/
client.py
1699 lines (1532 loc) · 70.3 KB
/
client.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
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
"""The main Client class for the Python client."""
from __future__ import annotations
import concurrent.futures
import json
import os
import re
import secrets
import tempfile
import threading
import time
import urllib.parse
import uuid
import warnings
from concurrent.futures import Future
from dataclasses import dataclass
from datetime import datetime
from pathlib import Path
from threading import Lock
from typing import Any, Callable, Literal
import httpx
import huggingface_hub
import websockets
from huggingface_hub import CommitOperationAdd, SpaceHardware, SpaceStage
from huggingface_hub.utils import (
RepositoryNotFoundError,
build_hf_headers,
send_telemetry,
)
from packaging import version
from gradio_client import serializing, utils
from gradio_client.documentation import document, set_documentation_group
from gradio_client.exceptions import SerializationSetupError
from gradio_client.utils import (
Communicator,
JobStatus,
Message,
QueueError,
ServerMessage,
Status,
StatusUpdate,
)
set_documentation_group("py-client")
DEFAULT_TEMP_DIR = os.environ.get("GRADIO_TEMP_DIR") or str(
Path(tempfile.gettempdir()) / "gradio"
)
@document("predict", "submit", "view_api", "duplicate", "deploy_discord")
class Client:
"""
The main Client class for the Python client. This class is used to connect to a remote Gradio app and call its API endpoints.
Example:
from gradio_client import Client
client = Client("abidlabs/whisper-large-v2") # connecting to a Hugging Face Space
client.predict("test.mp4", api_name="/predict")
>> What a nice recording! # returns the result of the remote API call
client = Client("https://bec81a83-5b5c-471e.gradio.live") # connecting to a temporary Gradio share URL
job = client.submit("hello", api_name="/predict") # runs the prediction in a background thread
job.result()
>> 49 # returns the result of the remote API call (blocking call)
"""
def __init__(
self,
src: str,
hf_token: str | None = None,
max_workers: int = 40,
serialize: bool = True,
output_dir: str | Path = DEFAULT_TEMP_DIR,
verbose: bool = True,
auth: tuple[str, str] | None = None,
):
"""
Parameters:
src: Either the name of the Hugging Face Space to load, (e.g. "abidlabs/whisper-large-v2") or the full URL (including "http" or "https") of the hosted Gradio app to load (e.g. "http://mydomain.com/app" or "https://bec81a83-5b5c-471e.gradio.live/").
hf_token: The Hugging Face token to use to access private Spaces. Automatically fetched if you are logged in via the Hugging Face Hub CLI. Obtain from: https://huggingface.co/settings/token
max_workers: The maximum number of thread workers that can be used to make requests to the remote Gradio app simultaneously.
serialize: Whether the client should serialize the inputs and deserialize the outputs of the remote API. If set to False, the client will pass the inputs and outputs as-is, without serializing/deserializing them. E.g. you if you set this to False, you'd submit an image in base64 format instead of a filepath, and you'd get back an image in base64 format from the remote API instead of a filepath.
output_dir: The directory to save files that are downloaded from the remote API. If None, reads from the GRADIO_TEMP_DIR environment variable. Defaults to a temporary directory on your machine.
verbose: Whether the client should print statements to the console.
"""
self.verbose = verbose
self.hf_token = hf_token
self.serialize = serialize
self.headers = build_hf_headers(
token=hf_token,
library_name="gradio_client",
library_version=utils.__version__,
)
self.space_id = None
self.cookies: dict[str, str] = {}
self.output_dir = (
str(output_dir) if isinstance(output_dir, Path) else output_dir
)
if src.startswith("http://") or src.startswith("https://"):
_src = src if src.endswith("/") else src + "/"
else:
_src = self._space_name_to_src(src)
if _src is None:
raise ValueError(
f"Could not find Space: {src}. If it is a private Space, please provide an hf_token."
)
self.space_id = src
self.src = _src
state = self._get_space_state()
if state == SpaceStage.BUILDING:
if self.verbose:
print("Space is still building. Please wait...")
while self._get_space_state() == SpaceStage.BUILDING:
time.sleep(2) # so we don't get rate limited by the API
pass
if state in utils.INVALID_RUNTIME:
raise ValueError(
f"The current space is in the invalid state: {state}. "
"Please contact the owner to fix this."
