-
Notifications
You must be signed in to change notification settings - Fork 1.3k
Expand file tree
/
Copy pathfeature_store.py
More file actions
4307 lines (3765 loc) · 170 KB
/
feature_store.py
File metadata and controls
4307 lines (3765 loc) · 170 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# Copyright 2019 The Feast Authors
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import asyncio
import copy
import itertools
import logging
import os
import time
import warnings
from datetime import datetime, timedelta
from pathlib import Path
from typing import (
TYPE_CHECKING,
Any,
Dict,
Iterable,
List,
Mapping,
Optional,
Sequence,
Tuple,
Union,
cast,
)
if TYPE_CHECKING:
from feast.diff.apply_progress import ApplyProgressContext
import pandas as pd
import pyarrow as pa
from colorama import Fore, Style
from fastapi.concurrency import run_in_threadpool
from google.protobuf.timestamp_pb2 import Timestamp
from tqdm import tqdm
from feast import feature_server, flags_helper, ui_server, utils
from feast.base_feature_view import BaseFeatureView
from feast.batch_feature_view import BatchFeatureView
from feast.data_source import (
DataSource,
KafkaSource,
KinesisSource,
PushMode,
PushSource,
)
from feast.diff.infra_diff import InfraDiff, diff_infra_protos
from feast.diff.registry_diff import RegistryDiff, apply_diff_to_registry, diff_between
from feast.dqm.errors import ValidationFailed
from feast.entity import Entity
from feast.errors import (
ConflictingFeatureViewNames,
DataFrameSerializationError,
DataSourceRepeatNamesException,
FeatureViewNotFoundException,
PushSourceNotFoundException,
RequestDataNotFoundInEntityDfException,
)
from feast.feast_object import FeastObject
from feast.feature_service import FeatureService
from feast.feature_view import (
DUMMY_ENTITY,
DUMMY_ENTITY_NAME,
FeatureView,
FeatureViewState,
)
from feast.inference import (
update_data_sources_with_inferred_event_timestamp_col,
update_feature_views_with_inferred_features_and_entities,
)
from feast.infra.infra_object import Infra
from feast.infra.offline_stores.offline_utils import (
DEFAULT_ENTITY_DF_EVENT_TIMESTAMP_COL,
)
from feast.infra.provider import Provider, RetrievalJob, get_provider
from feast.infra.registry.base_registry import BaseRegistry
from feast.infra.registry.registry import Registry
from feast.infra.registry.sql import SqlRegistry
from feast.labeling.label_view import LabelView
from feast.on_demand_feature_view import OnDemandFeatureView
from feast.online_response import OnlineResponse
from feast.permissions.permission import Permission
from feast.project import Project
from feast.protos.feast.serving.ServingService_pb2 import (
FieldStatus,
GetOnlineFeaturesResponse,
)
from feast.protos.feast.types.EntityKey_pb2 import EntityKey
from feast.protos.feast.types.Value_pb2 import RepeatedValue, Value
from feast.protos.feast.types.Value_pb2 import Value as ValueProto
from feast.repo_config import RepoConfig, load_repo_config
from feast.repo_contents import RepoContents
from feast.saved_dataset import SavedDataset, SavedDatasetStorage, ValidationReference
from feast.ssl_ca_trust_store_setup import configure_ca_trust_store_env_variables
from feast.stream_feature_view import StreamFeatureView
from feast.transformation.pandas_transformation import PandasTransformation
from feast.transformation.python_transformation import PythonTransformation
from feast.utils import _get_feature_view_vector_field_metadata, _utc_now
from feast.version_utils import parse_version
try:
from datetime import timezone as _timezone
except ImportError:
_timezone = None # type: ignore[assignment,misc]
_track_materialization = None # Lazy-loaded on first materialization call
_track_materialization_loaded = False
_logger = logging.getLogger(__name__)
def _get_track_materialization():
"""Lazy-import feast.metrics only when materialization tracking is needed.
Avoids importing the metrics module (and its prometheus_client /
psutil dependencies plus temp-dir creation) for every FeatureStore
usage such as ``feast apply`` or simple SDK reads.
