|
1
|
package eu.dnetlib.msro.openaireplus.workflows.nodes.objectStore;
|
|
2
|
|
|
3
|
import java.io.File;
|
|
4
|
import java.util.List;
|
|
5
|
import javax.xml.ws.wsaddressing.W3CEndpointReference;
|
|
6
|
|
|
7
|
import com.googlecode.sarasvati.NodeToken;
|
|
8
|
import eu.dnetlib.data.objectstore.rmi.ObjectStoreService;
|
|
9
|
import eu.dnetlib.data.objectstore.rmi.Protocols;
|
|
10
|
import eu.dnetlib.enabling.resultset.IterableResultSetFactory;
|
|
11
|
import eu.dnetlib.enabling.tools.blackboard.BlackboardJob;
|
|
12
|
import eu.dnetlib.msro.workflows.nodes.BlackboardJobNode;
|
|
13
|
import eu.dnetlib.msro.workflows.nodes.ProgressJobNode;
|
|
14
|
import eu.dnetlib.msro.workflows.resultset.ProcessCountingResultSetFactory;
|
|
15
|
import eu.dnetlib.msro.workflows.util.ProgressProvider;
|
|
16
|
import eu.dnetlib.msro.workflows.util.ResultsetProgressProvider;
|
|
17
|
import org.apache.commons.io.FileUtils;
|
|
18
|
import org.apache.commons.io.filefilter.TrueFileFilter;
|
|
19
|
|
|
20
|
public class ImportFilesJobNode extends BlackboardJobNode implements ProgressJobNode {
|
|
21
|
|
|
22
|
private String basePath;
|
|
23
|
private String repositoryPrefix;
|
|
24
|
private String oaiPrefix;
|
|
25
|
private String objectStoreId;
|
|
26
|
private String contentDescription;
|
|
27
|
|
|
28
|
private ResultsetProgressProvider progressProvider;
|
|
29
|
private ProcessCountingResultSetFactory processCountingResultSetFactory;
|
|
30
|
private IterableResultSetFactory iterableResultSetFactory;
|
|
31
|
|
|
32
|
public String getRepositoryPrefix() {
|
|
33
|
return repositoryPrefix;
|
|
34
|
}
|
|
35
|
|
|
36
|
public void setRepositoryPrefix(final String repositoryPrefix) {
|
|
37
|
this.repositoryPrefix = repositoryPrefix;
|
|
38
|
}
|
|
39
|
|
|
40
|
public String getBasePath() {
|
|
41
|
return basePath;
|
|
42
|
}
|
|
43
|
|
|
44
|
public void setBasePath(final String basePath) {
|
|
45
|
this.basePath = basePath;
|
|
46
|
}
|
|
47
|
|
|
48
|
@Override
|
|
49
|
public ProgressProvider getProgressProvider() {
|
|
50
|
return progressProvider;
|
|
51
|
}
|
|
52
|
|
|
53
|
@Override
|
|
54
|
protected String obtainServiceId(final NodeToken token) {
|
|
55
|
return getServiceLocator().getServiceId(ObjectStoreService.class);
|
|
56
|
}
|
|
57
|
|
|
58
|
@Override
|
|
59
|
protected void prepareJob(final BlackboardJob job, final NodeToken token) throws Exception {
|
|
60
|
File dir = new File(basePath);
|
|
61
|
|
|
62
|
System.out.println("Getting all files in " + dir.getCanonicalPath() + " including those in subdirectories");
|
|
63
|
List<File> files = (List<File>) FileUtils.listFiles(dir, TrueFileFilter.INSTANCE, TrueFileFilter.INSTANCE);
|
|
64
|
IteratorRepository irepo;
|
|
65
|
if (repositoryPrefix.equals("od________18")) {
|
|
66
|
irepo = new ArxivIteratorRepository(files, repositoryPrefix, oaiPrefix);
|
|
67
|
} else if (repositoryPrefix.equals("od______2367")) {
|
|
68
|
irepo = new PUMAIteratorRepository(files, repositoryPrefix, oaiPrefix);
|
|
69
|
} else {
|
|
70
|
irepo = new RepositoryNoURLIterator(files, repositoryPrefix, oaiPrefix);
|
|
71
|
}
|
|
72
|
final W3CEndpointReference epr = iterableResultSetFactory.createIterableResultSet(irepo);
|
|
73
|
this.progressProvider = processCountingResultSetFactory.createProgressProvider(token.getProcess(), epr);
|
|
74
|
job.setAction("FEED");
|
|
75
|
job.getParameters().put("obsID", getObjectStoreId());
|
|
76
|
job.getParameters().put("mime", getContentDescription());
|
|
77
|
job.getParameters().put("protocol", Protocols.File_System.toString());
|
|
78
|
job.getParameters().put("epr", progressProvider.getEpr().toString());
|
|
79
|
|
|
80
|
}
|
|
81
|
|
|
82
|
public IterableResultSetFactory getIterableResultSetFactory() {
|
|
83
|
return iterableResultSetFactory;
|
|
84
|
}
|
|
85
|
|
|
86
|
public void setIterableResultSetFactory(final IterableResultSetFactory iterableResultSetFactory) {
|
|
87
|
this.iterableResultSetFactory = iterableResultSetFactory;
|
|
88
|
}
|
|
89
|
|
|
90
|
public void setProgressProvider(final ResultsetProgressProvider progressProvider) {
|
|
91
|
this.progressProvider = progressProvider;
|
|
92
|
}
|
|
93
|
|
|
94
|
public ProcessCountingResultSetFactory getProcessCountingResultSetFactory() {
|
|
95
|
return processCountingResultSetFactory;
|
|
96
|
}
|
|
97
|
|
|
98
|
public void setProcessCountingResultSetFactory(final ProcessCountingResultSetFactory processCountingResultSetFactory) {
|
|
99
|
this.processCountingResultSetFactory = processCountingResultSetFactory;
|
|
100
|
}
|
|
101
|
|
|
102
|
public String getObjectStoreId() {
|
|
103
|
return objectStoreId;
|
|
104
|
}
|
|
105
|
|
|
106
|
public void setObjectStoreId(final String objectStoreId) {
|
|
107
|
this.objectStoreId = objectStoreId;
|
|
108
|
}
|
|
109
|
|
|
110
|
public String getContentDescription() {
|
|
111
|
return contentDescription;
|
|
112
|
}
|
|
113
|
|
|
114
|
public void setContentDescription(final String contentDescription) {
|
|
115
|
this.contentDescription = contentDescription;
|
|
116
|
}
|
|
117
|
|
|
118
|
public String getOaiPrefix() {
|
|
119
|
return oaiPrefix;
|
|
120
|
}
|
|
121
|
|
|
122
|
public void setOaiPrefix(final String oaiPrefix) {
|
|
123
|
this.oaiPrefix = oaiPrefix;
|
|
124
|
}
|
|
125
|
}
|