-
Notifications
You must be signed in to change notification settings - Fork 1
/
__init__.py
598 lines (487 loc) · 19.8 KB
/
__init__.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
import fiftyone as fo
import fiftyone.operators as foo
from fiftyone.operators import types
from fiftyone.brain import Similarity
import time
import requests
import glob
from pprint import pprint
import os
class SemanticFrameSearch(foo.Operator):
@property
def config(self):
return foo.OperatorConfig(
name="semantic_frames_search",
label="semantic frames search",
description="semantic video search using only frames as context",
dynamic=True,
icon="/assets/search.svg",
)
def resolve_input(self, ctx):
inputs = types.Object()
get_brain_key(ctx, inputs)
sort_choices = ["Sort Frames", "Sort Videos"] # replace with your choices
sort_group = types.RadioGroup()
for choice in sort_choices:
sort_group.add_choice(choice, label=choice)
inputs.enum(
"sort_group",
sort_group.values(),
label="Select which to sort",
description="Show individual frames or whole videos",
view=types.RadioView(),
default='Sort Videos',
required=True,
)
inputs.str("prompt", label="Prompt", required=True)
return types.Property(inputs)
def execute(self, ctx):
frames = ctx.dataset.to_frames(sample_frames=True)
sort_group = ctx.params.get("sort_group")
prompt = ctx.params.get("prompt")
brain_key = ctx.params.get("brain_key")
view = frames.sort_by_similarity(prompt, brain_key=brain_key)
if sort_group == "Sort Frames":
ctx.trigger("set_view", {"view": view._serialize()})
else:
sample_ids = view.distinct("sample_id")
sorted_list = []
for sample in view:
if sample.sample_id not in sorted_list:
sorted_list.append(sample.sample_id)
ctx.trigger("set_view", {"view": ctx.dataset.select(sorted_list,ordered=True)._serialize()})
return {}
class SemanticVideoBackend(foo.Operator):
@property
def config(self):
return foo.OperatorConfig(
name="create_semantic_video_index",
label="create semantic video index",
description="create a semantic video index backend with Twelve Labs",
dynamic=True,
icon="/assets/search.svg",
)
def resolve_input(self, ctx):
inputs = types.Object()
API_URL = ctx.secret("TWELVE_API_URL")
API_KEY = ctx.secret("TWELVE_API_KEY")
#API_URL = os.getenv("TWELVE_API_URL")
#API_KEY = os.getenv("TWELVE_API_KEY")
if API_URL is None or API_KEY is None:
inputs.view(
"warning",
types.Warning(label="Twelve Lab keys undefined",
description="Please define the enviroment variables TWELVE_API_KEY and TWELVE_API_URL and reload")
)
target_view = get_target_view(ctx, inputs)
inputs.message(
"Notice",
label="Create Semantic Video Index",
description="Choose your index name and the modalities of embeddings you will use!"