"""
global _track_materialization, _track_materialization_loaded
if not _track_materialization_loaded:
_track_materialization_loaded = True
try:
from feast.metrics import track_materialization
_track_materialization = track_materialization
except Exception: # pragma: no cover
_track_materialization = None
return _track_materialization
warnings.simplefilter("once", DeprecationWarning)
_UNSET = object()
class FeatureStore:
"""
A FeatureStore object is used to define, create, and retrieve features.
Attributes:
config: The config for the feature store.
repo_path: The path to the feature repo.
_registry: The registry for the feature store.
_provider: The provider for the feature store.
_openlineage_emitter: Optional OpenLineage emitter for lineage tracking.
"""
config: RepoConfig
repo_path: Path
_registry: Optional[BaseRegistry]
_provider: Optional[Provider]
_openlineage_emitter: Optional[Any] = None
_feature_service_cache: Dict[str, List[str]]
def __init__(
self,
repo_path: Optional[str] = None,
config: Optional[RepoConfig] = None,
fs_yaml_file: Optional[Path] = None,
):
"""
Creates a FeatureStore object.
Args:
repo_path (optional): Path to the feature repo. Defaults to the current working directory.
config (optional): Configuration object used to configure the feature store.
fs_yaml_file (optional): Path to the `feature_store.yaml` file used to configure the feature store.
At most one of 'fs_yaml_file' and 'config' can be set.
Raises:
ValueError: If both or neither of repo_path and config are specified.
"""
if fs_yaml_file is not None and config is not None:
raise ValueError("You cannot specify both fs_yaml_file and config.")
configure_ca_trust_store_env_variables()
if repo_path:
self.repo_path = Path(repo_path)
else:
self.repo_path = Path(os.getcwd())
# If config is specified, or fs_yaml_file is specified, those take precedence over
# the default feature_store.yaml location under repo_path.
if config is not None:
self.config = config
elif fs_yaml_file is not None:
self.config = load_repo_config(self.repo_path, fs_yaml_file)
else:
self.config = load_repo_config(
self.repo_path, utils.get_default_yaml_file_path(self.repo_path)
)
# Initialize lazy-loaded components as None
self._registry = None
self._provider = None
self._openlineage_emitter = None
# Initialize feature service cache for performance optimization
self._feature_service_cache = {}
# Cache for _resolve_feature_service_name lookups
self._fs_name_cache: Dict[frozenset, Optional[str]] = {}
self._fs_name_index: Dict[frozenset, str] = {}
self._fs_name_index_ts: float = -self._FS_NAME_INDEX_TTL_SECONDS
self._mlflow_client: Any = _UNSET
def _init_mlflow(self) -> Optional[Any]:
"""Bootstrap MLflow integration on first access.
Checks the config, imports the module, and creates the integration
client. Returns the client or ``None`` if MLflow is disabled or
unavailable.
"""
try:
mlflow_cfg = getattr(self.config, "mlflow", None)
if mlflow_cfg is None or not mlflow_cfg.enabled:
return None
from feast.mlflow import _register_store
_register_store(self)
from feast.mlflow_integration.client import FeastMlflowClient
return FeastMlflowClient(self)
except ImportError:
return None
except Exception as e:
warnings.warn(f"Failed to configure MLflow tracking: {e}")
return None
@property
def mlflow(self) -> Any:
"""Access the Feast–MLflow integration client.
Lazily initializes on first access. Returns ``None`` when MLflow
integration is not enabled, allowing callers to guard with
``if store.mlflow:``.
