Skip to content

Fix Redis key serialization in java feature server #2264

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
feast-ci-bot merged 1 commit into from
Feb 2, 2022
Merged

Fix Redis key serialization in java feature server #2264

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions java/serving/src/test/java/feast/serving/it/ServingBaseTests.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -157,5 +157,28 @@ public void shouldRefreshRegistryAndServeNewFeatures() throws InterruptedExcepti
equalTo(3));
}

/** https://github.com/feast-dev/feast/issues/2253 */
@Test
public void shouldGetOnlineFeaturesWithStringEntity() {
Map<String, ValueProto.RepeatedValue> entityRows =
ImmutableMap.of(
"entity",
ValueProto.RepeatedValue.newBuilder()
.addVal(DataGenerator.createStrValue("key-1"))
.build());

ImmutableList<String> featureReferences =
ImmutableList.of("feature_view_0:feature_0", "feature_view_0:feature_1");

ServingAPIProto.GetOnlineFeaturesRequest req =
TestUtils.createOnlineFeatureRequest(featureReferences, entityRows);

ServingAPIProto.GetOnlineFeaturesResponse resp = servingStub.getOnlineFeatures(req);

for (final int featureIdx : List.of(0, 1)) {
assertEquals(FieldStatus.PRESENT, resp.getResults(featureIdx).getStatuses(0));
}
}

abstract void updateRegistryFile(RegistryProto.Registry registry);
}
3 changes: 2 additions & 1 deletion java/serving/src/test/java/feast/serving/it/ServingBenchmarkIT.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,8 @@ protected ServingAPIProto.GetOnlineFeaturesRequest buildOnlineRequest(
int rowsCount, int featuresCount) {
List<ValueProto.Value> entities =
IntStream.range(0, rowsCount)
.mapToObj(i -> DataGenerator.createInt64Value(rand.nextInt(1000)))
.mapToObj(
i -> DataGenerator.createStrValue(String.format("key-%s", rand.nextInt(1000))))
.collect(Collectors.toList());

List<String> featureReferences =
Expand Down
2 changes: 1 addition & 1 deletion java/serving/src/test/resources/docker-compose/feast10/definitions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,7 @@ def transformed_conv_rate(features_df: pd.DataFrame) -> pd.DataFrame:

entity = Entity(
name="entity",
value_type=ValueType.INT64,
value_type=ValueType.STRING,
)

benchmark_feature_views = [
Expand Down
12 changes: 8 additions & 4 deletions java/serving/src/test/resources/docker-compose/feast10/materialize.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,22 +28,26 @@
# for more info.
df.to_parquet("driver_stats.parquet")


# For Benchmarks
# Please read more in Feast RFC-031
# (link https://docs.google.com/document/d/12UuvTQnTTCJhdRgy6h10zSbInNGSyEJkIxpOcgOen1I/edit)
# about this benchmark setup
def generate_data(num_rows: int, num_features: int, key_space: int, destination: str) -> pd.DataFrame:
def generate_data(num_rows: int, num_features: int, destination: str) -> pd.DataFrame:
features = [f"feature_{i}" for i in range(num_features)]
columns = ["entity", "event_timestamp"] + features
df = pd.DataFrame(0, index=np.arange(num_rows), columns=columns)
df["event_timestamp"] = datetime.utcnow()
for column in ["entity"] + features:
df[column] = np.random.randint(1, key_space, num_rows)
for column in features:
df[column] = np.random.randint(1, num_rows, num_rows)

df["entity"] = "key-" + \
pd.Series(np.arange(1, num_rows + 1)).astype(pd.StringDtype())

df.to_parquet(destination)


generate_data(10**3, 250, 10**3, "benchmark_data.parquet")
generate_data(10**3, 250, "benchmark_data.parquet")


fs = FeatureStore(".")
Expand Down
89 changes: 43 additions & 46 deletions ...s/redis/src/main/java/feast/storage/connectors/redis/retriever/EntityKeySerializerV2.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,16 +16,14 @@
*/
package feast.storage.connectors.redis.retriever;

import com.google.common.primitives.UnsignedBytes;
import com.google.protobuf.ProtocolStringList;
import feast.proto.storage.RedisProto;
import feast.proto.types.ValueProto;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.nio.charset.StandardCharsets;
import java.util.ArrayList;
import java.util.Comparator;
import java.util.List;
import java.util.*;
import org.apache.commons.lang3.ArrayUtils;
import org.apache.commons.lang3.tuple.Pair;

// This is derived from
Expand All @@ -48,70 +46,52 @@ public byte[] serialize(RedisProto.RedisKeyV2 entityKey) {
}
tuples.sort(Comparator.comparing(Pair::getLeft));

