1
|
package eu.dnetlib.data.transform;
|
2
|
|
3
|
import java.io.StringReader;
|
4
|
import java.util.Arrays;
|
5
|
import java.util.List;
|
6
|
import java.util.stream.Collectors;
|
7
|
import javax.xml.transform.Transformer;
|
8
|
import javax.xml.transform.stream.StreamSource;
|
9
|
|
10
|
import com.google.common.base.Function;
|
11
|
import com.ximpleware.AutoPilot;
|
12
|
import com.ximpleware.VTDGen;
|
13
|
import com.ximpleware.VTDNav;
|
14
|
import eu.dnetlib.data.graph.model.DNGFDecoder;
|
15
|
import eu.dnetlib.data.transform.VtdUtilityParser.Node;
|
16
|
import org.apache.commons.codec.binary.Base64;
|
17
|
import org.apache.commons.lang3.StringUtils;
|
18
|
import org.apache.commons.logging.Log;
|
19
|
import org.apache.commons.logging.LogFactory;
|
20
|
import org.dom4j.io.DocumentResult;
|
21
|
|
22
|
public class XsltRowTransformer implements Function<String, List<Row>> {
|
23
|
|
24
|
private static final Log log = LogFactory.getLog(XsltRowTransformer.class); // NOPMD by marko on 11/24/08 5:02 PM
|
25
|
|
26
|
public static final String DATA_XPATH = "/ROWS/ROW";
|
27
|
|
28
|
public static final List<String> ATTRS = Arrays.asList("value", "mode");
|
29
|
|
30
|
private Transformer transformer;
|
31
|
|
32
|
public XsltRowTransformer(final Transformer transformer) {
|
33
|
this.transformer = transformer;
|
34
|
log.info(String.format("using trasformer: '%s'", getTransformerClassName()));
|
35
|
}
|
36
|
|
37
|
private List<Row> transform(final String xml) {
|
38
|
final DocumentResult result = new DocumentResult();
|
39
|
try {
|
40
|
transformer.transform(new StreamSource(new StringReader(xml)), result);
|
41
|
|
42
|
final VTDGen parser = new VTDGen();
|
43
|
parser.setDoc(result.getDocument().asXML().getBytes());
|
44
|
parser.parse(true);
|
45
|
|
46
|
final VTDNav nav = parser.getNav();
|
47
|
final AutoPilot autoPilot = new AutoPilot(nav);
|
48
|
return VtdUtilityParser.getTextValuesWithAttributes(autoPilot, nav, DATA_XPATH, ATTRS)
|
49
|
.stream()
|
50
|
.map(this::asRow)
|
51
|
.collect(Collectors.toList());
|
52
|
} catch (Exception e) {
|
53
|
log.error("Error parsing xml:\n" + xml, e);
|
54
|
throw new RuntimeException("Unable to parse document:\n" + xml, e);
|
55
|
}
|
56
|
}
|
57
|
|
58
|
private Row asRow(final Node base64Node) {
|
59
|
final String base64 = base64Node.getAttributes().get("value");
|
60
|
if (StringUtils.isBlank(base64))
|
61
|
throw new IllegalArgumentException("empty protocolbuffer value, check xslt");
|
62
|
|
63
|
final DNGFDecoder d = DNGFDecoder.decode(Base64.decodeBase64(base64));
|
64
|
final String qualifier = getQualifier(base64Node, d);
|
65
|
return new Row(d.getColumnFamily(), d.getRowkey(), Arrays.asList(new Column<>(qualifier, d.getDNGF().toByteArray())));
|
66
|
}
|
67
|
|
68
|
private String getQualifier(final Node base64Node, final DNGFDecoder d) {
|
69
|
if("update".equals(base64Node.getAttributes().get("mode"))) {
|
70
|
return "update_" + System.nanoTime();
|
71
|
}
|
72
|
return d.getQualifier();
|
73
|
}
|
74
|
|
75
|
public String getTransformerClassName() {
|
76
|
return transformer != null ? transformer.getClass().getName() : "unknown";
|
77
|
}
|
78
|
|
79
|
@Override
|
80
|
public List<Row> apply(final String xml) {
|
81
|
return transform(xml);
|
82
|
}
|
83
|
|
84
|
}
|