mistralai 1.3.0__py3-none-any.whl → 1.4.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- mistralai/__init__.py +10 -1
- mistralai/_version.py +4 -1
- mistralai/agents.py +58 -14
- mistralai/chat.py +58 -14
- mistralai/classifiers.py +32 -20
- mistralai/embeddings.py +16 -10
- mistralai/files.py +94 -34
- mistralai/fim.py +30 -14
- mistralai/jobs.py +80 -32
- mistralai/mistral_jobs.py +64 -24
- mistralai/models/__init__.py +5 -0
- mistralai/models/agentscompletionrequest.py +5 -0
- mistralai/models/agentscompletionstreamrequest.py +5 -0
- mistralai/models/chatcompletionrequest.py +5 -0
- mistralai/models/chatcompletionstreamrequest.py +5 -0
- mistralai/models/fileschema.py +3 -2
- mistralai/models/function.py +3 -0
- mistralai/models/prediction.py +26 -0
- mistralai/models/retrievefileout.py +3 -2
- mistralai/models/toolcall.py +3 -0
- mistralai/models/uploadfileout.py +3 -2
- mistralai/models_.py +92 -48
- mistralai/sdk.py +2 -1
- mistralai/sdkconfiguration.py +10 -4
- {mistralai-1.3.0.dist-info → mistralai-1.4.0.dist-info}/METADATA +9 -41
- {mistralai-1.3.0.dist-info → mistralai-1.4.0.dist-info}/RECORD +30 -29
- mistralai_azure/_hooks/custom_user_agent.py +1 -1
- mistralai_gcp/sdk.py +1 -2
- {mistralai-1.3.0.dist-info → mistralai-1.4.0.dist-info}/LICENSE +0 -0
- {mistralai-1.3.0.dist-info → mistralai-1.4.0.dist-info}/WHEEL +0 -0
mistralai/__init__.py
CHANGED
|
@@ -1,9 +1,18 @@
|
|
|
1
1
|
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
|
|
2
2
|
|
|
3
|
-
from ._version import
|
|
3
|
+
from ._version import (
|
|
4
|
+
__title__,
|
|
5
|
+
__version__,
|
|
6
|
+
__openapi_doc_version__,
|
|
7
|
+
__gen_version__,
|
|
8
|
+
__user_agent__,
|
|
9
|
+
)
|
|
4
10
|
from .sdk import *
|
|
5
11
|
from .sdkconfiguration import *
|
|
6
12
|
from .models import *
|
|
7
13
|
|
|
8
14
|
|
|
9
15
|
VERSION: str = __version__
|
|
16
|
+
OPENAPI_DOC_VERSION = __openapi_doc_version__
|
|
17
|
+
SPEAKEASY_GENERATOR_VERSION = __gen_version__
|
|
18
|
+
USER_AGENT = __user_agent__
|
mistralai/_version.py
CHANGED
|
@@ -3,7 +3,10 @@
|
|
|
3
3
|
import importlib.metadata
|
|
4
4
|
|
|
5
5
|
__title__: str = "mistralai"
|
|
6
|
-
__version__: str = "1.
|
|
6
|
+
__version__: str = "1.4.0"
|
|
7
|
+
__openapi_doc_version__: str = "0.0.2"
|
|
8
|
+
__gen_version__: str = "2.493.32"
|
|
9
|
+
__user_agent__: str = "speakeasy-sdk/python 1.4.0 2.493.32 0.0.2 mistralai"
|
|
7
10
|
|
|
8
11
|
try:
|
|
9
12
|
if __package__ is not None:
|
mistralai/agents.py
CHANGED
|
@@ -43,11 +43,14 @@ class Agents(BaseSDK):
|
|
|
43
43
|
presence_penalty: Optional[float] = None,
|
|
44
44
|
frequency_penalty: Optional[float] = None,
|
|
45
45
|
n: OptionalNullable[int] = UNSET,
|
|
46
|
+
prediction: Optional[
|
|
47
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
48
|
+
] = None,
|
|
46
49
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
47
50
|
server_url: Optional[str] = None,
|
|
48
51
|
timeout_ms: Optional[int] = None,
|
|
49
52
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
50
|
-
) ->
|
|
53
|
+
) -> models.ChatCompletionResponse:
|
|
51
54
|
r"""Agents Completion
|
|
52
55
|
|
|
53
56
|
:param messages: The prompt(s) to generate completions for, encoded as a list of dict with role and content.