)
if self.verbose:
print(f"Loaded as API: {self.src} ✔")
if auth is not None:
self._login(auth)
self.config = self._get_config()
self.protocol: str = self.config.get("protocol", "ws")
self.api_url = urllib.parse.urljoin(self.src, utils.API_URL)
self.sse_url = urllib.parse.urljoin(
self.src, utils.SSE_URL_V0 if self.protocol == "sse" else utils.SSE_URL
)
self.sse_data_url = urllib.parse.urljoin(
self.src,
utils.SSE_DATA_URL_V0 if self.protocol == "sse" else utils.SSE_DATA_URL,
)
self.ws_url = urllib.parse.urljoin(
self.src.replace("http", "ws", 1), utils.WS_URL
)
self.upload_url = urllib.parse.urljoin(self.src, utils.UPLOAD_URL)
self.reset_url = urllib.parse.urljoin(self.src, utils.RESET_URL)
self.app_version = version.parse(self.config.get("version", "2.0"))
self._info = self._get_api_info()
self.session_hash = str(uuid.uuid4())
endpoint_class = (
Endpoint if self.protocol.startswith("sse") else EndpointV3Compatibility
)
self.endpoints = [
endpoint_class(self, fn_index, dependency, self.protocol)
for fn_index, dependency in enumerate(self.config["dependencies"])
]
# Create a pool of threads to handle the requests
self.executor = concurrent.futures.ThreadPoolExecutor(max_workers=max_workers)
# Disable telemetry by setting the env variable HF_HUB_DISABLE_TELEMETRY=1
threading.Thread(target=self._telemetry_thread).start()
self.stream_open = False
self.streaming_future: Future | None = None
self.pending_messages_per_event: dict[str, list[Message | None]] = {}
self.pending_event_ids: set[str] = set()
async def stream_messages(self) -> None:
try:
async with httpx.AsyncClient(timeout=httpx.Timeout(timeout=None)) as client:
async with client.stream(
"GET",
self.sse_url,
params={"session_hash": self.session_hash},
headers=self.headers,
cookies=self.cookies,
) as response:
async for line in response.aiter_lines():
line = line.rstrip("\n")
if not len(line):
continue
if line.startswith("data:"):
resp = json.loads(line[5:])
if resp["msg"] == ServerMessage.heartbeat:
continue
elif resp["msg"] == ServerMessage.server_stopped:
for (
pending_messages
) in self.pending_messages_per_event.values():
pending_messages.append(resp)
return
event_id = resp["event_id"]
if event_id not in self.pending_messages_per_event:
self.pending_messages_per_event[event_id] = []
self.pending_messages_per_event[event_id].append(resp)
if resp["msg"] == ServerMessage.process_completed:
self.pending_event_ids.remove(event_id)
if len(self.pending_event_ids) == 0:
self.stream_open = False
return
else:
raise ValueError(f"Unexpected SSE line: '{line}'")
except BaseException as e:
import traceback
traceback.print_exc()
raise e
async def send_data(self, data, hash_data):
async with httpx.AsyncClient() as client:
req = await client.post(
self.sse_data_url,
json={**data, **hash_data},
headers=self.headers,
cookies=self.cookies,
)
if req.status_code == 503:
raise QueueError("Queue is full! Please try again.")
req.raise_for_status()
resp = req.json()
event_id = resp["event_id"]
if not self.stream_open:
self.stream_open = True
def open_stream():
return utils.synchronize_async(self.stream_messages)
def close_stream(_):
self.stream_open = False
for _, pending_messages in self.pending_messages_per_event.items():
pending_messages.append(None)
if self.streaming_future is None or self.streaming_future.done():
self.streaming_future = self.executor.submit(open_stream)
self.streaming_future.add_done_callback(close_stream)
return event_id
@classmethod
def duplicate(
cls,
from_id: str,
to_id: str | None = None,
hf_token: str | None = None,
private: bool = True,
hardware: Literal[
"cpu-basic",
"cpu-upgrade",
"t4-small",
"t4-medium",
"a10g-small",
"a10g-large",
"a100-large",
]
| SpaceHardware
| None = None,
secrets: dict[str, str] | None = None,
sleep_timeout: int = 5,
max_workers: int = 40,
verbose: bool = True,
):
"""
Duplicates a Hugging Face Space under your account and returns a Client object
for the new Space. No duplication is created if the Space already exists in your
account (to override this, provide a new name for the new Space using `to_id`).