"""
if self._mlflow_client is _UNSET:
self._mlflow_client = self._init_mlflow()
return self._mlflow_client
@staticmethod
def _count_entities(entity_rows: Any) -> int:
"""Count entities from either a list or columnar mapping."""
if isinstance(entity_rows, list):
return len(entity_rows)
if isinstance(entity_rows, Mapping):
try:
_first_col = next(iter(entity_rows.values()))
if isinstance(_first_col, RepeatedValue):
return len(_first_col.val)
return len(_first_col)
except Exception:
return 0
return 0
_FS_NAME_INDEX_TTL_SECONDS = 300
def _rebuild_fs_name_index(self) -> None:
"""Rebuild the {frozenset(refs) → service_name} index from the registry."""
index: Dict[frozenset, str] = {}
for fs in self.registry.list_feature_services(self.project, allow_cache=True):
fs_refs = frozenset(
f"{p.name_to_use()}:{f.name}"
for p in fs.feature_view_projections
for f in p.features
)
index[fs_refs] = fs.name
self._fs_name_index = index
self._fs_name_cache = {}
self._fs_name_index_ts = time.monotonic()
def _resolve_feature_service_name(self, feature_refs: List[str]) -> Optional[str]:
"""Find the best-matching feature service for the given feature refs.
Resolution: exact match wins immediately; otherwise the smallest
superset (fewest extra features) is returned. The full index is
rebuilt from the registry every _FS_NAME_INDEX_TTL_SECONDS and
per-query results are cached for O(1) repeated lookups.
"""
try:
now = time.monotonic()
if (now - self._fs_name_index_ts) >= self._FS_NAME_INDEX_TTL_SECONDS:
self._rebuild_fs_name_index()
ref_key = frozenset(feature_refs)
if ref_key in self._fs_name_cache:
return self._fs_name_cache[ref_key]
if ref_key in self._fs_name_index:
self._fs_name_cache[ref_key] = self._fs_name_index[ref_key]
return self._fs_name_index[ref_key]
best_match = None
best_extra = float("inf")
for fs_refs, fs_name in self._fs_name_index.items():
if ref_key.issubset(fs_refs):
extra = len(fs_refs) - len(ref_key)
if extra < best_extra:
best_match = fs_name
best_extra = extra
self._fs_name_cache[ref_key] = best_match
return best_match
except Exception as e:
_logger.debug("Failed to resolve feature service name: %s", e)
return None
def _log_entity_df_metadata(self, entity_df, start_date=None, end_date=None):
"""Log lightweight entity_df metadata to MLflow."""
try:
if self.mlflow is not None:
self.mlflow.log_entity_df_metadata(entity_df, start_date, end_date)
except Exception as e:
_logger.debug("Failed to log entity_df metadata to MLflow: %s", e)
def _log_entity_df_artifact(self, entity_df):
"""Upload entity DataFrame as a parquet artifact to MLflow."""
try:
if self.mlflow is not None:
self.mlflow.log_entity_df_artifact(entity_df)
except Exception as e:
_logger.debug("Failed to log entity_df artifact to MLflow: %s", e)
def _init_openlineage_emitter(self) -> Optional[Any]:
"""Initialize OpenLineage emitter if configured and enabled."""
try:
if (
hasattr(self.config, "openlineage")
and self.config.openlineage is not None
and self.config.openlineage.enabled
):
from feast.openlineage import FeastOpenLineageEmitter
ol_config = self.config.openlineage.to_openlineage_config()
emitter = FeastOpenLineageEmitter(ol_config)
if emitter.is_enabled:
return emitter
except ImportError:
# OpenLineage not installed, silently skip
pass
except Exception as e:
warnings.warn(f"Failed to initialize OpenLineage emitter: {e}")
return None
def __repr__(self) -> str:
# Show lazy loading status without triggering initialization
registry_status = "not loaded" if self._registry is None else "loaded"
provider_status = "not loaded" if self._provider is None else "loaded"
return (
f"FeatureStore(\n"
f" repo_path={self.repo_path!r},\n"
f" config={self.config!r},\n"
f" registry={registry_status},\n"
f" provider={provider_status}\n"
f")"
)
@property
def registry(self) -> BaseRegistry:
"""Gets the registry of this feature store."""
if self._registry is None:
self._registry = self._create_registry()
# Add feature service cache to registry for performance optimization
if self._registry and not hasattr(self._registry, "_feature_service_cache"):
setattr(
self._registry,
"_feature_service_cache",
self._feature_service_cache,
)
if self._registry is None:
raise RuntimeError("Registry failed to initialize properly")
return self._registry
def _create_registry(self) -> BaseRegistry:
"""Create and initialize the registry."""
