-
Notifications
You must be signed in to change notification settings - Fork 511
/
query.py
691 lines (570 loc) · 19.2 KB
/
query.py
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
"""
FiftyOne Server queries.
| Copyright 2017-2024, Voxel51, Inc.
| `voxel51.com <https://voxel51.com/>`_
|
"""
from dataclasses import asdict
from datetime import date, datetime
from enum import Enum
import logging
import os
import typing as t
import eta.core.serial as etas
import eta.core.utils as etau
import strawberry as gql
from bson import ObjectId, json_util
import fiftyone as fo
import fiftyone.brain as fob # pylint: disable=import-error,no-name-in-module
import fiftyone.constants as foc
import fiftyone.core.context as focx
import fiftyone.core.dataset as fod
import fiftyone.core.media as fom
from fiftyone.core.odm import SavedViewDocument
import fiftyone.core.stages as fosg
from fiftyone.core.state import SampleField, serialize_fields
import fiftyone.core.uid as fou
from fiftyone.core.utils import run_sync_task
import fiftyone.core.view as fov
import fiftyone.server.aggregate as fosa
from fiftyone.server.aggregations import aggregate_resolver
from fiftyone.server.color import ColorBy, ColorScheme
from fiftyone.server.data import Info
from fiftyone.server.dataloader import get_dataloader_resolver
from fiftyone.server.indexes import Index, from_dict as indexes_from_dict
from fiftyone.server.lightning import lightning_resolver
from fiftyone.server.metadata import MediaType
from fiftyone.server.paginator import Connection, get_paginator_resolver
from fiftyone.server.samples import (
SampleFilter,
SampleItem,
paginate_samples,
)
from fiftyone.server.scalars import BSON, BSONArray, JSON
from fiftyone.server.stage_definitions import stage_definitions
from fiftyone.server.utils import from_dict
ID = gql.scalar(
t.NewType("ID", str),
serialize=lambda v: str(v),
parse_value=lambda v: ObjectId(v),
)
DATASET_FILTER = [{"sample_collection_name": {"$regex": "^samples\\."}}]
DATASET_FILTER_STAGE = [{"$match": DATASET_FILTER[0]}]
@gql.type
class Group:
name: str
media_type: MediaType
@gql.type
class Target:
target: str
value: str
@gql.type
class NamedTargets:
name: str
targets: t.List[Target]
@gql.interface
class RunConfig:
cls: str
@gql.interface
class Run:
key: str
version: t.Optional[str]
timestamp: t.Optional[datetime]
config: t.Optional[RunConfig]
view_stages: t.Optional[t.List[str]]
@gql.enum
class BrainRunType(Enum):
similarity = "similarity"
visualization = "visualization"
@gql.type
class BrainRunConfig(RunConfig):
embeddings_field: t.Optional[str]
method: t.Optional[str]
patches_field: t.Optional[str]
supports_prompts: t.Optional[bool]
@gql.field
def type(self) -> t.Optional[BrainRunType]:
try:
if issubclass(fob.SimilarityConfig, etau.get_class(self.cls)):
return BrainRunType.similarity
if issubclass(fob.VisualizationConfig, etau.get_class(self.cls)):
return BrainRunType.visualization
except:
pass
return None
@gql.field
def max_k(self) -> t.Optional[int]:
config = self._create_config()
return getattr(config, "max_k", None)
@gql.field
def supports_least_similarity(self) -> t.Optional[bool]:
config = self._create_config()
return getattr(config, "supports_least_similarity", None)
def _create_config(self):
try:
cls = etau.get_class(self.cls)
return cls(
embeddings_field=self.embeddings_field,
patches_field=self.patches_field,
)
except:
return None
@gql.type
class BrainRun(Run):
config: t.Optional[BrainRunConfig]
@gql.type
class EvaluationRunConfig(RunConfig):
gt_field: t.Optional[str]
pred_field: t.Optional[str]
method: t.Optional[str]
@gql.type
class EvaluationRun(Run):
config: t.Optional[EvaluationRunConfig]
@gql.type
class SavedView:
id: t.Optional[str]
dataset_id: t.Optional[str]
name: t.Optional[str]
description: t.Optional[str]
color: t.Optional[str]
slug: t.Optional[str]
view_stages: t.Optional[t.List[str]]
created_at: t.Optional[datetime]
last_modified_at: t.Optional[datetime]
last_loaded_at: t.Optional[datetime]
@gql.field
def view_name(self) -> t.Optional[str]:
if isinstance(self, ObjectId):
return None
return self.name
@gql.field
def stage_dicts(self) -> t.Optional[BSONArray]:
return [json_util.loads(x) for x in self.view_stages]
@classmethod
def from_doc(cls, doc: SavedViewDocument):
stage_dicts = [json_util.loads(x) for x in doc.view_stages]
data = doc.to_dict()
data["id"] = str(data.pop("_id"))
data["dataset_id"] = str(data.pop("_dataset_id"))
saved_view = from_dict(data_class=cls, data=data)
saved_view.stage_dicts = stage_dicts
return saved_view
@gql.type
class SidebarGroup:
name: str
paths: t.Optional[t.List[str]]
expanded: t.Optional[bool] = None
@gql.type
class KeypointSkeleton:
labels: t.Optional[t.List[str]]
edges: t.List[t.List[int]]
@gql.type
class NamedKeypointSkeleton(KeypointSkeleton):
name: str
@gql.enum
class SidebarMode(Enum):
all = "all"
best = "best"
fast = "fast"
@gql.type
class DatasetAppConfig:
color_scheme: t.Optional[ColorScheme]
media_fields: t.Optional[t.List[str]]
plugins: t.Optional[JSON]
sidebar_groups: t.Optional[t.List[SidebarGroup]]
sidebar_mode: t.Optional[SidebarMode]
spaces: t.Optional[JSON]
grid_media_field: str = "filepath"
modal_media_field: str = "filepath"
@gql.type
class Dataset:
id: gql.ID
dataset_id: gql.ID
name: str
created_at: t.Optional[date]
last_loaded_at: t.Optional[datetime]
persistent: bool
group_media_types: t.Optional[t.List[Group]]
group_field: t.Optional[str]
default_group_slice: t.Optional[str]
media_type: t.Optional[MediaType]
parent_media_type: t.Optional[MediaType]
mask_targets: t.List[NamedTargets]
default_mask_targets: t.Optional[t.List[Target]]
sample_fields: t.List[SampleField]
frame_fields: t.Optional[t.List[SampleField]]
brain_methods: t.Optional[t.List[BrainRun]]
evaluations: t.Optional[t.List[EvaluationRun]]
saved_view_slug: t.Optional[str]
saved_views: t.Optional[t.List[SavedView]]
version: t.Optional[str]
view_cls: t.Optional[str]
view_name: t.Optional[str]
default_skeleton: t.Optional[KeypointSkeleton]
skeletons: t.List[NamedKeypointSkeleton]
app_config: t.Optional[DatasetAppConfig]
info: t.Optional[JSON]
estimated_frame_count: t.Optional[int]
estimated_sample_count: t.Optional[int]
frame_indexes: t.Optional[t.List[Index]]
sample_indexes: t.Optional[t.List[Index]]
frame_collection_name: gql.Private[t.Optional[str]]
sample_collection_name: gql.Private[t.Optional[str]]
@gql.field
def stages(
self, slug: t.Optional[str] = None, view: t.Optional[BSONArray] = None
) -> t.Optional[BSONArray]:
if slug:
for view in self.saved_views:
if view.slug == slug:
return view.stage_dicts()
return view or []
@gql.field
async def estimated_sample_count(self, info: Info = None) -> int:
return await info.context.db[
self.sample_collection_name
].estimated_document_count()
@gql.field
async def estimated_frame_count(
self, info: Info = None
) -> t.Optional[int]:
if self.frame_collection_name:
return await info.context.db[
self.frame_collection_name
].estimated_document_count()
@staticmethod
def modifier(doc: dict) -> dict:
doc["id"] = doc.pop("_id")
doc["dataset_id"] = doc["id"]
doc["default_mask_targets"] = _convert_targets(
doc.get("default_mask_targets", {})
)
doc["mask_targets"] = [
NamedTargets(name=name, targets=_convert_targets(targets))
for name, targets in doc.get("mask_targets", {}).items()
]
flat = _flatten_fields([], doc.get("sample_fields", []))
doc["sample_fields"] = flat
doc["frame_fields"] = _flatten_fields([], doc.get("frame_fields", []))
doc["brain_methods"] = list(doc.get("brain_methods", {}).values())
doc["evaluations"] = list(doc.get("evaluations", {}).values())
doc["saved_views"] = doc.get("saved_views", [])
doc["skeletons"] = list(
dict(name=name, **data)
for name, data in doc.get("skeletons", {}).items()
)
doc["group_media_types"] = [
Group(name=name, media_type=media_type)
for name, media_type in doc.get("group_media_types", {}).items()
]
doc["default_skeletons"] = doc.get("default_skeletons", None)
# gql private fields must always be present
doc.setdefault("frame_collection_name", None)
return doc
@classmethod
async def resolver(
cls,
name: str,
info: Info = None,
saved_view_slug: t.Optional[str] = gql.UNSET,
view: t.Optional[BSONArray] = None,
) -> t.Optional["Dataset"]:
return await serialize_dataset(
dataset_name=name,
serialized_view=view,
saved_view_slug=saved_view_slug,