To use this method, you must provide an `hf_token` or be logged in via the Hugging
Face Hub CLI.
The new Space will be private by default and use the same hardware as the original
Space. This can be changed by using the `private` and `hardware` parameters. For
hardware upgrades (beyond the basic CPU tier), you may be required to provide
billing information on Hugging Face: https://huggingface.co/settings/billing
Parameters:
from_id: The name of the Hugging Face Space to duplicate in the format "{username}/{space_id}", e.g. "gradio/whisper".
to_id: The name of the new Hugging Face Space to create, e.g. "abidlabs/whisper-duplicate". If not provided, the new Space will be named "{your_HF_username}/{space_id}".
hf_token: The Hugging Face token to use to access private Spaces. Automatically fetched if you are logged in via the Hugging Face Hub CLI. Obtain from: https://huggingface.co/settings/token
private: Whether the new Space should be private (True) or public (False). Defaults to True.
hardware: The hardware tier to use for the new Space. Defaults to the same hardware tier as the original Space. Options include "cpu-basic", "cpu-upgrade", "t4-small", "t4-medium", "a10g-small", "a10g-large", "a100-large", subject to availability.
secrets: A dictionary of (secret key, secret value) to pass to the new Space. Defaults to None. Secrets are only used when the Space is duplicated for the first time, and are not updated if the duplicated Space already exists.
sleep_timeout: The number of minutes after which the duplicate Space will be puased if no requests are made to it (to minimize billing charges). Defaults to 5 minutes.
max_workers: The maximum number of thread workers that can be used to make requests to the remote Gradio app simultaneously.
verbose: Whether the client should print statements to the console.
Example:
import os
from gradio_client import Client
HF_TOKEN = os.environ.get("HF_TOKEN")
client = Client.duplicate("abidlabs/whisper", hf_token=HF_TOKEN)
client.predict("audio_sample.wav")
>> "This is a test of the whisper speech recognition model."
"""
try:
original_info = huggingface_hub.get_space_runtime(from_id, token=hf_token)
except RepositoryNotFoundError as rnfe:
raise ValueError(
f"Could not find Space: {from_id}. If it is a private Space, please provide an `hf_token`."
) from rnfe
if to_id:
if "/" in to_id:
to_id = to_id.split("/")[1]
space_id = huggingface_hub.get_full_repo_name(to_id, token=hf_token)
else:
space_id = huggingface_hub.get_full_repo_name(
from_id.split("/")[1], token=hf_token
)
try:
huggingface_hub.get_space_runtime(space_id, token=hf_token)
if verbose:
print(
f"Using your existing Space: {utils.SPACE_URL.format(space_id)} 🤗"
)
if secrets is not None:
warnings.warn(
"Secrets are only used when the Space is duplicated for the first time, and are not updated if the duplicated Space already exists."
)
except RepositoryNotFoundError:
if verbose:
print(f"Creating a duplicate of {from_id} for your own use... 🤗")
huggingface_hub.duplicate_space(
from_id=from_id,
to_id=space_id,
token=hf_token,
exist_ok=True,
private=private,
)
if secrets is not None:
for key, value in secrets.items():
huggingface_hub.add_space_secret(
space_id, key, value, token=hf_token
)
if verbose:
print(f"Created new Space: {utils.SPACE_URL.format(space_id)}")
current_info = huggingface_hub.get_space_runtime(space_id, token=hf_token)
current_hardware = (
current_info.hardware or huggingface_hub.SpaceHardware.CPU_BASIC
)
hardware = hardware or original_info.hardware
if current_hardware != hardware:
huggingface_hub.request_space_hardware(space_id, hardware, token=hf_token) # type: ignore
print(
f"-------\nNOTE: this Space uses upgraded hardware: {hardware}... see billing info at https://huggingface.co/settings/billing\n-------"
)
# Setting a timeout only works if the hardware is not basic
# so set it here after the hardware has been requested
if hardware != huggingface_hub.SpaceHardware.CPU_BASIC:
utils.set_space_timeout(
space_id, hf_token=hf_token, timeout_in_seconds=sleep_timeout * 60
)
if verbose:
print("")
client = cls(
space_id, hf_token=hf_token, max_workers=max_workers, verbose=verbose
)
return client
def _get_space_state(self):
if not self.space_id:
return None
info = huggingface_hub.get_space_runtime(self.space_id, token=self.hf_token)
return info.stage
def predict(
self,
*args,
api_name: str | None = None,
fn_index: int | None = None,
) -> Any:
"""
Calls the Gradio API and returns the result (this is a blocking call).
