1
|
package eu.dnetlib.data.mapreduce.hbase.dataexport;
|
2
|
|
3
|
import com.fasterxml.jackson.annotation.JsonInclude;
|
4
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
5
|
import com.google.gson.Gson;
|
6
|
import com.google.protobuf.InvalidProtocolBufferException;
|
7
|
import eu.dnetlib.data.mapreduce.util.DedupUtils;
|
8
|
import eu.dnetlib.data.mapreduce.util.OafRowKeyDecoder;
|
9
|
import eu.dnetlib.data.mapreduce.util.UpdateMerger;
|
10
|
import eu.dnetlib.data.proto.OafProtos;
|
11
|
import eu.dnetlib.data.proto.TypeProtos;
|
12
|
import eu.dnetlib.dhp.schema.oaf.Oaf;
|
13
|
import eu.dnetlib.dhp.schema.oaf.Relation;
|
14
|
import org.apache.commons.lang.StringUtils;
|
15
|
import org.apache.hadoop.hbase.client.Result;
|
16
|
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
|
17
|
import org.apache.hadoop.hbase.mapreduce.TableMapper;
|
18
|
import org.apache.hadoop.hbase.util.Bytes;
|
19
|
import org.apache.hadoop.io.Text;
|
20
|
import org.apache.hadoop.mapreduce.lib.output.MultipleOutputs;
|
21
|
|
22
|
import java.io.IOException;
|
23
|
import java.util.Map;
|
24
|
import java.util.Map.Entry;
|
25
|
import java.util.NavigableMap;
|
26
|
import java.util.stream.Collectors;
|
27
|
|
28
|
/**
|
29
|
* Exports Oaf objects as their json serialization.
|
30
|
*
|
31
|
* @author claudio
|
32
|
*
|
33
|
*/
|
34
|
public class ExportInformationSpaceMapper2DHP extends TableMapper<Text, Text> {
|
35
|
|
36
|
private Text keyOut;
|
37
|
|
38
|
private Text valueOut;
|
39
|
|
40
|
private MultipleOutputs multipleOutputs;
|
41
|
|
42
|
private ObjectMapper objectMapper;
|
43
|
|
44
|
@Override
|
45
|
protected void setup(final Context context) throws IOException, InterruptedException {
|
46
|
super.setup(context);
|
47
|
|
48
|
keyOut = new Text();
|
49
|
valueOut = new Text();
|
50
|
multipleOutputs = new MultipleOutputs(context);
|
51
|
objectMapper = new ObjectMapper()
|
52
|
.setSerializationInclusion(JsonInclude.Include.NON_NULL);
|
53
|
}
|
54
|
|
55
|
@Override
|
56
|
protected void cleanup(Context context) throws IOException, InterruptedException {
|
57
|
multipleOutputs.close();
|
58
|
}
|
59
|
|
60
|
@Override
|
61
|
protected void map(final ImmutableBytesWritable keyIn, final Result value, final Context context) throws IOException, InterruptedException {
|
62
|
try {
|
63
|
final OafRowKeyDecoder rkd = OafRowKeyDecoder.decode(keyIn.get());
|
64
|
|
65
|
final OafProtos.Oaf oaf = UpdateMerger.mergeBodyUpdates(context, value.getFamilyMap(Bytes.toBytes(rkd.getType().toString())));
|
66
|
|
67
|
if (oaf == null) {
|
68
|
return;
|
69
|
} else {
|
70
|
emit(context, oaf);
|
71
|
}
|
72
|
|
73
|
final Map<byte[], NavigableMap<byte[], byte[]>> row = value.getNoVersionMap();
|
74
|
|
75
|
for (byte[] cf : row.keySet()) {
|
76
|
|
77
|
for (Entry<byte[], byte[]> q : value.getFamilyMap(cf).entrySet().stream().filter(e -> {
|
78
|
final String key = new String(e.getKey());
|
79
|
boolean skip = key.startsWith("update") || key.equals(DedupUtils.BODY_S);
|
80
|
if (skip) {
|
81
|
context.getCounter("export", String.format("skipped %s", StringUtils.substring(key, 0, 6))).increment(1);
|
82
|
}
|
83
|
return !skip;
|
84
|
}).collect(Collectors.toList())) {
|
85
|
if (new String(q.getValue()).equals("")) {
|
86
|
context.getCounter("export", "skipped " + new String(cf)).increment(1);
|
87
|
} else {
|
88
|
emit(context, OafProtos.Oaf.parseFrom(q.getValue()));
|
89
|
}
|
90
|
}
|
91
|
}
|
92
|
} catch (final Throwable e) {
|
93
|
context.getCounter("export", "error: " + e.getClass().getName()).increment(1);
|
94
|
throw new RuntimeException(e);
|
95
|
}
|
96
|
}
|
97
|
|
98
|
private void emit(Context context, OafProtos.Oaf oaf) throws IOException, InterruptedException {
|
99
|
Oaf result = null;
|
100
|
try {
|
101
|
result = ProtoConverter.convert(oaf);
|
102
|
} catch (Throwable e) {
|
103
|
context.getCounter("export", "error:" + e.getClass().getName()).increment(1);
|
104
|
}
|
105
|
if (result != null) {
|
106
|
keyOut.set(result.getClass().getName());
|
107
|
valueOut.set(objectMapper.writeValueAsString(result));
|
108
|
|
109
|
final String namedOutput = result.getClass().getSimpleName().toLowerCase();
|
110
|
multipleOutputs.write(namedOutput, keyOut, valueOut, namedOutput + "/" + namedOutput);
|
111
|
|
112
|
context.getCounter("export", result.getClass().getName()).increment(1);
|
113
|
}
|
114
|
}
|
115
|
|
116
|
}
|