dicts=False,
update_last_loaded_at=True,
)
dataset_dataloader = get_dataloader_resolver(
Dataset, "datasets", "name", DATASET_FILTER
)
@gql.enum
class Theme(Enum):
browser = "browser"
dark = "dark"
light = "light"
@gql.type
class AppConfig:
color_by: ColorBy
color_pool: t.List[str]
colorscale: str
grid_zoom: int
lightning_threshold: t.Optional[int]
loop_videos: bool
multicolor_keypoints: bool
notebook_height: int
plugins: t.Optional[JSON]
show_confidence: bool
show_index: bool
show_label: bool
show_skeletons: bool
show_tooltip: bool
sidebar_mode: SidebarMode
theme: Theme
timezone: t.Optional[str]
use_frame_number: bool
spaces: t.Optional[JSON]
@gql.type
class SchemaResult:
field_schema: t.List[SampleField]
frame_field_schema: t.List[SampleField]
@gql.type
class Query(fosa.AggregateQuery):
aggregations = gql.field(resolver=aggregate_resolver)
lightning = gql.field(resolver=lightning_resolver)
@gql.field
def colorscale(self) -> t.Optional[t.List[t.List[int]]]:
if fo.app_config.colorscale:
return fo.app_config.get_colormap()
return None
@gql.field
def config(self) -> AppConfig:
d = fo.app_config.serialize()
d["timezone"] = fo.config.timezone
return from_dict(AppConfig, d)
@gql.field
def context(self) -> str:
return focx._get_context()
@gql.field
def dev(self) -> bool:
return foc.DEV_INSTALL or foc.RC_INSTALL
@gql.field
def do_not_track(self) -> bool:
return fo.config.do_not_track
@gql.field
async def estimated_dataset_count(self, info: Info = None) -> int:
return await info.context.db.datasets.estimated_document_count()
dataset: Dataset = gql.field(resolver=Dataset.resolver)
datasets: Connection[Dataset, str] = gql.field(
resolver=get_paginator_resolver(
Dataset, "created_at", DATASET_FILTER_STAGE, "datasets"
)
)
@gql.field
async def samples(
self,
dataset: str,
view: BSONArray,
first: t.Optional[int] = 20,
after: t.Optional[str] = None,
filter: t.Optional[SampleFilter] = None,
filters: t.Optional[BSON] = None,
extended_stages: t.Optional[BSON] = None,
pagination_data: t.Optional[bool] = True,
) -> Connection[SampleItem, str]:
return await paginate_samples(
dataset,
view,
filters,
first,
after,
sample_filter=filter,
extended_stages=extended_stages,
pagination_data=pagination_data,
)
@gql.field
async def sample(
self,
dataset: str,
view: BSONArray,
filter: SampleFilter,
filters: t.Optional[JSON] = None,
) -> t.Optional[SampleItem]:
samples = await paginate_samples(
dataset,
view,
filters,
1,
sample_filter=filter,
pagination_data=False,
)
if samples.edges:
return samples.edges[0].node
return None
stage_definitions = gql.field(stage_definitions)
@gql.field
def uid(self) -> str:
return fou.get_user_id()
@gql.field
def version(self) -> str:
return foc.VERSION
@gql.field
def saved_views(self, dataset_name: str) -> t.Optional[t.List[SavedView]]:
try:
ds = fod.load_dataset(dataset_name)
return [
SavedView.from_doc(view_doc)
for view_doc in ds._doc.saved_views
]
except:
return None
@gql.field
def schema_for_view_stages(
self,
dataset_name: str,
view_stages: BSONArray,
) -> SchemaResult:
try:
ds = fod.load_dataset(dataset_name)
if view_stages:
view = fov.DatasetView._build(ds, view_stages or [])
if ds.media_type == fom.VIDEO:
frame_schema = serialize_fields(
view.get_frame_field_schema(flat=True)
)
field_schema = serialize_fields(
view.get_field_schema(flat=True)
)
return SchemaResult(
field_schema=field_schema,
frame_field_schema=frame_schema,
)
return SchemaResult(
field_schema=serialize_fields(
view.get_field_schema(flat=True)
),
frame_field_schema=[],
)
if ds.media_type == fom.VIDEO:
frames_field_schema = serialize_fields(
ds.get_frame_field_schema(flat=True)
)
field_schema = serialize_fields(ds.get_field_schema(flat=True))
return SchemaResult(
field_schema=field_schema,
frame_field_schema=frames_field_schema,
)
return SchemaResult(
field_schema=serialize_fields(ds.get_field_schema(flat=True)),
frame_field_schema=[],
)
except Exception as e:
return SchemaResult(
field_schema=[],
frame_field_schema=[],
)
def _flatten_fields(
path: t.List[str], fields: t.List[t.Dict]
) -> t.List[t.Dict]:
result = []
for field in fields:
key = field.pop("name", None)
if key is None:
# Issues with concurrency can cause this to happen.
# Until it's fixed, just ignore these fields to avoid throwing hard
# errors when loading in the app.
logging.debug("Skipping field with no name: %s", field)
continue
field_path = path + [key]
field["path"] = ".".join(field_path)
result.append(field)
fields = field.pop("fields", None)
if fields:
result = result + _flatten_fields(field_path, fields)
return result
def _convert_targets(targets: t.Dict[str, str]) -> t.List[Target]:
return [Target(target=k, value=v) for k, v in targets.items()]
async def serialize_dataset(
dataset_name: str,
serialized_view: BSONArray,
saved_view_slug: t.Optional[str] = None,
dicts=True,
update_last_loaded_at=False,
) -> Dataset:
def run():
if not fod.dataset_exists(dataset_name):
return None
dataset = fod.load_dataset(dataset_name)
if update_last_loaded_at:
dataset._update_last_loaded_at(force=True)
dataset.reload()
view_name = None
try:
doc = dataset._get_saved_view_doc(saved_view_slug, slug=True)
view = dataset.load_saved_view(doc.name)
view_name = view.name
if serialized_view:
for stage in serialized_view:
view = view.add_stage(fosg.ViewStage._from_dict(stage))
except:
view = fov.DatasetView._build(dataset, serialized_view or [])
doc = dataset._doc.to_dict(no_dereference=True)
Dataset.modifier(doc)
data = from_dict(Dataset, doc)
data.view_cls = None
data.view_name = view_name
data.saved_view_slug = saved_view_slug
collection = dataset.view()
if view is not None:
# unique id for for the relay global store
#
# until a schema is with respect to a view and not a dataset this
# is required
data.id = ObjectId()
if view._dataset != dataset:
d = view._dataset._serialize()
data.media_type = d["media_type"]
data.view_cls = etau.get_class_name(view)
data.parent_media_type = view._parent_media_type
data.media_type = view.media_type
collection = view
data.sample_fields = serialize_fields(
collection.get_field_schema(flat=True)
)
data.frame_fields = serialize_fields(
collection.get_frame_field_schema(flat=True)
)
if dicts:
saved_views = []
for view in data.saved_views:
view_dict = asdict(view)
view_dict["view_name"] = view.view_name()
view_dict["stage_dicts"] = view.stage_dicts()
saved_views.append(view_dict)
data.saved_views = saved_views
for brain_method in data.brain_methods:
try:
type = brain_method.config.type().value
except:
type = None
try:
max_k = brain_method.config.max_k()
except:
max_k = None
try:
supports_least_similarity = (
brain_method.config.supports_least_similarity()
)
except:
supports_least_similarity = None
setattr(brain_method.config, "type", type)
setattr(brain_method.config, "max_k", max_k)
setattr(
brain_method.config,
"supports_least_similarity",
supports_least_similarity,
)
_assign_estimated_counts(data, dataset)
_assign_lightning_info(data, dataset)
return data
return await run_sync_task(run)
def _assign_estimated_counts(dataset: Dataset, fo_dataset: fo.Dataset):
setattr(
dataset,
"estimated_sample_count",
fo_dataset._sample_collection.estimated_document_count(),
)
setattr(
dataset,
"estimated_frame_count",
(
fo_dataset._frame_collection.estimated_document_count()
if fo_dataset._frame_collection_name
else None
),
)
def _assign_lightning_info(dataset: Dataset, fo_dataset: fo.Dataset):
dataset.sample_indexes, dataset.frame_indexes = indexes_from_dict(
fo_dataset.get_index_information()
)