|
case OPEN:
|
|
36
|
return input;
|
|
37
|
case CLOSED:
|
|
38
|
case RESTRICTED:
|
|
39
|
case UNKNOWN:
|
|
40
|
return null;
|
|
41
|
case EMBARGO:
|
|
42
|
if (input.getEmbargoDate() == null) return null;
|
|
43
|
LocalDate embargoDate = input.getEmbargoDate().toInstant().atZone(ZoneId.systemDefault()).toLocalDate();
|
|
44
|
LocalDate today = LocalDate.now();
|
|
45
|
final long between = ChronoUnit.DAYS.between(embargoDate, today);
|
|
46
|
if (between <= 0) return input;
|
|
47
|
return null;
|
|
48
|
}
|
|
49
|
}
|
|
50
|
return null;
|
|
51
|
}
|
|
52
|
|
|
53
|
public DownloadItem filterByRegexp(final DownloadItem input) {
|
|
54
|
if (this.regularExpression != null && this.regularExpression.size() > 0) {
|
|
55
|
final String baseURLs = input.getUrl();
|
|
56
|
final List<String> urlsList = new Gson().fromJson(baseURLs, List.class);
|
|
57
|
|
|
58
|
for (final String baseURL : urlsList) {
|
|
59
|
for (final String regExp : regularExpression) {
|
|
60
|
if (baseURL.matches(regExp))
|
|
61
|
return input;
|
|
62
|
}
|
|
63
|
}
|
|
64
|
return null;
|
|
65
|
} else return input;
|
|
66
|
}
|
|
67
|
|
|
68
|
protected boolean checkUrlsNotNull(DownloadItem input, List<String> urls) {
|
|
69
|
for (String s : urls) {
|
|
70
|
String newURL = extractURL(s);
|
|
71
|
if (newURL != null) {
|
|
72
|
input.setOriginalUrl(s);
|
|
73
|
input.setUrl(newURL);
|
|
74
|
return true;
|
|
75
|
} else {
|
|
76
|
input.setOriginalUrl(null);
|
|
77
|
input.setUrl(null);
|
|
78
|
}
|
|
79
|
}
|
|
80
|
return false;
|
|
81
|
}
|
|
82
|
|
|
83
|
public abstract String extractURL(final String baseURL) throws DownloadPluginException;
|
|
84
|
|
|
85
|
/**
|
|
86
|
* Gets the regular expression.
|
|
87
|
*
|
|
88
|
* @return the regular expression
|
|
89
|
*/
|
|
90
|
public List<String> getRegularExpression() {
|
|
91
|
return regularExpression;
|
|
92
|
}
|
|
93
|
|
|
94
|
/**
|
|
95
|
* Sets the regular expression.
|
|
96
|
*
|
|
97
|
* @param regularExpression the new regular expression
|
|
98
|
*/
|
|
99
|
public void setRegularExpression(final List<String> regularExpression) {
|
|
100
|
this.regularExpression = regularExpression;
|
|
101
|
}
|
|
102
|
|
|
103
|
}
|