private void writeContainerless(Record src, OutputStream dst) {
try {
GenericDatumWriter datumWriter = new GenericDatumWriter();
Encoder encoder = null;
Schema schema = null;
for (Object attachment : src.get(Fields.ATTACHMENT_BODY)) {
Preconditions.checkNotNull(attachment);
GenericContainer datum = (GenericContainer) attachment;
schema = getSchema(datum, schema);
assert schema != null;
datumWriter.setSchema(schema);
if (encoder == null) { // init
if (format == Format.containerlessJSON) {
encoder = EncoderFactory.get().jsonEncoder(schema, dst);
} else {
encoder = EncoderFactory.get().binaryEncoder(dst, null);
}
assert encoder != null;
}
datumWriter.write(datum, encoder);
}
encoder.flush();
} catch (IOException e) {
throw new MorphlineRuntimeException(e);
}
}
public void toStream(final OutputStream stream, final Object object) throws IOException {
final Object generic = encodeNode(object);
final Encoder encoder = EncoderFactory.get().directBinaryEncoder(stream, null);
final DatumWriter<Object> writer = new GenericDatumWriter<Object>(Schemas.NODE);
writer.write(generic, encoder);
encoder.flush();
}
public static void tweetToAvro(ByteArrayOutputStream out, Status status, Schema schema)
throws IOException {
GenericDatumWriter<GenericRecord> gdw = new GenericDatumWriter<>(schema);
Encoder e = EncoderFactory.get().binaryEncoder(out, null);
GenericRecord tweet = buildTweet(schema, status);
gdw.write(tweet, e);
e.flush();
}
public void toStream(
final OutputStream stream,
final Record object,
@Nullable final Set<URI> propertiesToSerialize)
throws IOException {
final Object generic = encodeRecord(object, propertiesToSerialize);
final Encoder encoder = EncoderFactory.get().directBinaryEncoder(stream, null);
final DatumWriter<Object> writer = new GenericDatumWriter<Object>(Schemas.NODE);
writer.write(generic, encoder);
encoder.flush();
}
@Override
@SuppressWarnings({"rawtypes", "unchecked"})
protected void writeMap(Schema schema, Object datum, Encoder out) throws IOException {
if (writeDirtyBits) {
// write extra state information for maps
StatefulMap<Utf8, ?> map = (StatefulMap) datum;
out.writeInt(map.states().size());
for (Entry<Utf8, State> e2 : map.states().entrySet()) {
out.writeString(e2.getKey());
out.writeInt(e2.getValue().ordinal());
}
}
super.writeMap(schema, datum, out);
}
public static void checkBlockingBinary(
Schema schema, Object datum, DatumWriter<Object> writer, DatumReader<Object> reader)
throws IOException {
ByteArrayOutputStream out = new ByteArrayOutputStream();
writer.setSchema(schema);
Encoder encoder = EncoderFactory.get().blockingBinaryEncoder(out, null);
writer.write(datum, encoder);
encoder.flush();
byte[] data = out.toByteArray();
reader.setSchema(schema);
Object decoded = reader.read(null, DecoderFactory.get().binaryDecoder(data, null));
assertEquals("Decoded data does not match.", datum, decoded);
}
/**
* Given an entity, an avro schema, and an encoder, write the entity to the encoder's underlying
* output stream.
*
* @param entity The entity we want to encode.
* @param encoder The Avro Encoder we will write to.
* @param writer The DatumWriter we'll use to encode the entity to the encoder.