registry_config = self.config.registry
if registry_config.registry_type == "sql":
return SqlRegistry(registry_config, self.config.project, None)
elif registry_config.registry_type == "snowflake.registry":
from feast.infra.registry.snowflake import SnowflakeRegistry
return SnowflakeRegistry(registry_config, self.config.project, None)
elif registry_config and registry_config.registry_type == "remote":
from feast.infra.registry.remote import RemoteRegistry
return RemoteRegistry(
registry_config, self.config.project, None, self.config.auth_config
)
else:
return Registry(
self.config.project,
registry_config,
repo_path=self.repo_path,
auth_config=self.config.auth_config,
)
@property
def project(self) -> str:
"""Gets the project of this feature store."""
return self.config.project
@property
def provider(self) -> Provider:
"""Gets the provider of this feature store."""
if self._provider is None:
self._provider = get_provider(self.config)
return self._provider
def _get_provider(self) -> Provider:
# TODO: Bake self.repo_path into self.config so that we dont only have one interface to paths
return self.provider
@property
def openlineage_emitter(self) -> Optional[Any]:
"""Gets the OpenLineage emitter of this feature store."""
if self._openlineage_emitter is None:
self._openlineage_emitter = self._init_openlineage_emitter()
return self._openlineage_emitter
def _clear_feature_service_cache(self):
"""Clear feature service cache to avoid stale data after registry refresh."""
self._feature_service_cache.clear()
if hasattr(self.registry, "_feature_service_cache"):
getattr(self.registry, "_feature_service_cache").clear()
self._fs_name_cache.clear()
self._fs_name_index.clear()
self._fs_name_index_ts = -self._FS_NAME_INDEX_TTL_SECONDS
def refresh_registry(self):
"""Fetches and caches a copy of the feature registry in memory.
Explicitly calling this method allows for direct control of the state of the registry cache. Every time this
method is called the complete registry state will be retrieved from the remote registry store backend
(e.g., GCS, S3), and the cache timer will be reset. If refresh_registry() is run before get_online_features()
is called, then get_online_features() will use the cached registry instead of retrieving (and caching) the
registry itself.
Additionally, the TTL for the registry cache can be set to infinity (by setting it to 0), which means that
refresh_registry() will become the only way to update the cached registry. If the TTL is set to a value
greater than 0, then once the cache becomes stale (more time than the TTL has passed), a new cache will be
downloaded synchronously, which may increase latencies if the triggering method is get_online_features().
"""
self.registry.refresh(self.project)
self._clear_feature_service_cache()
def list_entities(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[Entity]:
"""
Retrieves the list of entities from the registry.
Args:
allow_cache: Whether to allow returning entities from a cached registry.
tags: Filter by tags.
Returns:
A list of entities.
"""
return self._list_entities(allow_cache, tags=tags)
def _list_entities(
self,
allow_cache: bool = False,
hide_dummy_entity: bool = True,
tags: Optional[dict[str, str]] = None,
) -> List[Entity]:
all_entities = self.registry.list_entities(
self.project, allow_cache=allow_cache, tags=tags
)
return [
entity
for entity in all_entities
if entity.name != DUMMY_ENTITY_NAME or not hide_dummy_entity
]
def list_feature_services(
self, tags: Optional[dict[str, str]] = None
) -> List[FeatureService]:
"""
Retrieves the list of feature services from the registry.
Args:
tags: Filter by tags.
Returns:
A list of feature services.
"""
return self.registry.list_feature_services(self.project, tags=tags)
def _list_all_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[BaseFeatureView]:
feature_views = []
for fv in self.registry.list_all_feature_views(
self.project, allow_cache=allow_cache, tags=tags
):
if (
isinstance(fv, FeatureView)
and fv.entities
and fv.entities[0] == DUMMY_ENTITY_NAME
):
fv.entities = []
fv.entity_columns = []
feature_views.append(fv)
return feature_views
def list_all_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[BaseFeatureView]:
"""
Retrieves the list of feature views from the registry.