ByteBuffer stringBytes = ByteBuffer.allocate(Integer.BYTES);
stringBytes.order(ByteOrder.LITTLE_ENDIAN);
stringBytes.putInt(ValueProto.ValueType.Enum.STRING.getNumber());

for (Pair<String, ValueProto.Value> pair : tuples) {
for (final byte b : stringBytes.array()) {
buffer.add(b);
}
for (final byte b : pair.getLeft().getBytes(StandardCharsets.UTF_8)) {
buffer.add(b);
}
buffer.addAll(encodeInteger(ValueProto.ValueType.Enum.STRING.getNumber()));
buffer.addAll(encodeString(pair.getLeft()));
}

for (Pair<String, ValueProto.Value> pair : tuples) {
final ValueProto.Value val = pair.getRight();
switch (val.getValCase()) {
case STRING_VAL:
buffer.add(UnsignedBytes.checkedCast(ValueProto.ValueType.Enum.STRING.getNumber()));
buffer.add(
UnsignedBytes.checkedCast(
val.getStringVal().getBytes(StandardCharsets.UTF_8).length));
for (final byte b : val.getStringVal().getBytes(StandardCharsets.UTF_8)) {
buffer.add(b);
}
String stringVal = val.getStringVal();

buffer.addAll(encodeInteger(ValueProto.ValueType.Enum.STRING.getNumber()));
buffer.addAll(encodeInteger(stringVal.length()));
buffer.addAll(encodeString(stringVal));

break;
case BYTES_VAL:
buffer.add(UnsignedBytes.checkedCast(ValueProto.ValueType.Enum.BYTES.getNumber()));
for (final byte b : val.getBytesVal().toByteArray()) {
buffer.add(b);
}
byte[] bytes = val.getBytesVal().toByteArray();

buffer.addAll(encodeInteger(ValueProto.ValueType.Enum.BYTES.getNumber()));
buffer.addAll(encodeInteger(bytes.length));
buffer.addAll(encodeBytes(bytes));

break;
case INT32_VAL:
ByteBuffer int32ByteBuffer =
ByteBuffer.allocate(Integer.BYTES + Integer.BYTES + Integer.BYTES);
int32ByteBuffer.order(ByteOrder.LITTLE_ENDIAN);
int32ByteBuffer.putInt(ValueProto.ValueType.Enum.INT32.getNumber());
int32ByteBuffer.putInt(Integer.BYTES);
int32ByteBuffer.putInt(val.getInt32Val());
for (final byte b : int32ByteBuffer.array()) {
buffer.add(b);
}
buffer.addAll(encodeInteger(ValueProto.ValueType.Enum.INT32.getNumber()));
buffer.addAll(encodeInteger(Integer.BYTES));
buffer.addAll(encodeInteger(val.getInt32Val()));

break;
case INT64_VAL:
ByteBuffer int64ByteBuffer =
ByteBuffer.allocate(Integer.BYTES + Integer.BYTES + Integer.BYTES);
int64ByteBuffer.order(ByteOrder.LITTLE_ENDIAN);
int64ByteBuffer.putInt(ValueProto.ValueType.Enum.INT64.getNumber());
int64ByteBuffer.putInt(Integer.BYTES);
buffer.addAll(encodeInteger(ValueProto.ValueType.Enum.INT64.getNumber()));
buffer.addAll(encodeInteger(Integer.BYTES));
/* This is super dumb - but in https://github.com/feast-dev/feast/blob/dcae1606f53028ce5413567fb8b66f92cfef0f8e/sdk/python/feast/infra/key_encoding_utils.py#L9
we use `struct.pack("<l", v.int64_val)` to get the bytes of an int64 val. This actually extracts only 4 bytes,
instead of 8 bytes as you'd expect from to serialize an int64 value.
*/
int64ByteBuffer.putInt(Long.valueOf(val.getInt64Val()).intValue());
for (final byte b : int64ByteBuffer.array()) {
buffer.add(b);
}
buffer.addAll(encodeInteger(((Long) val.getInt64Val()).intValue()));

break;
default:
throw new RuntimeException("Unable to serialize Entity Key");
}
}
for (final byte b : entityKey.getProject().getBytes(StandardCharsets.UTF_8)) {
buffer.add(b);
}

buffer.addAll(encodeString(entityKey.getProject()));

final byte[] bytes = new byte[buffer.size()];
for (int i = 0; i < buffer.size(); i++) {
Expand All @@ -120,4 +100,21 @@ public byte[] serialize(RedisProto.RedisKeyV2 entityKey) {

return bytes;
}

private List<Byte> encodeBytes(byte[] toByteArray) {
return Arrays.asList(ArrayUtils.toObject(toByteArray));
}

private List<Byte> encodeInteger(Integer value) {
ByteBuffer buffer = ByteBuffer.allocate(Integer.BYTES);
buffer.order(ByteOrder.LITTLE_ENDIAN);
buffer.putInt(value);

return Arrays.asList(ArrayUtils.toObject(buffer.array()));
}

private List<Byte> encodeString(String value) {
byte[] stringBytes = value.getBytes(StandardCharsets.UTF_8);
return encodeBytes(stringBytes);
}
}