*/
public static <T> void writeAvroEntity(T entity, Encoder encoder, DatumWriter<T> writer) {
try {
writer.write(entity, encoder);
encoder.flush();
} catch (IOException e) {
throw new SerializationException("Could not serialize Avro entity", e);
}
}
public static void checkBinaryJson(String json) throws Exception {
Object node = Json.parseJson(json);
ByteArrayOutputStream out = new ByteArrayOutputStream();
DatumWriter<Object> writer = new Json.ObjectWriter();
Encoder encoder = EncoderFactory.get().binaryEncoder(out, null);
encoder = EncoderFactory.get().validatingEncoder(Json.SCHEMA, encoder);
writer.write(node, encoder);
encoder.flush();
byte[] bytes = out.toByteArray();
DatumReader<Object> reader = new Json.ObjectReader();
Decoder decoder = DecoderFactory.get().binaryDecoder(bytes, null);
decoder = DecoderFactory.get().validatingDecoder(Json.SCHEMA, decoder);
Object decoded = reader.read(null, decoder);
assertEquals("Decoded json does not match.", Json.toString(node), Json.toString(decoded));
}
private static void checkJson(Schema schema, Object datum, String json) throws Exception {
ByteArrayOutputStream out = new ByteArrayOutputStream();
Encoder encoder = EncoderFactory.get().jsonEncoder(schema, out);
DatumWriter<Object> writer = new GenericDatumWriter<Object>();
writer.setSchema(schema);
writer.write(datum, encoder);
encoder.flush();
byte[] data = out.toByteArray();
String encoded = new String(data, "UTF-8");
assertEquals("Encoded data does not match.", json, encoded);
DatumReader<Object> reader = new GenericDatumReader<Object>();
reader.setSchema(schema);
Object decoded =
reader.read(null, DecoderFactory.get().jsonDecoder(schema, new ByteArrayInputStream(data)));
assertEquals("Decoded data does not match.", datum, decoded);
}
private static void checkJson(
Schema schema, Object datum, DatumWriter<Object> writer, DatumReader<Object> reader)
throws IOException {
ByteArrayOutputStream out = new ByteArrayOutputStream();
Encoder encoder = EncoderFactory.get().jsonEncoder(schema, out);
writer.setSchema(schema);
writer.write(datum, encoder);
writer.write(datum, encoder);
encoder.flush();
byte[] data = out.toByteArray();
reader.setSchema(schema);
Decoder decoder = DecoderFactory.get().jsonDecoder(schema, new ByteArrayInputStream(data));
Object decoded = reader.read(null, decoder);
assertEquals("Decoded data does not match.", datum, decoded);
decoded = reader.read(decoded, decoder);
assertEquals("Decoded data does not match.", datum, decoded);
}
@Test
public void testEnumMismatch() throws Exception {
Schema actual = Schema.parse("{\"type\":\"enum\",\"name\":\"E\",\"symbols\":[\"X\",\"Y\"]}");
Schema expected = Schema.parse("{\"type\":\"enum\",\"name\":\"E\",\"symbols\":[\"Y\",\"Z\"]}");
ByteArrayOutputStream out = new ByteArrayOutputStream();
DatumWriter<Object> writer = new GenericDatumWriter<Object>(actual);
Encoder encoder = EncoderFactory.get().directBinaryEncoder(out, null);
writer.write(new GenericData.EnumSymbol(actual, "Y"), encoder);
writer.write(new GenericData.EnumSymbol(actual, "X"), encoder);
encoder.flush();
byte[] data = out.toByteArray();
Decoder decoder = DecoderFactory.get().binaryDecoder(data, null);
DatumReader<String> in = new GenericDatumReader<String>(actual, expected);
assertEquals("Wrong value", new GenericData.EnumSymbol(expected, "Y"), in.read(null, decoder));
try {
in.read(null, decoder);
fail("Should have thrown exception.");
} catch (AvroTypeException e) {
// expected
}
}
public void serializeGeneric() throws IOException {
// Create a datum to serialize.
Schema schema = new Schema.Parser().parse(getClass().getResourceAsStream("/MyPair.avsc"));
GenericRecord datum = new GenericData.Record(schema);
datum.put("left", new Utf8("dog"));
datum.put("right", new Utf8("cat"));
// Serialize it.
ByteArrayOutputStream out = new ByteArrayOutputStream();
DatumWriter<GenericRecord> writer = new GenericDatumWriter<GenericRecord>(schema);
Encoder encoder = EncoderFactory.get().binaryEncoder(out, null);
writer.write(datum, encoder);
encoder.flush();
out.close();
System.out.println("Serialization: " + out);
// Deserialize it.