Args:
allow_cache: Whether to allow returning entities from a cached registry.
Returns:
A list of feature views.
"""
return self._list_all_feature_views(allow_cache, tags=tags)
def list_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[FeatureView]:
"""
Retrieves the list of feature views from the registry.
Args:
allow_cache: Whether to allow returning entities from a cached registry.
tags: Filter by tags.
Returns:
A list of feature views.
"""
return utils._list_feature_views(
self.registry, self.project, allow_cache, tags=tags
)
def list_batch_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[FeatureView]:
"""
Retrieves the list of feature views from the registry.
Args:
allow_cache: Whether to allow returning entities from a cached registry.
tags: Filter by tags.
Returns:
A list of feature views.
"""
return self._list_batch_feature_views(allow_cache=allow_cache, tags=tags)
def _list_batch_feature_views(
self,
allow_cache: bool = False,
hide_dummy_entity: bool = True,
tags: Optional[dict[str, str]] = None,
) -> List[FeatureView]:
feature_views = []
for fv in self.registry.list_feature_views(
self.project, allow_cache=allow_cache, tags=tags
):
if (
hide_dummy_entity
and fv.entities
and fv.entities[0] == DUMMY_ENTITY_NAME
):
fv.entities = []
fv.entity_columns = []
feature_views.append(fv)
return feature_views
def _list_stream_feature_views(
self,
allow_cache: bool = False,
hide_dummy_entity: bool = True,
tags: Optional[dict[str, str]] = None,
) -> List[StreamFeatureView]:
stream_feature_views = []
for sfv in self.registry.list_stream_feature_views(
self.project, allow_cache=allow_cache, tags=tags
):
if hide_dummy_entity and sfv.entities[0] == DUMMY_ENTITY_NAME:
sfv.entities = []
sfv.entity_columns = []
stream_feature_views.append(sfv)
return stream_feature_views
def list_on_demand_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[OnDemandFeatureView]:
"""
Retrieves the list of on demand feature views from the registry.
Args:
allow_cache: Whether to allow returning entities from a cached registry.
tags: Filter by tags.
Returns:
A list of on demand feature views.
"""
return self.registry.list_on_demand_feature_views(
self.project, allow_cache=allow_cache, tags=tags
)
def list_stream_feature_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[StreamFeatureView]:
"""
Retrieves the list of stream feature views from the registry.
Returns:
A list of stream feature views.
"""
return self._list_stream_feature_views(allow_cache, tags=tags)
def list_label_views(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[LabelView]:
"""
Retrieves the list of label views from the registry.
Args:
allow_cache: Whether to allow returning label views from a cached registry.
tags: Filter by tags.
Returns:
A list of label views.
"""
return self.registry.list_label_views(
self.project, allow_cache=allow_cache, tags=tags
)
def get_label_view(
self, name: str, allow_registry_cache: bool = False
) -> LabelView:
"""
Retrieves a label view by name.
Args:
name: Name of the label view.
allow_registry_cache: Whether to allow returning the label view from a cached registry.
Returns:
The specified label view.
Raises:
FeatureViewNotFoundException: The label view could not be found.
"""
return self.registry.get_label_view(
name, self.project, allow_cache=allow_registry_cache
)
def list_data_sources(
self, allow_cache: bool = False, tags: Optional[dict[str, str]] = None
) -> List[DataSource]:
"""
Retrieves the list of data sources from the registry.
Args:
allow_cache: Whether to allow returning data sources from a cached registry.
tags: Filter by tags.
Returns:
A list of data sources.
"""
return self.registry.list_data_sources(
self.project, allow_cache=allow_cache, tags=tags
)
def get_entity(self, name: str, allow_registry_cache: bool = False) -> Entity:
"""
Retrieves an entity.
Args:
name: Name of entity.
allow_registry_cache: (Optional) Whether to allow returning this entity from a cached registry
Returns:
The specified entity.