Parameters:
args: The arguments to pass to the remote API. The order of the arguments must match the order of the inputs in the Gradio app.
api_name: The name of the API endpoint to call starting with a leading slash, e.g. "/predict". Does not need to be provided if the Gradio app has only one named API endpoint.
fn_index: As an alternative to api_name, this parameter takes the index of the API endpoint to call, e.g. 0. Both api_name and fn_index can be provided, but if they conflict, api_name will take precedence.
Returns:
The result of the API call. Will be a Tuple if the API has multiple outputs.
Example:
from gradio_client import Client
client = Client(src="gradio/calculator")
client.predict(5, "add", 4, api_name="/predict")
>> 9.0
"""
inferred_fn_index = self._infer_fn_index(api_name, fn_index)
if self.endpoints[inferred_fn_index].is_continuous:
raise ValueError(
"Cannot call predict on this function as it may run forever. Use submit instead."
)
return self.submit(*args, api_name=api_name, fn_index=fn_index).result()
def new_helper(self, fn_index: int) -> Communicator:
return Communicator(
Lock(),
JobStatus(),
self.endpoints[fn_index].process_predictions,
self.reset_url,
)
def submit(
self,
*args,
api_name: str | None = None,
fn_index: int | None = None,
result_callbacks: Callable | list[Callable] | None = None,
) -> Job:
"""
Creates and returns a Job object which calls the Gradio API in a background thread. The job can be used to retrieve the status and result of the remote API call.
Parameters:
args: The arguments to pass to the remote API. The order of the arguments must match the order of the inputs in the Gradio app.
api_name: The name of the API endpoint to call starting with a leading slash, e.g. "/predict". Does not need to be provided if the Gradio app has only one named API endpoint.
fn_index: As an alternative to api_name, this parameter takes the index of the API endpoint to call, e.g. 0. Both api_name and fn_index can be provided, but if they conflict, api_name will take precedence.
result_callbacks: A callback function, or list of callback functions, to be called when the result is ready. If a list of functions is provided, they will be called in order. The return values from the remote API are provided as separate parameters into the callback. If None, no callback will be called.
Returns:
A Job object that can be used to retrieve the status and result of the remote API call.
Example:
from gradio_client import Client
client = Client(src="gradio/calculator")
job = client.submit(5, "add", 4, api_name="/predict")
job.status()
>> <Status.STARTING: 'STARTING'>
job.result() # blocking call
>> 9.0
"""
inferred_fn_index = self._infer_fn_index(api_name, fn_index)
helper = None
if self.endpoints[inferred_fn_index].protocol in ("ws", "sse", "sse_v1"):
helper = self.new_helper(inferred_fn_index)
end_to_end_fn = self.endpoints[inferred_fn_index].make_end_to_end_fn(helper)
future = self.executor.submit(end_to_end_fn, *args)
job = Job(
future, communicator=helper, verbose=self.verbose, space_id=self.space_id
)
if result_callbacks:
if isinstance(result_callbacks, Callable):
result_callbacks = [result_callbacks]
def create_fn(callback) -> Callable:
def fn(future):
if isinstance(future.result(), tuple):
callback(*future.result())
else:
callback(future.result())
return fn
for callback in result_callbacks:
job.add_done_callback(create_fn(callback))
return job
def _get_api_info(self):
if self.serialize:
api_info_url = urllib.parse.urljoin(self.src, utils.API_INFO_URL)
else:
api_info_url = urllib.parse.urljoin(self.src, utils.RAW_API_INFO_URL)
if self.app_version > version.Version("3.36.1"):
r = httpx.get(api_info_url, headers=self.headers, cookies=self.cookies)
if r.is_success:
info = r.json()
else:
raise ValueError(f"Could not fetch api info for {self.src}: {r.text}")
else:
fetch = httpx.post(
utils.SPACE_FETCHER_URL,
json={"config": json.dumps(self.config), "serialize": self.serialize},
)
if fetch.is_success:
info = fetch.json()["api"]
else:
raise ValueError(
f"Could not fetch api info for {self.src}: {fetch.text}"
)
return info
def view_api(
self,
all_endpoints: bool | None = None,
print_info: bool = True,
return_format: Literal["dict", "str"] | None = None,
) -> dict | str | None:
"""
Prints the usage info for the API. If the Gradio app has multiple API endpoints, the usage info for each endpoint will be printed separately. If return_format="dict" the info is returned in dictionary format, as shown in the example below.
