@Override
public DataBag exec(Tuple input) throws IOException {
if (input == null || input.size() == 0) {
return null;
}
try {
DataByteArray dba = null;
try {
dba = (DataByteArray) input.get(0);
} catch (ExecException e) {
logger.error("Error in reading field:", e);
throw e;
}
DocumentWrapper dm = null;
try {
dm = DocumentWrapper.parseFrom(dba.get());
} catch (Exception e) {
logger.error("Error in reading ByteArray to DocumentMetadata:", e);
throw e;
}
DataBag ret = new DefaultDataBag();
DataByteArray metadata = new DataByteArray(dm.getDocumentMetadata().toByteArray());
List<Author> authors = dm.getDocumentMetadata().getBasicMetadata().getAuthorList();
for (int i = 0; i < authors.size(); i++) {
String sname = authors.get(i).getSurname();
Object[] to = new Object[] {sname, metadata, i};
Tuple t = TupleFactory.getInstance().newTuple(Arrays.asList(to));
ret.add(t);
}
return ret;
} catch (Exception e) {
logger.error("Error in processing input row:", e);
throw new IOException(
"Caught exception processing input row:\n" + StackTraceExtractor.getStackTrace(e));
}
}
@Override
public void map(
Writable key,
BytesWritable value,
Mapper<Writable, BytesWritable, Text, BytesWritable>.Context context)
throws IOException, InterruptedException {
DocumentWrapper docWrapper = DocumentProtos.DocumentWrapper.parseFrom(value.copyBytes());
String docKey = keyGen.generateKey(docWrapper.getDocumentMetadata(), 0);
if (!docKey.isEmpty()) {
DocumentWrapper thinDocWrapper = DocumentWrapperUtils.cloneDocumentMetadata(docWrapper);
context.write(new Text(docKey), new BytesWritable(thinDocWrapper.toByteArray()));
}
}
@Override
protected void map(Writable key, BytesWritable value, Context context)
throws IOException, InterruptedException {
int percentOfWritten = context.getConfiguration().getInt("percentOfWritten", 100);
DocumentWrapper docWrapper = DocumentProtos.DocumentWrapper.parseFrom(value.copyBytes());
log.info(
"work title = "
+ docWrapper.getDocumentMetadata().getBasicMetadata().getTitle(0).getText());
if ((i % 101) > 100 - percentOfWritten) {
log.info("writing...");
context.write(new Text(docWrapper.getRowId()), new BytesWritable(value.copyBytes()));
}
i++;
}