|
|
@@ -62,6 +65,7 @@ class Agents(BaseSDK):
|
|
|
62
65
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
63
66
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
64
67
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
68
|
+
:param prediction:
|
|
65
69
|
:param retries: Override the default retry configuration for this method
|
|
66
70
|
:param server_url: Override the default server URL for this method
|
|
67
71
|
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
|
|
@@ -93,6 +97,9 @@ class Agents(BaseSDK):
|
|
|
93
97
|
presence_penalty=presence_penalty,
|
|
94
98
|
frequency_penalty=frequency_penalty,
|
|
95
99
|
n=n,
|
|
100
|
+
prediction=utils.get_pydantic_model(
|
|
101
|
+
prediction, Optional[models.Prediction]
|
|
102
|
+
),
|
|
96
103
|
agent_id=agent_id,
|
|
97
104
|
)
|
|
98
105
|
|
|
@@ -138,13 +145,16 @@ class Agents(BaseSDK):
|
|
|
138
145
|
|
|
139
146
|
data: Any = None
|
|
140
147
|
if utils.match_response(http_res, "200", "application/json"):
|
|
141
|
-
return utils.unmarshal_json(
|
|
142
|
-
http_res.text, Optional[models.ChatCompletionResponse]
|
|
143
|
-
)
|
|
148
|
+
return utils.unmarshal_json(http_res.text, models.ChatCompletionResponse)
|
|
144
149
|
if utils.match_response(http_res, "422", "application/json"):
|
|
145
150
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
146
151
|
raise models.HTTPValidationError(data=data)
|
|
147
|
-
if utils.match_response(http_res,
|
|
152
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
153
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
154
|
+
raise models.SDKError(
|
|
155
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
156
|
+
)
|
|
157
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
148
158
|
http_res_text = utils.stream_to_text(http_res)
|
|
149
159
|
raise models.SDKError(
|
|
150
160
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -191,11 +201,14 @@ class Agents(BaseSDK):
|
|
|
191
201
|
presence_penalty: Optional[float] = None,
|
|
192
202
|
frequency_penalty: Optional[float] = None,
|
|
193
203
|
n: OptionalNullable[int] = UNSET,
|
|
204
|
+
prediction: Optional[
|
|
205
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
206
|
+
] = None,
|
|
194
207
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
195
208
|
server_url: Optional[str] = None,
|
|
196
209
|
timeout_ms: Optional[int] = None,
|
|
197
210
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
198
|
-
) ->
|
|
211
|
+
) -> models.ChatCompletionResponse:
|
|
199
212
|
r"""Agents Completion
|
|
200
213
|
|
|
201
214
|
:param messages: The prompt(s) to generate completions for, encoded as a list of dict with role and content.
|
|
@@ -210,6 +223,7 @@ class Agents(BaseSDK):
|
|
|
210
223
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
211
224
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
212
225
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
226
|
+
:param prediction:
|
|
213
227
|
:param retries: Override the default retry configuration for this method
|
|
214
228
|
:param server_url: Override the default server URL for this method
|
|
215
229
|
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
|
|
@@ -241,6 +255,9 @@ class Agents(BaseSDK):
|
|
|
241
255
|
presence_penalty=presence_penalty,
|
|
242
256
|
frequency_penalty=frequency_penalty,
|
|
243
257
|
n=n,
|
|
258
|
+
prediction=utils.get_pydantic_model(
|
|
259
|
+
prediction, Optional[models.Prediction]
|
|
260
|
+
),
|
|
244
261
|
agent_id=agent_id,
|
|
245
262
|
)
|
|
246
263
|
|
|
@@ -286,13 +303,16 @@ class Agents(BaseSDK):
|
|
|
286
303
|
|
|
287
304
|
data: Any = None
|
|
288
305
|
if utils.match_response(http_res, "200", "application/json"):
|
|
289
|
-
return utils.unmarshal_json(
|
|
290
|
-
http_res.text, Optional[models.ChatCompletionResponse]
|
|
291
|
-
)
|
|
306
|
+
return utils.unmarshal_json(http_res.text, models.ChatCompletionResponse)
|
|
292
307
|
if utils.match_response(http_res, "422", "application/json"):
|
|
293
308
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
294
309
|
raise models.HTTPValidationError(data=data)
|
|
295
|
-
if utils.match_response(http_res,
|
|
310
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
311
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
312
|
+
raise models.SDKError(
|
|
313
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
314
|
+
)
|
|