Parameters:
all_endpoints: If True, prints information for both named and unnamed endpoints in the Gradio app. If False, will only print info about named endpoints. If None (default), will print info about named endpoints, unless there aren't any -- in which it will print info about unnamed endpoints.
print_info: If True, prints the usage info to the console. If False, does not print the usage info.
return_format: If None, nothing is returned. If "str", returns the same string that would be printed to the console. If "dict", returns the usage info as a dictionary that can be programmatically parsed, and *all endpoints are returned in the dictionary* regardless of the value of `all_endpoints`. The format of the dictionary is in the docstring of this method.
Example:
from gradio_client import Client
client = Client(src="gradio/calculator")
client.view_api(return_format="dict")
>> {
'named_endpoints': {
'/predict': {
'parameters': [
{
'label': 'num1',
'type_python': 'int | float',
'type_description': 'numeric value',
'component': 'Number',
'example_input': '5'
},
{
'label': 'operation',
'type_python': 'str',
'type_description': 'string value',
'component': 'Radio',
'example_input': 'add'
},
{
'label': 'num2',
'type_python': 'int | float',
'type_description': 'numeric value',
'component': 'Number',
'example_input': '5'
},
],
'returns': [
{
'label': 'output',
'type_python': 'int | float',
'type_description': 'numeric value',
'component': 'Number',
},
]
},
'/flag': {
'parameters': [
...
],
'returns': [
...
]
}
}
'unnamed_endpoints': {
2: {
'parameters': [
...
],
'returns': [
...
]
}
}
}
}
"""
num_named_endpoints = len(self._info["named_endpoints"])
num_unnamed_endpoints = len(self._info["unnamed_endpoints"])
if num_named_endpoints == 0 and all_endpoints is None:
all_endpoints = True
human_info = "Client.predict() Usage Info\n---------------------------\n"
human_info += f"Named API endpoints: {num_named_endpoints}\n"
for api_name, endpoint_info in self._info["named_endpoints"].items():
human_info += self._render_endpoints_info(api_name, endpoint_info)
if all_endpoints:
human_info += f"\nUnnamed API endpoints: {num_unnamed_endpoints}\n"
for fn_index, endpoint_info in self._info["unnamed_endpoints"].items():
# When loading from json, the fn_indices are read as strings
# because json keys can only be strings
human_info += self._render_endpoints_info(int(fn_index), endpoint_info)
else:
if num_unnamed_endpoints > 0:
human_info += f"\nUnnamed API endpoints: {num_unnamed_endpoints}, to view, run Client.view_api(all_endpoints=True)\n"
if print_info:
print(human_info)
if return_format == "str":
return human_info
elif return_format == "dict":
return self._info
def reset_session(self) -> None:
self.session_hash = str(uuid.uuid4())
def _render_endpoints_info(
self,
name_or_index: str | int,
endpoints_info: dict[str, list[dict[str, Any]]],
) -> str:
parameter_names = [p["label"] for p in endpoints_info["parameters"]]
parameter_names = [utils.sanitize_parameter_names(p) for p in parameter_names]
rendered_parameters = ", ".join(parameter_names)
if rendered_parameters:
rendered_parameters = rendered_parameters + ", "
return_values = [p["label"] for p in endpoints_info["returns"]]
return_values = [utils.sanitize_parameter_names(r) for r in return_values]
rendered_return_values = ", ".join(return_values)
if len(return_values) > 1:
rendered_return_values = f"({rendered_return_values})"
if isinstance(name_or_index, str):
final_param = f'api_name="{name_or_index}"'
elif isinstance(name_or_index, int):
final_param = f"fn_index={name_or_index}"
else:
raise ValueError("name_or_index must be a string or integer")
human_info = f"\n - predict({rendered_parameters}{final_param}) -> {rendered_return_values}\n"
human_info += " Parameters:\n"