)
inputs.str("index_name", label="Index_Name", required=True)
inputs.view(
"header",
types.Header(label="Select modalities for index", description="Select one or more from the below to extract embeddings from your videos", divider=True)
)
inputs.bool(
"visual",
label="visual",
description="",
view=types.CheckboxView(),
)
inputs.bool(
"conversation",
label="conversation",
description="Video must have audio to work!",
view=types.CheckboxView(),
)
inputs.bool(
"text_in_video",
label="text_in_video",
description="",
view=types.CheckboxView(),
)
inputs.bool(
"logo",
label="logo",
description="",
view=types.CheckboxView(),
)
inputs.view(
"header2",
types.Header(label="Advise delegating to avoid timeout", description="", divider=True)
)
_execution_mode(ctx, inputs)
return types.Property(inputs)
def resolve_delegation(self, ctx):
return ctx.params.get("delegate", False)
def execute(self, ctx):
ctx.dataset.compute_metadata()
target = ctx.params.get("target", None)
target_view = _get_target_view(ctx, target)
API_URL = ctx.secret("TWELVE_API_URL")
API_KEY = ctx.secret("TWELVE_API_KEY")
INDEX_NAME = ctx.params.get("index_name")
INDEXES_URL = f"{API_URL}/indexes"
headers = {
"x-api-key": API_KEY
}
so = []
if ctx.params.get("visual"):
so.append("visual")
if ctx.params.get("logo"):
so.append("logo")
if ctx.params.get("text_in_video"):
so.append("text_in_video")
if ctx.params.get("conversation"):
so.append("conversation")
data = {
"engine_id": "marengo2.5",
"index_options": so,
"index_name": INDEX_NAME,
}
response = requests.post(INDEXES_URL, headers=headers, json=data)
headers = {
"x-api-key": API_KEY,
"Content-Type": "application/json"
}
INDEX_ID = get_twelve_id_from_name(INDEXES_URL, headers, INDEX_NAME)
TASKS_URL = f"{API_URL}/tasks"
videos = target_view
for sample in videos:
if sample.metadata.duration < 4:
continue
else:
file_name = sample.filepath.split("/")[-1]
file_path = sample.filepath
file_stream = open(file_path,"rb")
headers = {
"x-api-key": API_KEY
}
data = {
"index_id": INDEX_ID,
"language": "en"
}
file_param=[
("video_file", (file_name, file_stream, "application/octet-stream")),]
response = requests.post(TASKS_URL, headers=headers, data=data, files=file_param)
TASK_ID = response.json().get("_id")
print (f"Status code: {response.status_code}")
pprint (response.json())
TASK_STATUS_URL = f"{API_URL}/tasks/{TASK_ID}"
while True:
response = requests.get(TASK_STATUS_URL, headers=headers)
STATUS = response.json().get("status")
if STATUS == "ready":
break
time.sleep(10)
VIDEO_ID = response.json().get('video_id')
sample["Twelve Labs " + INDEX_NAME] = VIDEO_ID
sample.save()
return {}
class SemanticVideoSearch(foo.Operator):
@property
def config(self):
return foo.OperatorConfig(
name="semantic_video_search",
label="semantic video search",
description="semantic video search using videos as context",
dynamic=True,
icon="/assets/search.svg",
)
def resolve_input(self, ctx):
inputs = types.Object()
API_URL = ctx.secret("TWELVE_API_URL")
API_KEY = ctx.secret("TWELVE_API_KEY")
#API_URL = os.getenv("TWELVE_API_URL")
#API_KEY = os.getenv("TWELVE_API_KEY")
if API_URL is None or API_KEY is None:
inputs.view(
"warning",
types.Warning(label="Twelve Lab keys undefined",
description="Please define the enviroment variables TWELVE_API_KEY and TWELVE_API_URL and reload")
)
else:
target_view = get_target_view(ctx, inputs)
INDEXES_URL = f"{API_URL}/indexes"
headers = {
"x-api-key": API_KEY,
"Content-Type": "application/json"
}
response = requests.get(INDEXES_URL, headers=headers)
if response.json()["data"] == []:
inputs.view(
"No Index",
types.Warning(label="No Indexes detected",
description="Please run `create semantic video index` first in order to semantic search on your dataset!")
