-
Notifications
You must be signed in to change notification settings - Fork 24
Expand file tree
/
Copy pathendpoints.py
More file actions
612 lines (505 loc) · 20.4 KB
/
endpoints.py
File metadata and controls
612 lines (505 loc) · 20.4 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
from __future__ import annotations
import warnings
from typing import Dict, List, Literal, Optional, Union
from together.abstract import api_requestor
from together.together_response import TogetherResponse
from together.types import TogetherClient, TogetherRequest
from together.types.endpoints import DedicatedEndpoint, HardwareWithStatus, ListEndpoint
class Endpoints:
def __init__(self, client: TogetherClient) -> None:
self._client = client
def list(
self,
type: Optional[Literal["dedicated", "serverless"]] = None,
usage_type: Optional[Literal["on-demand", "reserved"]] = None,
mine: Optional[bool] = None,
) -> List[ListEndpoint]:
"""
List all endpoints, can be filtered by endpoint type and ownership.
Args:
type (str, optional): Filter endpoints by endpoint type ("dedicated" or "serverless"). Defaults to None.
usage_type (str, optional): Filter endpoints by usage type ("on-demand" or "reserved"). Defaults to None.
mine (bool, optional): If True, return only endpoints owned by the caller. Defaults to None.
Returns:
List[ListEndpoint]: List of endpoint objects
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
params: Dict[
str,
Union[
Literal["dedicated", "serverless"],
Literal["on-demand", "reserved"],
bool,
],
] = {}
if type is not None:
params["type"] = type
if usage_type is not None:
params["usage_type"] = usage_type
if mine is not None:
params["mine"] = mine
response, _, _ = requestor.request(
options=TogetherRequest(
method="GET",
url="endpoints",
params=params,
),
stream=False,
)
response.data = response.data["data"]
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, list)
return [ListEndpoint(**endpoint) for endpoint in response.data]
def create(
self,
*,
model: str,
hardware: str,
min_replicas: int,
max_replicas: int,
display_name: Optional[str] = None,
disable_prompt_cache: bool = True,
disable_speculative_decoding: bool = True,
state: Literal["STARTED", "STOPPED"] = "STARTED",
inactive_timeout: Optional[int] = None,
availability_zone: Optional[str] = None,
) -> DedicatedEndpoint:
"""
Create a new dedicated endpoint.
Args:
model (str): The model to deploy on this endpoint
hardware (str): The hardware configuration to use for this endpoint
min_replicas (int): The minimum number of replicas to maintain
max_replicas (int): The maximum number of replicas to scale up to
display_name (str, optional): A human-readable name for the endpoint
disable_prompt_cache (bool, optional): Whether to disable the prompt cache. Defaults to False.
disable_speculative_decoding (bool, optional): Whether to disable speculative decoding. Defaults to False.
state (str, optional): The desired state of the endpoint. Defaults to "STARTED".
inactive_timeout (int, optional): The number of minutes of inactivity after which the endpoint will be automatically stopped. Set to 0 to disable automatic timeout.
availability_zone (str, optional): Start endpoint in specified availability zone (e.g., us-central-4b).
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
if disable_prompt_cache:
warnings.warn(
"The 'disable_prompt_cache' parameter (CLI flag: '--no-prompt-cache') is deprecated and will be removed in a future version.",
stacklevel=2,
)
requestor = api_requestor.APIRequestor(
client=self._client,
)
data: Dict[str, Union[str, bool, Dict[str, int], int]] = {
"model": model,
"hardware": hardware,
"autoscaling": {
"min_replicas": min_replicas,
"max_replicas": max_replicas,
},
"disable_prompt_cache": disable_prompt_cache,
"disable_speculative_decoding": disable_speculative_decoding,
"state": state,
}
if display_name is not None:
data["display_name"] = display_name
if inactive_timeout is not None:
data["inactive_timeout"] = inactive_timeout
if availability_zone is not None:
data["availability_zone"] = availability_zone
response, _, _ = requestor.request(
options=TogetherRequest(
method="POST",
url="endpoints",
params=data,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
def get(self, endpoint_id: str) -> DedicatedEndpoint:
"""
Get details of a specific endpoint.
Args:
endpoint_id (str): ID of the endpoint to retrieve
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
response, _, _ = requestor.request(
options=TogetherRequest(
method="GET",
url=f"endpoints/{endpoint_id}",
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
def delete(self, endpoint_id: str) -> None:
"""
Delete a specific endpoint.