if endpoints_info["parameters"]:
for info in endpoints_info["parameters"]:
desc = (
f" ({info['python_type']['description']})"
if info["python_type"].get("description")
else ""
)
type_ = info["python_type"]["type"]
human_info += f" - [{info['component']}] {utils.sanitize_parameter_names(info['label'])}: {type_}{desc} \n"
else:
human_info += " - None\n"
human_info += " Returns:\n"
if endpoints_info["returns"]:
for info in endpoints_info["returns"]:
desc = (
f" ({info['python_type']['description']})"
if info["python_type"].get("description")
else ""
)
type_ = info["python_type"]["type"]
human_info += f" - [{info['component']}] {utils.sanitize_parameter_names(info['label'])}: {type_}{desc} \n"
else:
human_info += " - None\n"
return human_info
def __repr__(self):
return self.view_api(print_info=False, return_format="str")
def __str__(self):
return self.view_api(print_info=False, return_format="str")
def _telemetry_thread(self) -> None:
# Disable telemetry by setting the env variable HF_HUB_DISABLE_TELEMETRY=1
data = {
"src": self.src,
}
try:
send_telemetry(
topic="py_client/initiated",
library_name="gradio_client",
library_version=utils.__version__,
user_agent=data,
)
except Exception:
pass
def _infer_fn_index(self, api_name: str | None, fn_index: int | None) -> int:
inferred_fn_index = None
if api_name is not None:
for i, d in enumerate(self.config["dependencies"]):
config_api_name = d.get("api_name")
if config_api_name is None or config_api_name is False:
continue
if "/" + config_api_name == api_name:
inferred_fn_index = i
break
else:
error_message = f"Cannot find a function with `api_name`: {api_name}."
if not api_name.startswith("/"):
error_message += " Did you mean to use a leading slash?"
raise ValueError(error_message)
elif fn_index is not None:
inferred_fn_index = fn_index
if (
inferred_fn_index >= len(self.endpoints)
or not self.endpoints[inferred_fn_index].is_valid
):
raise ValueError(f"Invalid function index: {fn_index}.")
else:
valid_endpoints = [
e
for e in self.endpoints
if e.is_valid
and e.api_name is not None
and e.backend_fn is not None
and e.show_api
]
if len(valid_endpoints) == 1:
inferred_fn_index = valid_endpoints[0].fn_index
else:
raise ValueError(
"This Gradio app might have multiple endpoints. Please specify an `api_name` or `fn_index`"
)
return inferred_fn_index
def __del__(self):
if hasattr(self, "executor"):
self.executor.shutdown(wait=True)
def _space_name_to_src(self, space) -> str | None:
return huggingface_hub.space_info(space, token=self.hf_token).host # type: ignore
def _login(self, auth: tuple[str, str]):
resp = httpx.post(
urllib.parse.urljoin(self.src, utils.LOGIN_URL),
data={"username": auth[0], "password": auth[1]},
)
if not resp.is_success:
raise ValueError(f"Could not login to {self.src}")
self.cookies = {
name: value for name, value in resp.cookies.items() if value is not None
}
def _get_config(self) -> dict:
r = httpx.get(
urllib.parse.urljoin(self.src, utils.CONFIG_URL),
headers=self.headers,
cookies=self.cookies,
)
if r.is_success:
return r.json()
elif r.status_code == 401:
raise ValueError(f"Could not load {self.src}. Please login.")
else: # to support older versions of Gradio
r = httpx.get(self.src, headers=self.headers, cookies=self.cookies)
if not r.is_success:
raise ValueError(f"Could not fetch config for {self.src}")
# some basic regex to extract the config
result = re.search(r"window.gradio_config = (.*?);[\s]*</script>", r.text)
try:
config = json.loads(result.group(1)) # type: ignore
except AttributeError as ae:
raise ValueError(
f"Could not get Gradio config from: {self.src}"
) from ae
if "allow_flagging" in config:
raise ValueError(
"Gradio 2.x is not supported by this client. Please upgrade your Gradio app to Gradio 3.x or higher."