)
else:
vis_flag = False
logo_flag = False
convo_flag = False
text_flag = False
index_info = {}
for item in response.json()['data']:
if "visual" in item["index_options"]:
vis_flag = True
if "logo" in item["index_options"]:
logo_flag = True
if "text_in_video" in item["index_options"]:
text_flag = True
if "conversation" in item["index_options"]:
convo_flag = True
index_info[item['index_name']] = item['_id']
choices = index_info.keys()
choices_compare = [x[12:] for x in ctx.dataset.get_field_schema().keys()] #change if ever add more than Twelve Labs
common_index = list(set(choices_compare) & set(choices))
if len(common_index) < 1:
inputs.view(
"warning2",
types.Warning(label="Twelve Lab Video ID Missing From Sample",
description="Samples need to have a Twelve Lab Video ID associated with them.\
They are found in a field called Twelve Labs (index name). If this is missing from your dataset,\
make sure your dataset is persisent and to avoid losing between runs")
)
else:
radio_group = types.RadioGroup()
for choice in common_index:
radio_group.add_choice(choice, label=choice)
inputs.message(
"Notice",
label="Semantic Video Search",
description="Search through your video dataset with a prompt. If you haven't yet \
generated a similarity index with Twelve Labs, run the creaet semantic video index \
operator first! Note, you can only search on modalities within your chosen index!"
)
if len(choices) != len(common_index):
inputs.view(
"warning3",
types.Warning(label="Only showing indexes that have Video IDs on the dataset.",
description="To add video IDs, run create_semantic_video_index to regenerate the index. \
It will store the Video IDs in a field called Twelve Labs (index_name). If this is missing from your dataset,\
make sure your dataset is persisent and to avoid losing between runs")
)
inputs.enum(
"index_name",
radio_group.values(),
label="Pick an index",
description="",
view=types.DropdownView(),
required=True,
)
inputs.str("prompt", label="Prompt", required=True)
inputs.view(
"header",
types.Header(label="Select modalities for search", description="Select one or more from the below to search through your videos. Note: your index must have this modality!", divider=True)
)
if vis_flag:
inputs.bool(
"visual",
label="visual",
description="",
view=types.CheckboxView(),
)
if convo_flag:
inputs.bool(
"conversation",
label="conversation",
description="Video must have audio to work!",
view=types.CheckboxView(),
)
if text_flag:
inputs.bool(
"text_in_video",
label="text_in_video",
description="",
view=types.CheckboxView(),
)
if logo_flag:
inputs.bool(
"logo",
label="logo",
description="",
view=types.CheckboxView(),
)
_execution_mode(ctx, inputs)
return types.Property(inputs)
def resolve_delegation(self, ctx):
return ctx.params.get("delegate", False)
def execute(self, ctx):
API_URL = ctx.secret("TWELVE_API_URL")
API_KEY = ctx.secret("TWELVE_API_KEY")
assert API_KEY, "Env variable TWELVE_API_KEY not defined."
assert API_URL, "Env variable TWELVE_API_URL not defined."
target = ctx.params.get("target", None)
target_view = _get_target_view(ctx, target)
INDEX_NAME = ctx.params.get("index_name")
INDEXES_URL = f"{API_URL}/indexes"
headers = {
"x-api-key": API_KEY,
"Content-Type": "application/json"
}
INDEX_ID = get_twelve_id_from_name(INDEXES_URL, headers, INDEX_NAME)
prompt = ctx.params.get("prompt")
SEARCH_URL = f"{API_URL}/search"
headers = {
"x-api-key": API_KEY
}
so = []
if ctx.params.get("visual"):
so.append("visual")
if ctx.params.get("logo"):
so.append("logo")
if ctx.params.get("text_in_video"):
so.append("text_in_video")
if ctx.params.get("conversation"):
so.append("conversation")
data = {
"query": prompt,
"index_id": INDEX_ID,
"search_options": so,
}
response = requests.post(SEARCH_URL, headers=headers, json=data)
video_ids = [entry['video_id'] for entry in response.json()['data']]
print(response.json())
samples = []
view1 = target_view.select_by("Twelve Labs " + INDEX_NAME, video_ids,ordered=True)