315
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
296
316
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
297
317
|
raise models.SDKError(
|
|
298
318
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -339,11 +359,14 @@ class Agents(BaseSDK):
|
|
|
339
359
|
presence_penalty: Optional[float] = None,
|
|
340
360
|
frequency_penalty: Optional[float] = None,
|
|
341
361
|
n: OptionalNullable[int] = UNSET,
|
|
362
|
+
prediction: Optional[
|
|
363
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
364
|
+
] = None,
|
|
342
365
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
343
366
|
server_url: Optional[str] = None,
|
|
344
367
|
timeout_ms: Optional[int] = None,
|
|
345
368
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
346
|
-
) ->
|
|
369
|
+
) -> eventstreaming.EventStream[models.CompletionEvent]:
|
|
347
370
|
r"""Stream Agents completion
|
|
348
371
|
|
|
349
372
|
Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
|
|
@@ -360,6 +383,7 @@ class Agents(BaseSDK):
|
|
|
360
383
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
361
384
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
362
385
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
386
|
+
:param prediction:
|
|
363
387
|
:param retries: Override the default retry configuration for this method
|
|
364
388
|
:param server_url: Override the default server URL for this method
|
|
365
389
|
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
|
|
@@ -391,6 +415,9 @@ class Agents(BaseSDK):
|
|
|
391
415
|
presence_penalty=presence_penalty,
|
|
392
416
|
frequency_penalty=frequency_penalty,
|
|
393
417
|
n=n,
|
|
418
|
+
prediction=utils.get_pydantic_model(
|
|
419
|
+
prediction, Optional[models.Prediction]
|
|
420
|
+
),
|
|
394
421
|
agent_id=agent_id,
|
|
395
422
|
)
|
|
396
423
|
|
|
@@ -446,7 +473,12 @@ class Agents(BaseSDK):
|
|
|
446
473
|
http_res_text = utils.stream_to_text(http_res)
|
|
447
474
|
data = utils.unmarshal_json(http_res_text, models.HTTPValidationErrorData)
|
|
448
475
|
raise models.HTTPValidationError(data=data)
|
|
449
|
-
if utils.match_response(http_res,
|
|
476
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
477
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
478
|
+
raise models.SDKError(
|
|
479
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
480
|
+
)
|
|
481
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
450
482
|
http_res_text = utils.stream_to_text(http_res)
|
|
451
483
|
raise models.SDKError(
|
|
452
484
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -493,11 +525,14 @@ class Agents(BaseSDK):
|
|
|
493
525
|
presence_penalty: Optional[float] = None,
|
|
494
526
|
frequency_penalty: Optional[float] = None,
|
|
495
527
|
n: OptionalNullable[int] = UNSET,
|
|
528
|
+
prediction: Optional[
|
|
529
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
530
|
+
] = None,
|
|
496
531
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
497
532
|
server_url: Optional[str] = None,
|
|
498
533
|
timeout_ms: Optional[int] = None,
|
|
499
534
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
500
|
-
) ->
|
|
535
|
+
) -> eventstreaming.EventStreamAsync[models.CompletionEvent]:
|
|
501
536
|
r"""Stream Agents completion
|
|
502
537
|
|
|
503
538
|
Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
|
|
@@ -514,6 +549,7 @@ class Agents(BaseSDK):
|
|
|
514
549
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
515
550
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
516
551
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
552
|
+
:param prediction:
|
|
517
553
|
:param retries: Override the default retry configuration for this method
|
|
518
554
|
:param server_url: Override the default server URL for this method
|
|
519
555
|
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
|
|
@@ -545,6 +581,9 @@ class Agents(BaseSDK):
|
|
|
545
581
|
presence_penalty=presence_penalty,
|
|
546
582
|
frequency_penalty=frequency_penalty,
|
|
547
583
|
n=n,
|
|
584
|
+
prediction=utils.get_pydantic_model(
|
|
585
|
+
prediction, Optional[models.Prediction]
|
|
586
|
+
),
|
|
548
587
|
agent_id=agent_id,
|
|
549
588
|
)
|
|
550
589
|
|
|
@@ -600,7 +639,12 @@ class Agents(BaseSDK):
|
|
|
600
639
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
601
640
|
data = utils.unmarshal_json(http_res_text, models.HTTPValidationErrorData)