DatumReader<GenericRecord> reader = new GenericDatumReader<GenericRecord>(schema);
BinaryDecoder decoder = DecoderFactory.get().binaryDecoder(out.toByteArray(), null);
GenericRecord result = reader.read(null, decoder);
System.out.printf("Left: %s, Right: %s\n", result.get("left"), result.get("right"));
}
/**
* Gives the output message
*
* @param outputType output data type
* @param result mapping result
* @return the output as a String
* @throws IOException
*/
public String getOutputMessage(String outputType, GenericRecord result)
throws SynapseException, IOException {
DatumWriter<GenericRecord> writer = null;
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
Encoder encoder = new DummyEncoder(byteArrayOutputStream);
// OMElement outMessage = null;
String outMessage = null;
try {
writer = WriterRegistry.getInstance().get(outputType).newInstance();
writer.setSchema(result.getSchema());
writer.write(result, encoder);
if (log.isDebugEnabled()) {
log.debug("Output received from datum writer.." + byteArrayOutputStream.toString());
}
} catch (Exception e) {
handleException("Data coversion Failed at JSONWriter..", e);
} finally {
encoder.flush();
}
/*try {
// Converts the result into an OMElement
outMessage = getOutputResult(byteArrayOutputStream.toString());
} catch (XMLStreamException e) {
handleException(
"Failed at generating the OMElement for the JSON output received...",
e);
}*/
outMessage = byteArrayOutputStream.toString();
return outMessage;
}
@Test
public void experimentWithAvro() throws Exception {
String root = "org/hibernate/search/remote/codex/avro/v1_1/";
parseSchema(root + "attribute/TokenTrackingAttribute.avro", "attribute/TokenTrackingAttribute");
parseSchema(root + "attribute/CharTermAttribute.avro", "attribute/CharTermAttribute");
parseSchema(root + "attribute/PayloadAttribute.avro", "attribute/PayloadAttribute");
parseSchema(root + "attribute/KeywordAttribute.avro", "attribute/KeywordAttribute");
parseSchema(
root + "attribute/PositionIncrementAttribute.avro", "attribute/PositionIncrementAttribute");
parseSchema(root + "attribute/FlagsAttribute.avro", "attribute/FlagsAttribute");
parseSchema(root + "attribute/TypeAttribute.avro", "attribute/TypeAttribute");
parseSchema(root + "attribute/OffsetAttribute.avro", "attribute/OffsetAttribute");
parseSchema(root + "field/TermVector.avro", "field/TermVector");
parseSchema(root + "field/Index.avro", "field/Index");
parseSchema(root + "field/Store.avro", "field/Store");
parseSchema(root + "field/TokenStreamField.avro", "field/TokenStreamField");
parseSchema(root + "field/ReaderField.avro", "field/ReaderField");
parseSchema(root + "field/StringField.avro", "field/StringField");
parseSchema(root + "field/BinaryField.avro", "field/BinaryField");
parseSchema(root + "field/NumericIntField.avro", "field/NumericIntField");
parseSchema(root + "field/NumericLongField.avro", "field/NumericLongField");
parseSchema(root + "field/NumericFloatField.avro", "field/NumericFloatField");
parseSchema(root + "field/NumericDoubleField.avro", "field/NumericDoubleField");
parseSchema(root + "field/CustomFieldable.avro", "field/CustomFieldable");
parseSchema(root + "Document.avro", "Document");
parseSchema(root + "operation/Id.avro", "operation/Id");
parseSchema(root + "operation/OptimizeAll.avro", "operation/OptimizeAll");
parseSchema(root + "operation/PurgeAll.avro", "operation/PurgeAll");
parseSchema(root + "operation/Flush.avro", "operation/Flush");
parseSchema(root + "operation/Delete.avro", "operation/Delete");
parseSchema(root + "operation/Add.avro", "operation/Add");
parseSchema(root + "operation/Update.avro", "operation/Update");
parseSchema(root + "Message.avro", "Message");
String filename = root + "Works.avpr";
Protocol protocol = parseProtocol(filename, "Works");
final Schema termVectorSchema = protocol.getType("TermVector");
final Schema indexSchema = protocol.getType("Index");
final Schema storeSchema = protocol.getType("Store");
final Schema tokenTrackingAttribute = protocol.getType("TokenTrackingAttribute");
final Schema tokenStreamSchema = protocol.getType("TokenStreamField");
final Schema readerSchema = protocol.getType("ReaderField");
final Schema stringSchema = protocol.getType("StringField");
final Schema binarySchema = protocol.getType("BinaryField");
