1
|
package eu.dnetlib.efg.download.plugin;
|
2
|
|
3
|
import java.util.ArrayList;
|
4
|
import java.util.Arrays;
|
5
|
import java.util.List;
|
6
|
|
7
|
import com.google.gson.Gson;
|
8
|
import eu.dnetlib.enabling.tools.DnetStreamSupport;
|
9
|
import eu.dnetlib.rmi.data.AbstractDownloadPlugin;
|
10
|
import eu.dnetlib.rmi.data.DownloadItem;
|
11
|
import eu.dnetlib.rmi.data.DownloadPlugin;
|
12
|
import eu.dnetlib.rmi.data.DownloadPluginException;
|
13
|
|
14
|
public class EFGDownloadPlugin extends AbstractDownloadPlugin implements DownloadPlugin {
|
15
|
|
16
|
private final static List<String> suffix = Arrays.asList("jpg", "jpeg", "gif", "png", "pdf", "ps", "djvu", "scanpix");
|
17
|
|
18
|
@Override
|
19
|
public String extractURL(final String baseURL) throws DownloadPluginException {
|
20
|
if (baseURL == null)
|
21
|
return null;
|
22
|
for (String s : suffix) {
|
23
|
if (baseURL.trim().endsWith(s))
|
24
|
return baseURL;
|
25
|
}
|
26
|
return null;
|
27
|
}
|
28
|
|
29
|
@Override
|
30
|
public void setBasePath(final String basePath) {
|
31
|
|
32
|
}
|
33
|
|
34
|
@Override
|
35
|
public DownloadItem retrieveUrl(final DownloadItem input) throws DownloadPluginException {
|
36
|
try {
|
37
|
String url = input.getOriginalUrl();
|
38
|
if (url == null || url.trim().length() == 0) {
|
39
|
return input;
|
40
|
}
|
41
|
List<String> urls = new Gson().fromJson(url, ArrayList.class);
|
42
|
if (urls == null || urls.size() == 0) {
|
43
|
input.setUrl(null);
|
44
|
input.setOriginalUrl(null);
|
45
|
return input;
|
46
|
}
|
47
|
if (checkUrlsNotNull(input, urls))
|
48
|
return input;
|
49
|
input.setOriginalUrl(null);
|
50
|
input.setUrl(null);
|
51
|
return input;
|
52
|
} catch (Throwable e) {
|
53
|
throw new DownloadPluginException("error on retrieve correct Url");
|
54
|
}
|
55
|
}
|
56
|
|
57
|
@Override
|
58
|
public Iterable<DownloadItem> retrieveUrls(final Iterable<DownloadItem> items) throws DownloadPluginException {
|
59
|
return DnetStreamSupport.generateStreamFromIterator(items.iterator()).map(it -> retrieveUrl(it))::iterator;
|
60
|
}
|
61
|
|
62
|
@Override
|
63
|
public String getPluginName() {
|
64
|
return "EFGDownloadPlugin";
|
65
|
}
|
66
|
}
|