1 |
41789
|
sandro.lab
|
package eu.dnetlib.data.mdstore.modular;
|
2 |
|
|
|
3 |
|
|
import java.io.StringReader;
|
4 |
|
|
import java.util.HashMap;
|
5 |
|
|
import java.util.Map;
|
6 |
|
|
import javax.xml.xpath.XPath;
|
7 |
|
|
import javax.xml.xpath.XPathFactory;
|
8 |
|
|
|
9 |
|
|
import org.apache.commons.logging.Log;
|
10 |
|
|
import org.apache.commons.logging.LogFactory;
|
11 |
|
|
import org.xml.sax.InputSource;
|
12 |
|
|
|
13 |
|
|
/**
|
14 |
|
|
* Terrible implementation of a record parser.
|
15 |
|
|
*
|
16 |
|
|
* @author marko
|
17 |
|
|
*/
|
18 |
|
|
public class SimpleRecordParser implements RecordParser {
|
19 |
|
|
|
20 |
|
|
static final Log log = LogFactory.getLog(SimpleRecordParser.class); // NOPMD by marko on 11/24/08 5:02 PM
|
21 |
|
|
|
22 |
|
|
@Override
|
23 |
|
|
public Map<String, String> parseRecord(String record) {
|
24 |
|
|
Map<String, String> props = new HashMap<String, String>();
|
25 |
|
|
|
26 |
|
|
try {
|
27 |
|
|
// DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder();
|
28 |
|
|
XPath xpath = XPathFactory.newInstance().newXPath();
|
29 |
|
|
|
30 |
|
|
// Document doc = builder.parse(new InputSource(new StringReader(record)));
|
31 |
|
|
InputSource doc = new InputSource(new StringReader(record));
|
32 |
|
|
|
33 |
|
|
props.put("id", xpath.evaluate("//*[local-name()='objIdentifier']", doc));
|
34 |
|
|
props.put("originalId", xpath.evaluate("//*[local-name()='efgEntity']/*/*[local-name()='identifier']", doc));
|
35 |
|
|
|
36 |
|
|
// String date = xpath.evaluate("//*[local-name()='dateOfCollection'][1]", doc);
|
37 |
|
|
// props.put("date", new Date(date).getTime());
|
38 |
|
|
|
39 |
|
|
} catch (Exception e) {
|
40 |
|
|
log.warn("got exception while parsing document", e);
|
41 |
|
|
log.warn("record is:");
|
42 |
|
|
log.warn(record);
|
43 |
|
|
log.warn("------------");
|
44 |
|
|
}
|
45 |
|
|
return props;
|
46 |
|
|
|
47 |
|
|
}
|
48 |
|
|
|
49 |
|
|
}
|