final Schema intFieldSchema = protocol.getType("NumericIntField");
final Schema longFieldSchema = protocol.getType("NumericLongField");
final Schema floatFieldSchema = protocol.getType("NumericFloatField");
final Schema doubleFieldSchema = protocol.getType("NumericDoubleField");
final Schema custonFieldableSchema = protocol.getType("CustomFieldable");
final Schema documentSchema = protocol.getType("Document");
final Schema idSchema = protocol.getType("Id");
final Schema optimizeAllSchema = protocol.getType("OptimizeAll");
final Schema purgeAllSchema = protocol.getType("PurgeAll");
final Schema flushSchema = protocol.getType("Flush");
final Schema deleteSchema = protocol.getType("Delete");
final Schema addSchema = protocol.getType("Add");
final Schema updateSchema = protocol.getType("Update");
Schema messageSchema = protocol.getType("Message");
final ByteArrayOutputStream out = new ByteArrayOutputStream();
GenericDatumWriter<GenericRecord> writer = new GenericDatumWriter<GenericRecord>(messageSchema);
Encoder encoder = EncoderFactory.get().directBinaryEncoder(out, null);
byte[] serializableSample = new byte[10];
for (int i = 0; i < 10; i++) {
serializableSample[i] = (byte) i;
}
List<String> classReferences = new ArrayList<String>();
classReferences.add(AvroTest.class.getName());
List<GenericRecord> fieldables = new ArrayList<GenericRecord>(1);
// custom fieldable
GenericRecord customFieldable = new GenericData.Record(custonFieldableSchema);
customFieldable.put("instance", ByteBuffer.wrap(serializableSample));
fieldables.add(customFieldable);
// numeric fields
GenericRecord numericField = createNumeric(intFieldSchema);
numericField.put("value", 3);
fieldables.add(numericField);
numericField = createNumeric(longFieldSchema);
numericField.put("value", 3l);
fieldables.add(numericField);
numericField = createNumeric(floatFieldSchema);
numericField.put("value", 2.3f);
fieldables.add(numericField);
numericField = createNumeric(doubleFieldSchema);
numericField.put("value", 2.3d);
fieldables.add(numericField);
// fields
GenericRecord field = createField(binarySchema);
field.put("offset", 0);
field.put("length", 10);
field.put("value", ByteBuffer.wrap(serializableSample));
fieldables.add(field);
field = createField(stringSchema);
field.put("value", stringSchema.getName());
field.put("store", "YES");
field.put("index", "ANALYZED");
field.put("termVector", "WITH_OFFSETS");
fieldables.add(field);
field = createField(tokenStreamSchema);
List<List<Object>> tokens = new ArrayList<List<Object>>();
List<Object> attrs = new ArrayList<Object>();
tokens.add(attrs);
GenericData.Record attr = new GenericData.Record(tokenTrackingAttribute);
List<Integer> positions = new ArrayList<Integer>();
positions.add(1);
positions.add(2);
positions.add(3);
positions.add(4);
attr.put("positions", positions);
attrs.add(attr);
attrs.add(ByteBuffer.wrap(serializableSample));
field.put("value", tokens);
field.put("termVector", "WITH_OFFSETS");
fieldables.add(field);
field = createField(readerSchema);
field.put("value", ByteBuffer.wrap(serializableSample));
field.put("termVector", "WITH_OFFSETS");
fieldables.add(field);
GenericRecord doc = new GenericData.Record(documentSchema);
doc.put("boost", 2.3f);
doc.put("fieldables", fieldables);
GenericRecord add = new GenericData.Record(addSchema);
add.put("class", classReferences.indexOf(AvroTest.class.getName()));
GenericRecord id = new GenericData.Record(idSchema);
id.put("value", ByteBuffer.wrap(serializableSample));
add.put("id", id);
add.put("document", doc);
Map<String, String> analyzers = new HashMap<String, String>();
analyzers.put("name", "ngram");
analyzers.put("description", "porter");
add.put("fieldToAnalyzerMap", analyzers);
GenericRecord delete = new GenericData.Record(deleteSchema);
delete.put("class", classReferences.indexOf(AvroTest.class.getName()));
id = new GenericData.Record(idSchema);
id.put("value", new Long(30));
delete.put("id", id);
GenericRecord purgeAll = new GenericData.Record(purgeAllSchema);
purgeAll.put("class", classReferences.indexOf(AvroTest.class.getName()));
GenericRecord optimizeAll = new GenericData.Record(optimizeAllSchema);
GenericRecord flush = new GenericData.Record(flushSchema);
List<GenericRecord> operations = new ArrayList<GenericRecord>(1);
operations.add(purgeAll);
operations.add(optimizeAll);
operations.add(flush);