|
|
602
641
|
raise models.HTTPValidationError(data=data)
|
|
603
|
-
if utils.match_response(http_res,
|
|
642
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
643
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
644
|
+
raise models.SDKError(
|
|
645
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
646
|
+
)
|
|
647
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
604
648
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
605
649
|
raise models.SDKError(
|
|
606
650
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
mistralai/chat.py
CHANGED
|
@@ -37,12 +37,15 @@ class Chat(BaseSDK):
|
|
|
37
37
|
presence_penalty: Optional[float] = None,
|
|
38
38
|
frequency_penalty: Optional[float] = None,
|
|
39
39
|
n: OptionalNullable[int] = UNSET,
|
|
40
|
+
prediction: Optional[
|
|
41
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
42
|
+
] = None,
|
|
40
43
|
safe_prompt: Optional[bool] = None,
|
|
41
44
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
42
45
|
server_url: Optional[str] = None,
|
|
43
46
|
timeout_ms: Optional[int] = None,
|
|
44
47
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
45
|
-
) ->
|
|
48
|
+
) -> models.ChatCompletionResponse:
|
|
46
49
|
r"""Chat Completion
|
|
47
50
|
|
|
48
51
|
:param model: ID of the model to use. You can use the [List Available Models](/api/#tag/models/operation/list_models_v1_models_get) API to see all of your available models, or see our [Model overview](/models) for model descriptions.
|
|
@@ -59,6 +62,7 @@ class Chat(BaseSDK):
|
|
|
59
62
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
60
63
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
61
64
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
65
|
+
:param prediction:
|
|
62
66
|
:param safe_prompt: Whether to inject a safety prompt before all conversations.
|
|
63
67
|
:param retries: Override the default retry configuration for this method
|
|
64
68
|
:param server_url: Override the default server URL for this method
|
|
@@ -92,6 +96,9 @@ class Chat(BaseSDK):
|
|
|
92
96
|
presence_penalty=presence_penalty,
|
|
93
97
|
frequency_penalty=frequency_penalty,
|
|
94
98
|
n=n,
|
|
99
|
+
prediction=utils.get_pydantic_model(
|
|
100
|
+
prediction, Optional[models.Prediction]
|
|
101
|
+
),
|
|
95
102
|
safe_prompt=safe_prompt,
|
|
96
103
|
)
|
|
97
104
|
|
|
@@ -137,13 +144,16 @@ class Chat(BaseSDK):
|
|
|
137
144
|
|
|
138
145
|
data: Any = None
|
|
139
146
|
if utils.match_response(http_res, "200", "application/json"):
|
|
140
|
-
return utils.unmarshal_json(
|
|
141
|
-
http_res.text, Optional[models.ChatCompletionResponse]
|
|
142
|
-
)
|
|
147
|
+
return utils.unmarshal_json(http_res.text, models.ChatCompletionResponse)
|
|
143
148
|
if utils.match_response(http_res, "422", "application/json"):
|
|
144
149
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
145
150
|
raise models.HTTPValidationError(data=data)
|
|
146
|
-
if utils.match_response(http_res,
|
|
151
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
152
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
153
|
+
raise models.SDKError(
|
|
154
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
155
|
+
)
|
|
156
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
147
157
|
http_res_text = utils.stream_to_text(http_res)
|
|
148
158
|
raise models.SDKError(
|
|
149
159
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -184,12 +194,15 @@ class Chat(BaseSDK):
|
|
|
184
194
|
presence_penalty: Optional[float] = None,
|
|
185
195
|
frequency_penalty: Optional[float] = None,
|
|
186
196
|
n: OptionalNullable[int] = UNSET,
|
|
197
|
+
prediction: Optional[
|
|
198
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
199
|
+
] = None,
|
|
187
200
|
safe_prompt: Optional[bool] = None,
|
|
188
201
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
189
202
|
server_url: Optional[str] = None,
|
|
190
203
|
timeout_ms: Optional[int] = None,
|
|
191
204
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
192
|
-
) ->
|
|
205
|
+
) -> models.ChatCompletionResponse:
|
|
193
206
|
r"""Chat Completion
|
|
194
207
|
|
|
195
208
|
:param model: ID of the model to use. You can use the [List Available Models](/api/#tag/models/operation/list_models_v1_models_get) API to see all of your available models, or see our [Model overview](/models) for model descriptions.