Args:
endpoint_id (str): ID of the endpoint to delete
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
requestor.request(
options=TogetherRequest(
method="DELETE",
url=f"endpoints/{endpoint_id}",
),
stream=False,
)
def update(
self,
endpoint_id: str,
*,
min_replicas: Optional[int] = None,
max_replicas: Optional[int] = None,
state: Optional[Literal["STARTED", "STOPPED"]] = None,
display_name: Optional[str] = None,
inactive_timeout: Optional[int] = None,
) -> DedicatedEndpoint:
"""
Update an endpoint's configuration.
Args:
endpoint_id (str): ID of the endpoint to update
min_replicas (int, optional): The minimum number of replicas to maintain
max_replicas (int, optional): The maximum number of replicas to scale up to
state (str, optional): The desired state of the endpoint ("STARTED" or "STOPPED")
display_name (str, optional): A human-readable name for the endpoint
inactive_timeout (int, optional): The number of minutes of inactivity after which the endpoint will be automatically stopped. Set to 0 to disable automatic timeout.
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
data: Dict[str, Union[str, Dict[str, int], int]] = {}
if min_replicas is not None or max_replicas is not None:
current_min = min_replicas
current_max = max_replicas
if current_min is None or current_max is None:
# Get current values if only one is specified
current = self.get(endpoint_id=endpoint_id)
current_min = current_min or current.autoscaling.min_replicas
current_max = current_max or current.autoscaling.max_replicas
data["autoscaling"] = {
"min_replicas": current_min,
"max_replicas": current_max,
}
if state is not None:
data["state"] = state
if display_name is not None:
data["display_name"] = display_name
if inactive_timeout is not None:
data["inactive_timeout"] = inactive_timeout
response, _, _ = requestor.request(
options=TogetherRequest(
method="PATCH",
url=f"endpoints/{endpoint_id}",
params=data,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
def list_hardware(self, model: Optional[str] = None) -> List[HardwareWithStatus]:
"""
List available hardware configurations.
Args:
model (str, optional): Filter hardware configurations by model compatibility. When provided,
the response includes availability status for each compatible configuration.
Returns:
List[HardwareWithStatus]: List of hardware configurations with their status
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
params = {}
if model is not None:
params["model"] = model
response, _, _ = requestor.request(
options=TogetherRequest(
method="GET",
url="hardware",
params=params,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, dict)
assert isinstance(response.data["data"], list)
return [HardwareWithStatus(**item) for item in response.data["data"]]
def list_avzones(self) -> List[str]:
"""
List all available availability zones.
Returns:
List[str]: List of unique availability zones
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
response, _, _ = requestor.request(
options=TogetherRequest(
method="GET",
url="clusters/availability-zones",
),
stream=False,
)
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, dict)
assert isinstance(response.data["avzones"], list)
return response.data["avzones"]
class AsyncEndpoints:
def __init__(self, client: TogetherClient) -> None:
self._client = client
async def list(
self,
type: Optional[Literal["dedicated", "serverless"]] = None,
usage_type: Optional[Literal["on-demand", "reserved"]] = None,
mine: Optional[bool] = None,
) -> List[ListEndpoint]:
"""
List all endpoints, can be filtered by type and ownership.
Args:
type (str, optional): Filter endpoints by type ("dedicated" or "serverless"). Defaults to None.
usage_type (str, optional): Filter endpoints by usage type ("on-demand" or "reserved"). Defaults to None.
mine (bool, optional): If True, return only endpoints owned by the caller. Defaults to None.
Returns:
List[ListEndpoint]: List of endpoint objects
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
params: Dict[
str,
Union[
Literal["dedicated", "serverless"],
Literal["on-demand", "reserved"],
bool,
],
] = {}
if type is not None:
params["type"] = type
if usage_type is not None:
params["usage_type"] = usage_type
if mine is not None:
params["mine"] = mine
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="GET",
url="endpoints",
params=params,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, list)
return [ListEndpoint(**endpoint) for endpoint in response.data]
async def create(
self,
*,
model: str,
hardware: str,
min_replicas: int,
max_replicas: int,
display_name: Optional[str] = None,
disable_prompt_cache: bool = True,
disable_speculative_decoding: bool = True,
state: Literal["STARTED", "STOPPED"] = "STARTED",
inactive_timeout: Optional[int] = None,
availability_zone: Optional[str] = None,
) -> DedicatedEndpoint:
"""
Create a new dedicated endpoint.