operations.add(delete);
operations.add(add);
GenericRecord message = new GenericData.Record(messageSchema);
message.put("classReferences", classReferences);
message.put("operations", operations);
writer.write(message, encoder);
encoder.flush();
ByteArrayInputStream inputStream = new ByteArrayInputStream(out.toByteArray());
Decoder decoder = DecoderFactory.get().binaryDecoder(inputStream, null);
GenericDatumReader<GenericRecord> reader = new GenericDatumReader<GenericRecord>(messageSchema);
while (true) {
try {
GenericRecord result = reader.read(null, decoder);
System.out.println(result);
assertThat(result).isNotNull();
// operations
assertThat(result.get("operations")).isNotNull().isInstanceOf(List.class);
List<?> ops = (List<?>) result.get("operations");
assertThat(ops).hasSize(5);
// Flush
assertThat(ops.get(2)).isInstanceOf(GenericRecord.class);
GenericRecord flushOp = (GenericRecord) ops.get(2);
assertThat(flushOp.getSchema().getName()).isEqualTo("Flush");
// Delete
assertThat(ops.get(3)).isInstanceOf(GenericRecord.class);
GenericRecord deleteOp = (GenericRecord) ops.get(3);
assertThat(deleteOp.getSchema().getName()).isEqualTo("Delete");
Object actual = ((GenericRecord) deleteOp.get("id")).get("value");
assertThat(actual).isInstanceOf(Long.class);
assertThat(actual).isEqualTo(Long.valueOf(30));
// Add
assertThat(ops.get(4)).isInstanceOf(GenericRecord.class);
GenericRecord addOp = (GenericRecord) ops.get(4);
assertThat(addOp.getSchema().getName()).isEqualTo("Add");
actual = ((GenericRecord) addOp.get("id")).get("value");
assertThat(actual).isInstanceOf(ByteBuffer.class);
ByteBuffer bb = (ByteBuffer) actual;
assertThat(bb.hasArray()).isTrue();
byte[] copy = new byte[bb.remaining()];
bb.get(copy);
assertThat(serializableSample).isEqualTo(copy);
// fieldToAnalyzerMap
assertThat(addOp.get("fieldToAnalyzerMap")).isInstanceOf(Map.class);
assertThat((Map) addOp.get("fieldToAnalyzerMap")).hasSize(2);
// document
assertThat(addOp.get("document")).isNotNull();
GenericRecord document = (GenericRecord) addOp.get("document");
assertThat(document.get("boost")).isEqualTo(2.3f);
// numeric fields
assertThat(document.get("fieldables")).isNotNull().isInstanceOf(List.class);
List<?> fields = (List<?>) document.get("fieldables");
assertThat(fields).hasSize(9); // custom + 4 numerics + 4 fields
field = (GenericRecord) fields.get(0);
assertThat(field.getSchema().getName()).isEqualTo("CustomFieldable");
field = (GenericRecord) fields.get(1);
assertThat(field.getSchema().getName()).isEqualTo("NumericIntField");
assertThat(field.get("value")).isEqualTo(3);
assertNumericField(field);
field = (GenericRecord) fields.get(2);
assertThat(field.getSchema().getName()).isEqualTo("NumericLongField");
assertThat(field.get("value")).isEqualTo(3l);
assertNumericField(field);
field = (GenericRecord) fields.get(3);
assertThat(field.getSchema().getName()).isEqualTo("NumericFloatField");
assertThat(field.get("value")).isEqualTo(2.3f);
assertNumericField(field);
field = (GenericRecord) fields.get(4);
assertThat(field.getSchema().getName()).isEqualTo("NumericDoubleField");
assertThat(field.get("value")).isEqualTo(2.3d);
assertNumericField(field);
// fields
field = (GenericRecord) fields.get(5);
assertThat(field.getSchema().getName()).isEqualTo("BinaryField");
assertThat(field.get("value")).isInstanceOf(ByteBuffer.class);
assertField(field);
field = (GenericRecord) fields.get(6);
assertThat(field.getSchema().getName()).isEqualTo("StringField");
assertThat(field.get("value")).isInstanceOf(Utf8.class);
assertTermVector(field);
assertIndexAndStore(field);
assertField(field);
field = (GenericRecord) fields.get(7);
assertThat(field.getSchema().getName()).isEqualTo("TokenStreamField");
assertThat(field.get("value")).isInstanceOf(List.class);
List<List<Object>> l1 = (List<List<Object>>) field.get("value");
assertThat(l1.get(0)).as("Wrong attribute impl list").hasSize(2);
Object object = l1.get(0).get(0);
assertThat(object).isNotNull();
assertTermVector(field);
assertField(field);
field = (GenericRecord) fields.get(8);
assertThat(field.getSchema().getName()).isEqualTo("ReaderField");
assertThat(field.get("value")).isInstanceOf(ByteBuffer.class);
assertTermVector(field);
assertField(field);
} catch (EOFException eof) {
break;
} catch (Exception ex) {
ex.printStackTrace();
throw ex;
}
}
}