|
|
@@ -206,6 +219,7 @@ class Chat(BaseSDK):
|
|
|
206
219
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
207
220
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
208
221
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
222
|
+
:param prediction:
|
|
209
223
|
:param safe_prompt: Whether to inject a safety prompt before all conversations.
|
|
210
224
|
:param retries: Override the default retry configuration for this method
|
|
211
225
|
:param server_url: Override the default server URL for this method
|
|
@@ -239,6 +253,9 @@ class Chat(BaseSDK):
|
|
|
239
253
|
presence_penalty=presence_penalty,
|
|
240
254
|
frequency_penalty=frequency_penalty,
|
|
241
255
|
n=n,
|
|
256
|
+
prediction=utils.get_pydantic_model(
|
|
257
|
+
prediction, Optional[models.Prediction]
|
|
258
|
+
),
|
|
242
259
|
safe_prompt=safe_prompt,
|
|
243
260
|
)
|
|
244
261
|
|
|
@@ -284,13 +301,16 @@ class Chat(BaseSDK):
|
|
|
284
301
|
|
|
285
302
|
data: Any = None
|
|
286
303
|
if utils.match_response(http_res, "200", "application/json"):
|
|
287
|
-
return utils.unmarshal_json(
|
|
288
|
-
http_res.text, Optional[models.ChatCompletionResponse]
|
|
289
|
-
)
|
|
304
|
+
return utils.unmarshal_json(http_res.text, models.ChatCompletionResponse)
|
|
290
305
|
if utils.match_response(http_res, "422", "application/json"):
|
|
291
306
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
292
307
|
raise models.HTTPValidationError(data=data)
|
|
293
|
-
if utils.match_response(http_res,
|
|
308
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
309
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
310
|
+
raise models.SDKError(
|
|
311
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
312
|
+
)
|
|
313
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
294
314
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
295
315
|
raise models.SDKError(
|
|
296
316
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -339,12 +359,15 @@ class Chat(BaseSDK):
|
|
|
339
359
|
presence_penalty: Optional[float] = None,
|
|
340
360
|
frequency_penalty: Optional[float] = None,
|
|
341
361
|
n: OptionalNullable[int] = UNSET,
|
|
362
|
+
prediction: Optional[
|
|
363
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
364
|
+
] = None,
|
|
342
365
|
safe_prompt: Optional[bool] = None,
|
|
343
366
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
344
367
|
server_url: Optional[str] = None,
|
|
345
368
|
timeout_ms: Optional[int] = None,
|
|
346
369
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
347
|
-
) ->
|
|
370
|
+
) -> eventstreaming.EventStream[models.CompletionEvent]:
|
|
348
371
|
r"""Stream chat completion
|
|
349
372
|
|
|
350
373
|
Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
|
|
@@ -363,6 +386,7 @@ class Chat(BaseSDK):
|
|
|
363
386
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
364
387
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
365
388
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
389
|
+
:param prediction:
|
|
366
390
|
:param safe_prompt: Whether to inject a safety prompt before all conversations.