)
return config
def deploy_discord(
self,
discord_bot_token: str | None = None,
api_names: list[str | tuple[str, str]] | None = None,
to_id: str | None = None,
hf_token: str | None = None,
private: bool = False,
):
"""
Deploy the upstream app as a discord bot. Currently only supports gr.ChatInterface.
Parameters:
discord_bot_token: This is the "password" needed to be able to launch the bot. Users can get a token by creating a bot app on the discord website. If run the method without specifying a token, the space will explain how to get one. See here: https://huggingface.co/spaces/freddyaboulton/test-discord-bot-v1.
api_names: The api_names of the app to turn into bot commands. This parameter currently has no effect as ChatInterface only has one api_name ('/chat').
to_id: The name of the space hosting the discord bot. If None, the name will be gradio-discord-bot-{random-substring}
hf_token: HF api token with write priviledges in order to upload the files to HF space. Can be ommitted if logged in via the HuggingFace CLI, unless the upstream space is private. Obtain from: https://huggingface.co/settings/token
private: Whether the space hosting the discord bot is private. The visibility of the discord bot itself is set via the discord website. See https://huggingface.co/spaces/freddyaboulton/test-discord-bot-v1
"""
if self.config["mode"] == "chat_interface" and not api_names:
api_names = [("chat", "chat")]
valid_list = isinstance(api_names, list) and (
isinstance(n, str)
or (
isinstance(n, tuple) and isinstance(n[0], str) and isinstance(n[1], str)
)
for n in api_names
)
if api_names is None or not valid_list:
raise ValueError(
f"Each entry in api_names must be either a string or a tuple of strings. Received {api_names}"
)
if len(api_names) != 1:
raise ValueError("Currently only one api_name can be deployed to discord.")
for i, name in enumerate(api_names):
if isinstance(name, str):
api_names[i] = (name, name)
fn = next(
(ep for ep in self.endpoints if ep.api_name == f"/{api_names[0][0]}"), None
)
if not fn:
raise ValueError(
f"api_name {api_names[0][0]} not present in {self.space_id or self.src}"
)
inputs = [inp for inp in fn.input_component_types if not inp.skip]
outputs = [inp for inp in fn.input_component_types if not inp.skip]
if not inputs == ["textbox"] and outputs == ["textbox"]:
raise ValueError(
"Currently only api_names with a single textbox as input and output are supported. "
f"Received {inputs} and {outputs}"
)
is_private = False
if self.space_id:
is_private = huggingface_hub.space_info(self.space_id).private
if is_private and not hf_token:
raise ValueError(
f"Since {self.space_id} is private, you must explicitly pass in hf_token "
"so that it can be added as a secret in the discord bot space."
)
if to_id:
if "/" in to_id:
to_id = to_id.split("/")[1]
space_id = huggingface_hub.get_full_repo_name(to_id, token=hf_token)
else:
if self.space_id:
space_id = f'{self.space_id.split("/")[1]}-gradio-discord-bot'
else:
space_id = f"gradio-discord-bot-{secrets.token_hex(4)}"
space_id = huggingface_hub.get_full_repo_name(space_id, token=hf_token)
api = huggingface_hub.HfApi()
try:
huggingface_hub.space_info(space_id)
first_upload = False
except huggingface_hub.utils.RepositoryNotFoundError:
first_upload = True
huggingface_hub.create_repo(
space_id,
repo_type="space",
space_sdk="gradio",
token=hf_token,
exist_ok=True,
private=private,
)
if first_upload:
huggingface_hub.metadata_update(
repo_id=space_id,
repo_type="space",
metadata={"tags": ["gradio-discord-bot"]},
)
with open(str(Path(__file__).parent / "templates" / "discord_chat.py")) as f:
app = f.read()
app = app.replace("<<app-src>>", self.src)
app = app.replace("<<api-name>>", api_names[0][0])
app = app.replace("<<command-name>>", api_names[0][1])
with tempfile.NamedTemporaryFile(mode="w", delete=False) as app_file:
with tempfile.NamedTemporaryFile(mode="w", delete=False) as requirements:
app_file.write(app)
requirements.write("\n".join(["discord.py==2.3.1"]))
operations = [
CommitOperationAdd(path_in_repo="app.py", path_or_fileobj=app_file.name),
CommitOperationAdd(
path_in_repo="requirements.txt", path_or_fileobj=requirements.name
),
]
api.create_commit(
repo_id=space_id,
commit_message="Deploy Discord Bot",
repo_type="space",
operations=operations,
token=hf_token,
)
if discord_bot_token:
huggingface_hub.add_space_secret(
space_id, "DISCORD_TOKEN", discord_bot_token, token=hf_token
)
if is_private:
huggingface_hub.add_space_secret(
space_id,
"HF_TOKEN",
hf_token, # type: ignore
token=hf_token,
)
url = f"https://huggingface.co/spaces/{space_id}"
print(f"See your discord bot here! {url}")
return url
@dataclass
class ComponentApiType:
skip: bool
value_is_file: bool
is_state: bool
@dataclass
class ReplaceMe:
index: int
class Endpoint:
"""Helper class for storing all the information about a single API endpoint."""