start = [entry['start'] for entry in response.json()['data']]
end = [entry['end'] for entry in response.json()['data']]
if "results" in ctx.dataset.get_field_schema().keys():
ctx.dataset.delete_sample_field("results")
i=0
for sample in view1:
support = [int(start[i]*sample.metadata.frame_rate)+1 ,int(end[i]*sample.metadata.frame_rate)+1]
sample["results"] = fo.TemporalDetection(label=prompt, support=tuple(support))
sample.save()
view2 = view1.to_clips("results")
ctx.trigger("set_view", {"view": view2._serialize()})
return {}
def get_target_view(ctx, inputs):
has_view = ctx.view != ctx.dataset.view()
has_selected = bool(ctx.selected)
default_target = None
if has_view or has_selected:
target_choices = types.RadioGroup(orientation="horizontal")
target_choices.add_choice(
"DATASET",
label="Entire dataset",
description="Process the entire dataset",
)
if has_view:
target_choices.add_choice(
"CURRENT_VIEW",
label="Current view",
description="Process the current view",
)
default_target = "CURRENT_VIEW"
if has_selected:
target_choices.add_choice(
"SELECTED_SAMPLES",
label="Selected samples",
description="Process only the selected samples",
)
default_target = "SELECTED_SAMPLES"
inputs.enum(
"target",
target_choices.values(),
default=default_target,
required=True,
label="Target view",
view=target_choices,
)
target = ctx.params.get("target", default_target)
return _get_target_view(ctx, target)
def _get_target_view(ctx, target):
if target == "SELECTED_SAMPLES":
return ctx.view.select(ctx.selected)
if target == "DATASET":
return ctx.dataset
return ctx.view
def get_twelve_id_from_name(INDEXES_URL, headers, INDEX_NAME):
response = requests.get(INDEXES_URL, headers=headers)
INDEX_ID = None
for index in response.json()["data"]:
if index["index_name"] == INDEX_NAME:
INDEX_ID = index["_id"]
return INDEX_ID
_BRAIN_RUN_TYPES = {
"similarity": Similarity,
}
def get_brain_key(
ctx,
inputs,
label="Brain key",
description="Select a brain key",
run_type="similarity",
show_default=True,
):
type = _BRAIN_RUN_TYPES.get(run_type, None)
brain_keys = ctx.dataset.list_brain_runs(type=type)
if not brain_keys:
message = "This dataset has no text similarity brain runs"
warning = types.Warning(
label=message,
description="https://docs.voxel51.com/user_guide/brain.html",
)
prop = inputs.view("warning", warning)
prop.invalid = True
return
text_brain_keys = []
for brain_key in brain_keys:
info = ctx.dataset.get_brain_info(brain_key)
if info.config.supports_prompts:
text_brain_keys.append(brain_key)
if len(text_brain_keys) < 1:
message = "This dataset has no text similarity brain runs"
warning = types.Warning(
label=message,
description="https://docs.voxel51.com/user_guide/brain.html",
)
prop = inputs.view("warning", warning)
prop.invalid = True
return
choices = types.DropdownView()
for brain_key in brain_keys:
choices.add_choice(brain_key, label=brain_key)
default = brain_keys[0] if show_default else None
inputs.str(
"brain_key",
default=default,
required=True,
label=label,
description=description,
view=choices,
)
return ctx.params.get("brain_key", None)
def _execution_mode(ctx, inputs):
delegate = ctx.params.get("delegate", False)
if delegate:
description = "Uncheck this box to execute the operation immediately"
else:
description = "Check this box to delegate execution of this task"
inputs.bool(
"delegate",
default=False,
required=True,
label="Delegate execution?",
description=description,
view=types.CheckboxView(),
)
if delegate:
inputs.view(
"notice",
types.Notice(
label=(
"You've chosen delegated execution. Note that you must "
"have a delegated operation service running in order for "
"this task to be processed. See "
"https://docs.voxel51.com/plugins/index.html#operators "
"for more information. Also, dont forget to set enviroment variables in the delegated enviroment as well!"
)
),
)
def register(plugin):
plugin.register(SemanticFrameSearch)
plugin.register(SemanticVideoSearch)
plugin.register(SemanticVideoBackend)