|
|
367
391
|
:param retries: Override the default retry configuration for this method
|
|
368
392
|
:param server_url: Override the default server URL for this method
|
|
@@ -398,6 +422,9 @@ class Chat(BaseSDK):
|
|
|
398
422
|
presence_penalty=presence_penalty,
|
|
399
423
|
frequency_penalty=frequency_penalty,
|
|
400
424
|
n=n,
|
|
425
|
+
prediction=utils.get_pydantic_model(
|
|
426
|
+
prediction, Optional[models.Prediction]
|
|
427
|
+
),
|
|
401
428
|
safe_prompt=safe_prompt,
|
|
402
429
|
)
|
|
403
430
|
|
|
@@ -453,7 +480,12 @@ class Chat(BaseSDK):
|
|
|
453
480
|
http_res_text = utils.stream_to_text(http_res)
|
|
454
481
|
data = utils.unmarshal_json(http_res_text, models.HTTPValidationErrorData)
|
|
455
482
|
raise models.HTTPValidationError(data=data)
|
|
456
|
-
if utils.match_response(http_res,
|
|
483
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
484
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
485
|
+
raise models.SDKError(
|
|
486
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
487
|
+
)
|
|
488
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
457
489
|
http_res_text = utils.stream_to_text(http_res)
|
|
458
490
|
raise models.SDKError(
|
|
459
491
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -502,12 +534,15 @@ class Chat(BaseSDK):
|
|
|
502
534
|
presence_penalty: Optional[float] = None,
|
|
503
535
|
frequency_penalty: Optional[float] = None,
|
|
504
536
|
n: OptionalNullable[int] = UNSET,
|
|
537
|
+
prediction: Optional[
|
|
538
|
+
Union[models.Prediction, models.PredictionTypedDict]
|
|
539
|
+
] = None,
|
|
505
540
|
safe_prompt: Optional[bool] = None,
|
|
506
541
|
retries: OptionalNullable[utils.RetryConfig] = UNSET,
|
|
507
542
|
server_url: Optional[str] = None,
|
|
508
543
|
timeout_ms: Optional[int] = None,
|
|
509
544
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
510
|
-
) ->
|
|
545
|
+
) -> eventstreaming.EventStreamAsync[models.CompletionEvent]:
|
|
511
546
|
r"""Stream chat completion
|
|
512
547
|
|
|
513
548
|
Mistral AI provides the ability to stream responses back to a client in order to allow partial results for certain requests. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Otherwise, the server will hold the request open until the timeout or until completion, with the response containing the full result as JSON.
|
|
@@ -526,6 +561,7 @@ class Chat(BaseSDK):
|
|
|
526
561
|
:param presence_penalty: presence_penalty determines how much the model penalizes the repetition of words or phrases. A higher presence penalty encourages the model to use a wider variety of words and phrases, making the output more diverse and creative.
|
|
527
562
|
:param frequency_penalty: frequency_penalty penalizes the repetition of words based on their frequency in the generated text. A higher frequency penalty discourages the model from repeating words that have already appeared frequently in the output, promoting diversity and reducing repetition.
|
|
528
563
|
:param n: Number of completions to return for each request, input tokens are only billed once.
|
|
564
|
+
:param prediction:
|
|
529
565
|
:param safe_prompt: Whether to inject a safety prompt before all conversations.
|
|
530
566
|
:param retries: Override the default retry configuration for this method
|
|
531
567
|
:param server_url: Override the default server URL for this method
|
|
@@ -561,6 +597,9 @@ class Chat(BaseSDK):
|
|
|
561
597
|
presence_penalty=presence_penalty,
|
|
562
598
|
frequency_penalty=frequency_penalty,
|
|
563
599
|
n=n,
|
|
600
|
+
prediction=utils.get_pydantic_model(
|
|
601
|
+
prediction, Optional[models.Prediction]
|
|
602
|
+
),
|
|
564
603
|
safe_prompt=safe_prompt,
|
|
565
604
|
)
|
|
566
605
|
|
|
@@ -616,7 +655,12 @@ class Chat(BaseSDK):
|
|
|
616
655
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
617
656
|
data = utils.unmarshal_json(http_res_text, models.HTTPValidationErrorData)
|
|
618
657
|
raise models.HTTPValidationError(data=data)
|
|
619
|
-
if utils.match_response(http_res,
|
|
658
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
659
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
660
|
+
raise models.SDKError(
|
|
661
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
662
|
+
)
|
|
663
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
620
664
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
621
665
|
raise models.SDKError(
|
|
622
666
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
mistralai/classifiers.py
CHANGED
|
@@ -23,7 +23,7 @@ class Classifiers(BaseSDK):
|
|
|
23
23
|
server_url: Optional[str] = None,
|
|
24
24
|
timeout_ms: Optional[int] = None,
|
|
25
25
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
26
|
-
) ->
|
|
26
|
+
) -> models.ClassificationResponse:
|
|
27
27
|
r"""Moderations
|
|
28
28
|
|
|
29
29
|
:param inputs: Text to classify.