Raises:
EntityNotFoundException: The entity could not be found.
"""
return self.registry.get_entity(
name, self.project, allow_cache=allow_registry_cache
)
def get_feature_service(
self, name: str, allow_cache: bool = False
) -> FeatureService:
"""
Retrieves a feature service.
Args:
name: Name of feature service.
allow_cache: Whether to allow returning feature services from a cached registry.
Returns:
The specified feature service.
Raises:
FeatureServiceNotFoundException: The feature service could not be found.
"""
return self.registry.get_feature_service(name, self.project, allow_cache)
def get_feature_view(
self, name: str, allow_registry_cache: bool = False
) -> FeatureView:
"""
Retrieves a feature view.
Args:
name: Name of feature view.
allow_registry_cache: (Optional) Whether to allow returning this entity from a cached registry
Returns:
The specified feature view.
Raises:
FeatureViewNotFoundException: The feature view could not be found.
"""
return self._get_feature_view(name, allow_registry_cache=allow_registry_cache)
def _get_feature_view(
self,
name: str,
hide_dummy_entity: bool = True,
allow_registry_cache: bool = False,
) -> FeatureView:
feature_view = self.registry.get_feature_view(
name, self.project, allow_cache=allow_registry_cache
)
if hide_dummy_entity and feature_view.entities[0] == DUMMY_ENTITY_NAME:
feature_view.entities = []
return feature_view
def list_feature_view_versions(self, name: str) -> List[Dict[str, Any]]:
"""
List version history for a feature view.
Args:
name: Name of feature view.
Returns:
List of version records.
"""
return self.registry.list_feature_view_versions(name, self.project)
def get_stream_feature_view(
self, name: str, allow_registry_cache: bool = False
) -> StreamFeatureView:
"""
Retrieves a stream feature view.
Args:
name: Name of stream feature view.
allow_registry_cache: (Optional) Whether to allow returning this entity from a cached registry
Returns:
The specified stream feature view.
Raises:
FeatureViewNotFoundException: The feature view could not be found.
"""
return self._get_stream_feature_view(
name, allow_registry_cache=allow_registry_cache
)
def _get_stream_feature_view(
self,
name: str,
hide_dummy_entity: bool = True,
allow_registry_cache: bool = False,
) -> StreamFeatureView:
stream_feature_view = self.registry.get_stream_feature_view(
name, self.project, allow_cache=allow_registry_cache
)
if hide_dummy_entity and stream_feature_view.entities[0] == DUMMY_ENTITY_NAME:
stream_feature_view.entities = []
return stream_feature_view
def get_on_demand_feature_view(
self, name: str, allow_registry_cache: bool = False
) -> OnDemandFeatureView:
"""
Retrieves a feature view.
Args:
name: Name of feature view.
allow_registry_cache: (Optional) Whether to allow returning this entity from a cached registry
Returns:
The specified feature view.
Raises:
FeatureViewNotFoundException: The feature view could not be found.
"""
return self.registry.get_on_demand_feature_view(
name, self.project, allow_cache=allow_registry_cache
)
def get_data_source(self, name: str) -> DataSource:
"""
Retrieves the list of data sources from the registry.
Args:
name: Name of the data source.
Returns:
The specified data source.
Raises:
DataSourceObjectNotFoundException: The data source could not be found.
"""
return self.registry.get_data_source(name, self.project)
def delete_feature_view(self, name: str):
"""
Deletes a feature view of any kind (FeatureView, OnDemandFeatureView, StreamFeatureView).
Args:
name: Name of feature view.
Raises:
FeatureViewNotFoundException: The feature view could not be found.
"""
return self.registry.delete_feature_view(name, self.project)
def enable_feature_view(self, name: str):
"""
Enable a feature view for serving and materialization.
Args:
name: Name of feature view.
"""
fv = self.registry.get_any_feature_view(name, self.project)
fv.enabled = True # type: ignore[attr-defined]
self.registry.apply_feature_view(fv, self.project)
def disable_feature_view(self, name: str):
"""
Disable a feature view to prevent serving and materialization.
