def _ingest_test_getfeaturetable_mocked_resp(file_url: str,
date_partition_col: str = None):
return GetFeatureTableResponse(table=FeatureTableProto(
spec=FeatureTableSpecProto(
name="ingest_featuretable",
max_age=Duration(seconds=3600),
features=[
FeatureSpecProto(
name="dev_feature_float",
value_type=ValueProto.ValueType.FLOAT,
),
FeatureSpecProto(
name="dev_feature_string",
value_type=ValueProto.ValueType.STRING,
),
],
entities=["dev_entity"],
batch_source=DataSourceProto(
file_options=DataSourceProto.FileOptions(
file_format=ParquetFormat().to_proto(), file_url=file_url),
event_timestamp_column="datetime",
created_timestamp_column="timestamp",
date_partition_column=date_partition_col
if date_partition_col is not None else None,
),
),
meta=FeatureTableMetaProto(),
))
def to_proto(self) -> FeatureTableProto:
"""
Converts an feature table object to its protobuf representation
Returns:
FeatureTableProto protobuf
"""
meta = FeatureTableMetaProto(
created_timestamp=self.created_timestamp,
last_updated_timestamp=self.last_updated_timestamp,
)
spec = FeatureTableSpecProto(
name=self.name,
entities=self.entities,
features=[
feature.to_proto() if type(feature) == Feature else feature
for feature in self.features
],
labels=self.labels,
max_age=self.max_age,
batch_source=(self.batch_source.to_proto() if issubclass(
type(self.batch_source), DataSource) else self.batch_source),
stream_source=(self.stream_source.to_proto() if issubclass(
type(self.stream_source), DataSource) else self.stream_source),
)
return FeatureTableProto(spec=spec, meta=meta)
def to_proto(self) -> FeatureTableProto:
"""
Converts an feature table object to its protobuf representation
Returns:
FeatureTableProto protobuf
"""
meta = FeatureTableMetaProto(
created_timestamp=self.created_timestamp,
last_updated_timestamp=self.last_updated_timestamp,
)
return FeatureTableProto(spec=self.to_spec_proto(), meta=meta)
def ApplyFeatureTable(self, request: ApplyFeatureTableRequest, context):
feature_table_spec = request.table_spec
feature_table_meta = FeatureTableMeta(
created_timestamp=Timestamp(seconds=10), )
applied_feature_table = FeatureTableProto(spec=feature_table_spec,
meta=feature_table_meta)
self._feature_tables[feature_table_spec.name] = applied_feature_table
_logger.info("registered feature table " + feature_table_spec.name +
" with " + str(len(feature_table_spec.entities)) +
" entities and " + str(len(feature_table_spec.features)) +
" features")
return ApplyFeatureTableResponse(table=applied_feature_table, )
def from_dict(cls, ft_dict):
"""
Creates a feature table from a dict
Args:
ft_dict: A dict representation of a feature table
Returns:
Returns a FeatureTable object based on the feature table dict
"""
feature_table_proto = json_format.ParseDict(
ft_dict, FeatureTableProto(), ignore_unknown_fields=True
)
return cls.from_proto(feature_table_proto)
def to_proto(self) -> FeatureTableProto:
"""
Converts an feature table object to its protobuf representation
Returns:
FeatureTableProto protobuf
"""
meta = FeatureTableMetaProto(
created_timestamp=self.created_timestamp,
last_updated_timestamp=self.last_updated_timestamp,
)
spec = FeatureTableSpecProto(
name=self.name,
entities=self.entities,
features=self.features,
labels=self.labels,
max_age=self.max_age,
batch_source=self.batch_source,
stream_source=self.stream_source,
)
return FeatureTableProto(spec=spec, meta=meta)