def __init__(
self, client: Client, fn_index: int, dependency: dict, protocol: str = "sse_v1"
):
self.client: Client = client
self.fn_index = fn_index
self.dependency = dependency
api_name = dependency.get("api_name")
self.api_name: str | Literal[False] | None = (
"/" + api_name if isinstance(api_name, str) else api_name
)
self.protocol = protocol
self.input_component_types = [
self._get_component_type(id_) for id_ in dependency["inputs"]
]
self.output_component_types = [
self._get_component_type(id_) for id_ in dependency["outputs"]
]
self.root_url = client.src + "/" if not client.src.endswith("/") else client.src
self.is_continuous = dependency.get("types", {}).get("continuous", False)
self.download_file = lambda d: self._download_file(
d,
save_dir=self.client.output_dir,
hf_token=self.client.hf_token,
root_url=self.root_url,
)
# Disallow hitting endpoints that the Gradio app has disabled
self.is_valid = self.api_name is not False
self.backend_fn = dependency.get("backend_fn")
self.show_api = dependency.get("show_api")
def _get_component_type(self, component_id: int):
component = next(
i for i in self.client.config["components"] if i["id"] == component_id
)
skip_api = component.get("skip_api", component["type"] in utils.SKIP_COMPONENTS)
return ComponentApiType(
skip_api,
self.value_is_file(component),
component["type"] == "state",
)
@staticmethod
def value_is_file(component: dict) -> bool:
# Hacky for now
if "api_info" not in component:
return False
return utils.value_is_file(component["api_info"])
def __repr__(self):
return f"Endpoint src: {self.client.src}, api_name: {self.api_name}, fn_index: {self.fn_index}"
def __str__(self):
return self.__repr__()
def make_end_to_end_fn(self, helper: Communicator | None = None):
_predict = self.make_predict(helper)
def _inner(*data):
if not self.is_valid:
raise utils.InvalidAPIEndpointError()
data = self.insert_state(*data)
if self.client.serialize:
data = self.serialize(*data)
predictions = _predict(*data)
predictions = self.process_predictions(*predictions)
# Append final output only if not already present
# for consistency between generators and not generators
if helper:
with helper.lock:
if not helper.job.outputs:
helper.job.outputs.append(predictions)
return predictions
return _inner
def make_predict(self, helper: Communicator | None = None):
def _predict(*data) -> tuple:
data = {
"data": data,
"fn_index": self.fn_index,
"session_hash": self.client.session_hash,
}
hash_data = {
"fn_index": self.fn_index,
"session_hash": self.client.session_hash,
}
if self.protocol == "sse":
result = utils.synchronize_async(
self._sse_fn_v0, data, hash_data, helper
)
elif self.protocol == "sse_v1":
event_id = utils.synchronize_async(
self.client.send_data, data, hash_data
)
self.client.pending_event_ids.add(event_id)
self.client.pending_messages_per_event[event_id] = []
result = utils.synchronize_async(self._sse_fn_v1, helper, event_id)