|
|
@@ -88,13 +88,16 @@ class Classifiers(BaseSDK):
|
|
|
88
88
|
|
|
89
89
|
data: Any = None
|
|
90
90
|
if utils.match_response(http_res, "200", "application/json"):
|
|
91
|
-
return utils.unmarshal_json(
|
|
92
|
-
http_res.text, Optional[models.ClassificationResponse]
|
|
93
|
-
)
|
|
91
|
+
return utils.unmarshal_json(http_res.text, models.ClassificationResponse)
|
|
94
92
|
if utils.match_response(http_res, "422", "application/json"):
|
|
95
93
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
96
94
|
raise models.HTTPValidationError(data=data)
|
|
97
|
-
if utils.match_response(http_res,
|
|
95
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
96
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
97
|
+
raise models.SDKError(
|
|
98
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
99
|
+
)
|
|
100
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
98
101
|
http_res_text = utils.stream_to_text(http_res)
|
|
99
102
|
raise models.SDKError(
|
|
100
103
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -121,7 +124,7 @@ class Classifiers(BaseSDK):
|
|
|
121
124
|
server_url: Optional[str] = None,
|
|
122
125
|
timeout_ms: Optional[int] = None,
|
|
123
126
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
124
|
-
) ->
|
|
127
|
+
) -> models.ClassificationResponse:
|
|
125
128
|
r"""Moderations
|
|
126
129
|
|
|
127
130
|
:param inputs: Text to classify.
|
|
@@ -186,13 +189,16 @@ class Classifiers(BaseSDK):
|
|
|
186
189
|
|
|
187
190
|
data: Any = None
|
|
188
191
|
if utils.match_response(http_res, "200", "application/json"):
|
|
189
|
-
return utils.unmarshal_json(
|
|
190
|
-
http_res.text, Optional[models.ClassificationResponse]
|
|
191
|
-
)
|
|
192
|
+
return utils.unmarshal_json(http_res.text, models.ClassificationResponse)
|
|
192
193
|
if utils.match_response(http_res, "422", "application/json"):
|
|
193
194
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
194
195
|
raise models.HTTPValidationError(data=data)
|
|
195
|
-
if utils.match_response(http_res,
|
|
196
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
197
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
198
|
+
raise models.SDKError(
|
|
199
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
200
|
+
)
|
|
201
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
196
202
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
197
203
|
raise models.SDKError(
|
|
198
204
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -219,7 +225,7 @@ class Classifiers(BaseSDK):
|
|
|
219
225
|
server_url: Optional[str] = None,
|
|
220
226
|
timeout_ms: Optional[int] = None,
|
|
221
227
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
222
|
-
) ->
|
|
228
|
+
) -> models.ClassificationResponse:
|
|
223
229
|
r"""Moderations Chat
|
|
224
230
|
|
|
225
231
|
:param inputs: Chat to classify
|
|
@@ -286,13 +292,16 @@ class Classifiers(BaseSDK):
|
|
|
286
292
|
|
|
287
293
|
data: Any = None
|
|
288
294
|
if utils.match_response(http_res, "200", "application/json"):
|
|
289
|
-
return utils.unmarshal_json(
|
|
290
|
-
http_res.text, Optional[models.ClassificationResponse]
|
|
291
|
-
)
|
|
295
|
+
return utils.unmarshal_json(http_res.text, models.ClassificationResponse)
|
|
292
296
|
if utils.match_response(http_res, "422", "application/json"):
|
|
293
297
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
294
298
|
raise models.HTTPValidationError(data=data)
|
|
295
|
-
if utils.match_response(http_res,
|
|
299
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
300
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
301
|
+
raise models.SDKError(
|
|
302
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
303
|
+
)
|
|
304
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
296
305
|
http_res_text = utils.stream_to_text(http_res)
|
|
297
306
|
raise models.SDKError(
|
|
298
307
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -319,7 +328,7 @@ class Classifiers(BaseSDK):
|
|
|
319
328
|
server_url: Optional[str] = None,
|
|
320
329
|
timeout_ms: Optional[int] = None,
|
|
321
330
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