Args:
model (str): The model to deploy on this endpoint
hardware (str): The hardware configuration to use for this endpoint
min_replicas (int): The minimum number of replicas to maintain
max_replicas (int): The maximum number of replicas to scale up to
display_name (str, optional): A human-readable name for the endpoint
disable_prompt_cache (bool, optional): Whether to disable the prompt cache. Defaults to False.
disable_speculative_decoding (bool, optional): Whether to disable speculative decoding. Defaults to False.
state (str, optional): The desired state of the endpoint. Defaults to "STARTED".
inactive_timeout (int, optional): The number of minutes of inactivity after which the endpoint will be automatically stopped. Set to 0 to disable automatic timeout.
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
if disable_prompt_cache:
warnings.warn(
"The 'disable_prompt_cache' parameter (CLI flag: '--no-prompt-cache') is deprecated and will be removed in a future version.",
stacklevel=2,
)
requestor = api_requestor.APIRequestor(
client=self._client,
)
data: Dict[str, Union[str, bool, Dict[str, int], int]] = {
"model": model,
"hardware": hardware,
"autoscaling": {
"min_replicas": min_replicas,
"max_replicas": max_replicas,
},
"disable_prompt_cache": disable_prompt_cache,
"disable_speculative_decoding": disable_speculative_decoding,
"state": state,
}
if display_name is not None:
data["display_name"] = display_name
if inactive_timeout is not None:
data["inactive_timeout"] = inactive_timeout
if availability_zone is not None:
data["availability_zone"] = availability_zone
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="POST",
url="endpoints",
params=data,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
async def get(self, endpoint_id: str) -> DedicatedEndpoint:
"""
Get details of a specific endpoint.
Args:
endpoint_id (str): ID of the endpoint to retrieve
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="GET",
url=f"endpoints/{endpoint_id}",
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
async def delete(self, endpoint_id: str) -> None:
"""
Delete a specific endpoint.
Args:
endpoint_id (str): ID of the endpoint to delete
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
await requestor.arequest(
options=TogetherRequest(
method="DELETE",
url=f"endpoints/{endpoint_id}",
),
stream=False,
)
async def update(
self,
endpoint_id: str,
*,
min_replicas: Optional[int] = None,
max_replicas: Optional[int] = None,
state: Optional[Literal["STARTED", "STOPPED"]] = None,
display_name: Optional[str] = None,
inactive_timeout: Optional[int] = None,
) -> DedicatedEndpoint:
"""
Update an endpoint's configuration.
Args:
endpoint_id (str): ID of the endpoint to update
min_replicas (int, optional): The minimum number of replicas to maintain
max_replicas (int, optional): The maximum number of replicas to scale up to
state (str, optional): The desired state of the endpoint ("STARTED" or "STOPPED")
display_name (str, optional): A human-readable name for the endpoint
inactive_timeout (int, optional): The number of minutes of inactivity after which the endpoint will be automatically stopped. Set to 0 to disable automatic timeout.
Returns:
DedicatedEndpoint: Object containing endpoint information
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
data: Dict[str, Union[str, Dict[str, int], int]] = {}
if min_replicas is not None or max_replicas is not None:
current_min = min_replicas
current_max = max_replicas
if current_min is None or current_max is None:
# Get current values if only one is specified
current = await self.get(endpoint_id=endpoint_id)
current_min = current_min or current.autoscaling.min_replicas
current_max = current_max or current.autoscaling.max_replicas
data["autoscaling"] = {
"min_replicas": current_min,
"max_replicas": current_max,
}
if state is not None:
data["state"] = state
if display_name is not None:
data["display_name"] = display_name
if inactive_timeout is not None:
data["inactive_timeout"] = inactive_timeout
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="PATCH",
url=f"endpoints/{endpoint_id}",
params=data,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
return DedicatedEndpoint(**response.data)
async def list_hardware(
self, model: Optional[str] = None
) -> List[HardwareWithStatus]:
"""
List available hardware configurations.
Args:
model (str, optional): Filter hardware configurations by model compatibility. When provided,
the response includes availability status for each compatible configuration.
Returns:
List[HardwareWithStatus]: List of hardware configurations with their status
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
params = {}
if model is not None:
params["model"] = model
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="GET",
url="hardware",
params=params,
),
stream=False,
)
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, dict)
assert isinstance(response.data["data"], list)
return [HardwareWithStatus(**item) for item in response.data["data"]]
async def list_avzones(self) -> List[str]:
"""
List all availability zones.
Returns:
List[str]: List of unique availability zones
"""
requestor = api_requestor.APIRequestor(
client=self._client,
)
response, _, _ = await requestor.arequest(
options=TogetherRequest(
method="GET",
url="clusters/availability-zones",
),
stream=False,
)
assert isinstance(response, TogetherResponse)
assert isinstance(response.data, dict)
assert isinstance(response.data["avzones"], list)
return response.data["avzones"]