1
|
package eu.dnetlib.index.solr.feed;
|
2
|
|
3
|
import java.text.ParseException;
|
4
|
import java.text.SimpleDateFormat;
|
5
|
import java.util.Arrays;
|
6
|
import java.util.Date;
|
7
|
import java.util.List;
|
8
|
import javax.xml.stream.XMLStreamException;
|
9
|
|
10
|
import org.apache.solr.common.SolrInputDocument;
|
11
|
import org.dom4j.DocumentException;
|
12
|
|
13
|
/**
|
14
|
*
|
15
|
* @author claudio
|
16
|
*
|
17
|
*/
|
18
|
public abstract class InputDocumentFactory {
|
19
|
|
20
|
public static final String INDEX_FIELD_PREFIX = "__";
|
21
|
|
22
|
public static final String DS_VERSION = INDEX_FIELD_PREFIX + "dsversion";
|
23
|
|
24
|
public static final String DS_ID = INDEX_FIELD_PREFIX + "dsid";
|
25
|
|
26
|
public static final String RESULT = "result";
|
27
|
|
28
|
public static final String INDEX_RESULT = INDEX_FIELD_PREFIX + RESULT;
|
29
|
|
30
|
public static final String INDEX_RECORD_ID = INDEX_FIELD_PREFIX + "indexrecordidentifier";
|
31
|
|
32
|
private static final String outFormat = new String("yyyy-MM-dd'T'hh:mm:ss'Z'");
|
33
|
|
34
|
private final static List<String> dateFormats = Arrays.asList("yyyy-MM-dd'T'hh:mm:ss", "yyyy-MM-dd", "dd-MM-yyyy", "dd/MM/yyyy", "yyyy");
|
35
|
|
36
|
public abstract SolrInputDocument parseDocument(final String version,
|
37
|
final String inputDocument,
|
38
|
final String dsId,
|
39
|
final String resultName) throws XMLStreamException;
|
40
|
|
41
|
public abstract SolrInputDocument parseDocument(final String version,
|
42
|
final String inputDocument,
|
43
|
final String dsId,
|
44
|
final String resultName,
|
45
|
final ResultTransformer resultTransformer) throws XMLStreamException;
|
46
|
|
47
|
/**
|
48
|
* method return a solr-compatible string representation of a date
|
49
|
*
|
50
|
* @param date
|
51
|
* @return the parsed date
|
52
|
* @throws DocumentException
|
53
|
* @throws ParseException
|
54
|
*/
|
55
|
public static String getParsedDateField(final String date) {
|
56
|
return new SimpleDateFormat(outFormat).format(tryParse(date));
|
57
|
}
|
58
|
|
59
|
public static Date tryParse(final String date) {
|
60
|
for (String formatString : dateFormats) {
|
61
|
try {
|
62
|
return new SimpleDateFormat(formatString).parse(date);
|
63
|
} catch (ParseException e) {}
|
64
|
}
|
65
|
throw new IllegalStateException("unable to parse date: " + date);
|
66
|
}
|
67
|
|
68
|
public String parseDate(final String date) {
|
69
|
return getParsedDateField(date);
|
70
|
}
|
71
|
|
72
|
}
|