Args:
name: Name of feature view.
"""
fv = self.registry.get_any_feature_view(name, self.project)
fv.enabled = False # type: ignore[attr-defined]
self.registry.apply_feature_view(fv, self.project)
def set_feature_view_state(self, name: str, state: FeatureViewState):
"""
Set the lifecycle state of a feature view.
Args:
name: Name of feature view.
state: Target state.
"""
fv = self.registry.get_any_feature_view(name, self.project)
if not fv.state.can_transition_to(state): # type: ignore[attr-defined]
raise ValueError(
f"Invalid state transition: {fv.state.name} -> {state.name}." # type: ignore[attr-defined]
)
fv.state = state # type: ignore[attr-defined]
self.registry.apply_feature_view(fv, self.project)
def delete_feature_service(self, name: str):
"""
Deletes a feature service.
Args:
name: Name of feature service.
Raises:
FeatureServiceNotFoundException: The feature view could not be found.
"""
return self.registry.delete_feature_service(name, self.project)
def _should_use_plan(self):
"""Returns True if plan and _apply_diffs should be used, False otherwise."""
# Currently only the local provider with sqlite online store supports plan and _apply_diffs.
return self.config.provider == "local" and (
self.config.online_store and self.config.online_store.type == "sqlite"
)
def _validate_all_feature_views(
self,
views_to_update: List[FeatureView],
odfvs_to_update: List[OnDemandFeatureView],
sfvs_to_update: List[StreamFeatureView],
lvs_to_update: Optional[List[LabelView]] = None,
):
"""Validates all feature views."""
if len(odfvs_to_update) > 0 and not flags_helper.is_test():
warnings.warn(
"On demand feature view is an experimental feature. "
"This API is stable, but the functionality does not scale well for offline retrieval",
RuntimeWarning,
)
_validate_feature_views(
[
*views_to_update,
*odfvs_to_update,
*sfvs_to_update,
*(lvs_to_update or []),
]
)
def _make_inferences(
self,
data_sources_to_update: List[DataSource],
entities_to_update: List[Entity],
views_to_update: List[FeatureView],
odfvs_to_update: List[OnDemandFeatureView],
sfvs_to_update: List[StreamFeatureView],
feature_services_to_update: List[FeatureService],
lvs_to_update: Optional[List[LabelView]] = None,
):
"""Makes inferences for entities, feature views, odfvs, and feature services."""
lvs_to_update = lvs_to_update or []
update_data_sources_with_inferred_event_timestamp_col(
data_sources_to_update, self.config
)
update_data_sources_with_inferred_event_timestamp_col(
[
view.batch_source
for view in views_to_update
if view.batch_source is not None
],
self.config,
)
update_data_sources_with_inferred_event_timestamp_col(
[
view.batch_source
for view in sfvs_to_update
if view.batch_source is not None
],
self.config,
)
update_data_sources_with_inferred_event_timestamp_col(
[lv.batch_source for lv in lvs_to_update if lv.batch_source is not None],
self.config,
)
# New feature views may reference previously applied entities.
entities = self._list_entities()
provider = self._get_provider()
update_feature_views_with_inferred_features_and_entities(
provider,
views_to_update,
entities + entities_to_update,
self.config,
)
update_feature_views_with_inferred_features_and_entities(
provider,
sfvs_to_update,
entities + entities_to_update,
self.config,
)
# We need to attach the time stamp fields to the underlying data sources
# and cascade the dependencies
update_feature_views_with_inferred_features_and_entities(
provider, odfvs_to_update, entities + entities_to_update, self.config
)
# TODO(kevjumba): Update schema inference
for sfv in sfvs_to_update:
if not sfv.schema:
raise ValueError(
f"schema inference not yet supported for stream feature views. please define schema for stream feature view: {sfv.name}"
)
for odfv in odfvs_to_update:
odfv.infer_features()
odfvs_to_write = [
odfv for odfv in odfvs_to_update if odfv.write_to_online_store
]
fvs_to_update_map = {
view.name: view
for view in [
*views_to_update,