Skip to content
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
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
Prev Previous commit
Next Next commit
Add new version of serialization and desrialization
Signed-off-by: cmuhao <sduxuhao@gmail.com>
  • Loading branch information
HaoXuAI committed Jun 16, 2024
commit c8f6c2bf4b5569c635bd4ae8b2178b0b08a183ba
29 changes: 15 additions & 14 deletions sdk/python/feast/infra/key_encoding_utils.py
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@


def _serialize_val(
value_type, v: ValueProto, entity_key_serialization_version=1
value_type, v: ValueProto, entity_key_serialization_version=1
) -> Tuple[bytes, int]:
if value_type == "string_val":
return v.string_val.encode("utf8"), ValueType.STRING
Expand All @@ -31,7 +31,7 @@ def _deserialize_value(value_type, value_bytes) -> ValueProto:
value = struct.unpack("<i", value_bytes)[0]
return ValueProto(int32_val=value)
elif value_type == ValueType.STRING:
value = value_bytes.decode('utf-8')
value = value_bytes.decode("utf-8")
return ValueProto(string_val=value)
elif value_type == ValueType.BYTES:
return ValueProto(bytes_val=value_bytes)
Expand All @@ -56,7 +56,7 @@ def serialize_entity_key_prefix(entity_keys: List[str]) -> bytes:


def serialize_entity_key(
entity_key: EntityKeyProto, entity_key_serialization_version=1
entity_key: EntityKeyProto, entity_key_serialization_version=1
) -> bytes:
"""
Serialize entity key to a bytestring so it can be used as a lookup key in a hash table.
Expand Down Expand Up @@ -92,7 +92,9 @@ def serialize_entity_key(
return b"".join(output)


def deserialize_entity_key(serialized_entity_key: bytes, entity_key_serialization_version=3) -> EntityKeyProto:
def deserialize_entity_key(
serialized_entity_key: bytes, entity_key_serialization_version=3
) -> EntityKeyProto:
"""
Deserialize entity key from a bytestring. This function can only be used with entity_key_serialization_version > 2.
Args:
Expand All @@ -114,32 +116,31 @@ def deserialize_entity_key(serialized_entity_key: bytes, entity_key_serializatio
offset += 4

# Read the length of the key
key_length = struct.unpack_from('<I', serialized_entity_key, offset)[0]
key_length = struct.unpack_from("<I", serialized_entity_key, offset)[0]
offset += 4

if key_type == ValueType.STRING:
key = struct.unpack_from(f'<{key_length}s', serialized_entity_key, offset)[0]
keys.append(key.decode('utf-8').rstrip('\x00'))
key = struct.unpack_from(f"<{key_length}s", serialized_entity_key, offset)[
0
]
keys.append(key.decode("utf-8").rstrip("\x00"))
offset += key_length
else:
raise ValueError(f"Unsupported key type: {key_type}")

value_type, = struct.unpack_from('<I', serialized_entity_key, offset)
(value_type,) = struct.unpack_from("<I", serialized_entity_key, offset)
offset += 4

value_length, = struct.unpack_from('<I', serialized_entity_key, offset)
(value_length,) = struct.unpack_from("<I", serialized_entity_key, offset)
offset += 4

# Read the value based on its type and length
value_bytes = serialized_entity_key[offset:offset + value_length]
value_bytes = serialized_entity_key[offset : offset + value_length]
value = _deserialize_value(value_type, value_bytes)
values.append(value)
offset += value_length

return EntityKeyProto(
join_keys=keys,
entity_values=values
)
return EntityKeyProto(join_keys=keys, entity_values=values)


def get_list_val_str(val):
Expand Down
16 changes: 9 additions & 7 deletions sdk/python/tests/unit/infra/test_key_encoding_utils.py
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
import pytest

from feast.infra.key_encoding_utils import serialize_entity_key, deserialize_entity_key
from feast.infra.key_encoding_utils import deserialize_entity_key, serialize_entity_key
from feast.protos.feast.types.EntityKey_pb2 import EntityKey as EntityKeyProto
from feast.protos.feast.types.Value_pb2 import Value as ValueProto

Expand All @@ -9,14 +9,14 @@ def test_serialize_entity_key():
# Should be fine
serialize_entity_key(
EntityKeyProto(
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2 ** 15))]
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2**15))]
),
entity_key_serialization_version=2,
)
# True int64, but should also be fine.
serialize_entity_key(
EntityKeyProto(
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2 ** 31))]
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2**31))]
),
entity_key_serialization_version=2,
)
Expand All @@ -25,20 +25,22 @@ def test_serialize_entity_key():
with pytest.raises(BaseException):
serialize_entity_key(
EntityKeyProto(
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2 ** 31))]
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2**31))]
),
)


def test_deserialize_entity_key():
serialized_entity_key = serialize_entity_key(
EntityKeyProto(
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2 ** 15))]
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2**15))]
),
entity_key_serialization_version=3,
)

deserialized_entity_key = deserialize_entity_key(serialized_entity_key, entity_key_serialization_version=3)
deserialized_entity_key = deserialize_entity_key(
serialized_entity_key, entity_key_serialization_version=3
)
assert deserialized_entity_key == EntityKeyProto(
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2 ** 15))]
join_keys=["user"], entity_values=[ValueProto(int64_val=int(2**15))]
)