322
|
-
) ->
|
|
331
|
+
) -> models.ClassificationResponse:
|
|
323
332
|
r"""Moderations Chat
|
|
324
333
|
|
|
325
334
|
:param inputs: Chat to classify
|
|
@@ -386,13 +395,16 @@ class Classifiers(BaseSDK):
|
|
|
386
395
|
|
|
387
396
|
data: Any = None
|
|
388
397
|
if utils.match_response(http_res, "200", "application/json"):
|
|
389
|
-
return utils.unmarshal_json(
|
|
390
|
-
http_res.text, Optional[models.ClassificationResponse]
|
|
391
|
-
)
|
|
398
|
+
return utils.unmarshal_json(http_res.text, models.ClassificationResponse)
|
|
392
399
|
if utils.match_response(http_res, "422", "application/json"):
|
|
393
400
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
394
401
|
raise models.HTTPValidationError(data=data)
|
|
395
|
-
if utils.match_response(http_res,
|
|
402
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
403
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
404
|
+
raise models.SDKError(
|
|
405
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
406
|
+
)
|
|
407
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
396
408
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
397
409
|
raise models.SDKError(
|
|
398
410
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
mistralai/embeddings.py
CHANGED
|
@@ -21,7 +21,7 @@ class Embeddings(BaseSDK):
|
|
|
21
21
|
server_url: Optional[str] = None,
|
|
22
22
|
timeout_ms: Optional[int] = None,
|
|
23
23
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
24
|
-
) ->
|
|
24
|
+
) -> models.EmbeddingResponse:
|
|
25
25
|
r"""Embeddings
|
|
26
26
|
|
|
27
27
|
Embeddings
|
|
@@ -90,13 +90,16 @@ class Embeddings(BaseSDK):
|
|
|
90
90
|
|
|
91
91
|
data: Any = None
|
|
92
92
|
if utils.match_response(http_res, "200", "application/json"):
|
|
93
|
-
return utils.unmarshal_json(
|
|
94
|
-
http_res.text, Optional[models.EmbeddingResponse]
|
|
95
|
-
)
|
|
93
|
+
return utils.unmarshal_json(http_res.text, models.EmbeddingResponse)
|
|
96
94
|
if utils.match_response(http_res, "422", "application/json"):
|
|
97
95
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
98
96
|
raise models.HTTPValidationError(data=data)
|
|
99
|
-
if utils.match_response(http_res,
|
|
97
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
98
|
+
http_res_text = utils.stream_to_text(http_res)
|
|
99
|
+
raise models.SDKError(
|
|
100
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
101
|
+
)
|
|
102
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
100
103
|
http_res_text = utils.stream_to_text(http_res)
|
|
101
104
|
raise models.SDKError(
|
|
102
105
|
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
@@ -121,7 +124,7 @@ class Embeddings(BaseSDK):
|
|
|
121
124
|
server_url: Optional[str] = None,
|
|
122
125
|
timeout_ms: Optional[int] = None,
|
|
123
126
|
http_headers: Optional[Mapping[str, str]] = None,
|
|
124
|
-
) ->
|
|
127
|
+
) -> models.EmbeddingResponse:
|
|
125
128
|
r"""Embeddings
|
|
126
129
|
|
|
127
130
|
Embeddings
|
|
@@ -190,13 +193,16 @@ class Embeddings(BaseSDK):
|
|
|
190
193
|
|
|
191
194
|
data: Any = None
|
|
192
195
|
if utils.match_response(http_res, "200", "application/json"):
|
|
193
|
-
return utils.unmarshal_json(
|
|
194
|
-
http_res.text, Optional[models.EmbeddingResponse]
|
|
195
|
-
)
|
|
196
|
+
return utils.unmarshal_json(http_res.text, models.EmbeddingResponse)
|
|
196
197
|
if utils.match_response(http_res, "422", "application/json"):
|
|
197
198
|
data = utils.unmarshal_json(http_res.text, models.HTTPValidationErrorData)
|
|
198
199
|
raise models.HTTPValidationError(data=data)
|
|
199
|
-
if utils.match_response(http_res,
|
|
200
|
+
if utils.match_response(http_res, "4XX", "*"):
|
|
201
|
+
http_res_text = await utils.stream_to_text_async(http_res)
|
|
202
|
+
raise models.SDKError(
|
|
203
|
+
"API error occurred", http_res.status_code, http_res_text, http_res
|
|
204
|
+
)
|
|
205
|
+
if utils.match_response(http_res, "5XX", "*"):
|
|
200
206
|
http_res_text = await utils.stream_to_text_async(http_res)
|
|
201
207
|
raise models.SDKError(
|
|
202
208
|
"API error occurred", http_res.status_code, http_res_text, http_res
|