File size: 17,951 Bytes
395201c |
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 |
from typing import Optional, Union, Any
import types, requests
from .base import BaseLLM
from litellm.utils import ModelResponse, Choices, Message, CustomStreamWrapper, convert_to_model_response_object
from typing import Callable, Optional
from litellm import OpenAIConfig
import litellm, json
import httpx
from openai import AzureOpenAI, AsyncAzureOpenAI
class AzureOpenAIError(Exception):
def __init__(self, status_code, message, request: Optional[httpx.Request]=None, response: Optional[httpx.Response]=None):
self.status_code = status_code
self.message = message
if request:
self.request = request
else:
self.request = httpx.Request(method="POST", url="https://api.openai.com/v1")
if response:
self.response = response
else:
self.response = httpx.Response(status_code=status_code, request=self.request)
super().__init__(
self.message
) # Call the base class constructor with the parameters it needs
class AzureOpenAIConfig(OpenAIConfig):
"""
Reference: https://platform.openai.com/docs/api-reference/chat/create
The class `AzureOpenAIConfig` provides configuration for the OpenAI's Chat API interface, for use with Azure. It inherits from `OpenAIConfig`. Below are the parameters::
- `frequency_penalty` (number or null): Defaults to 0. Allows a value between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, thereby minimizing repetition.
- `function_call` (string or object): This optional parameter controls how the model calls functions.
- `functions` (array): An optional parameter. It is a list of functions for which the model may generate JSON inputs.
- `logit_bias` (map): This optional parameter modifies the likelihood of specified tokens appearing in the completion.
- `max_tokens` (integer or null): This optional parameter helps to set the maximum number of tokens to generate in the chat completion.
- `n` (integer or null): This optional parameter helps to set how many chat completion choices to generate for each input message.
- `presence_penalty` (number or null): Defaults to 0. It penalizes new tokens based on if they appear in the text so far, hence increasing the model's likelihood to talk about new topics.
- `stop` (string / array / null): Specifies up to 4 sequences where the API will stop generating further tokens.
- `temperature` (number or null): Defines the sampling temperature to use, varying between 0 and 2.
- `top_p` (number or null): An alternative to sampling with temperature, used for nucleus sampling.
"""
def __init__(self,
frequency_penalty: Optional[int] = None,
function_call: Optional[Union[str, dict]]= None,
functions: Optional[list]= None,
logit_bias: Optional[dict]= None,
max_tokens: Optional[int]= None,
n: Optional[int]= None,
presence_penalty: Optional[int]= None,
stop: Optional[Union[str,list]]=None,
temperature: Optional[int]= None,
top_p: Optional[int]= None) -> None:
super().__init__(frequency_penalty,
function_call,
functions,
logit_bias,
max_tokens,
n,
presence_penalty,
stop,
temperature,
top_p)
class AzureChatCompletion(BaseLLM):
def __init__(self) -> None:
super().__init__()
def validate_environment(self, api_key, azure_ad_token):
headers = {
"content-type": "application/json",
}
if api_key is not None:
headers["api-key"] = api_key
elif azure_ad_token is not None:
headers["Authorization"] = f"Bearer {azure_ad_token}"
return headers
def completion(self,
model: str,
messages: list,
model_response: ModelResponse,
api_key: str,
api_base: str,
api_version: str,
api_type: str,
azure_ad_token: str,
print_verbose: Callable,
timeout,
logging_obj,
optional_params,
litellm_params,
logger_fn,
acompletion: bool = False,
headers: Optional[dict]=None,
client = None,
):
super().completion()
exception_mapping_worked = False
try:
if model is None or messages is None:
raise AzureOpenAIError(status_code=422, message=f"Missing model or messages")
max_retries = optional_params.pop("max_retries", 2)
### CHECK IF CLOUDFLARE AI GATEWAY ###
### if so - set the model as part of the base url
if "gateway.ai.cloudflare.com" in api_base:
## build base url - assume api base includes resource name
if client is None:
if not api_base.endswith("/"):
api_base += "/"
api_base += f"{model}"
azure_client_params = {
"api_version": api_version,
"base_url": f"{api_base}",
"http_client": litellm.client_session,
"max_retries": max_retries,
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if acompletion is True:
client = AsyncAzureOpenAI(**azure_client_params)
else:
client = AzureOpenAI(**azure_client_params)
data = {
"model": None,
"messages": messages,
**optional_params
}
else:
data = {
"model": model, # type: ignore
"messages": messages,
**optional_params
}
## LOGGING
logging_obj.pre_call(
input=messages,
api_key=api_key,
additional_args={
"headers": {
"api_key": api_key,
"azure_ad_token": azure_ad_token
},
"api_version": api_version,
"api_base": api_base,
"complete_input_dict": data,
},
)
if acompletion is True:
if optional_params.get("stream", False):
return self.async_streaming(logging_obj=logging_obj, api_base=api_base, data=data, model=model, api_key=api_key, api_version=api_version, azure_ad_token=azure_ad_token, timeout=timeout, client=client)
else:
return self.acompletion(api_base=api_base, data=data, model_response=model_response, api_key=api_key, api_version=api_version, model=model, azure_ad_token=azure_ad_token, timeout=timeout, client=client)
elif "stream" in optional_params and optional_params["stream"] == True:
return self.streaming(logging_obj=logging_obj, api_base=api_base, data=data, model=model, api_key=api_key, api_version=api_version, azure_ad_token=azure_ad_token, timeout=timeout, client=client)
else:
if not isinstance(max_retries, int):
raise AzureOpenAIError(status_code=422, message="max retries must be an int")
# init AzureOpenAI Client
azure_client_params = {
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"http_client": litellm.client_session,
"max_retries": max_retries,
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if client is None:
azure_client = AzureOpenAI(**azure_client_params)
else:
azure_client = client
response = azure_client.chat.completions.create(**data) # type: ignore
response.model = "azure/" + str(response.model)
return convert_to_model_response_object(response_object=json.loads(response.model_dump_json()), model_response_object=model_response)
except AzureOpenAIError as e:
exception_mapping_worked = True
raise e
except Exception as e:
raise e
async def acompletion(self,
api_key: str,
api_version: str,
model: str,
api_base: str,
data: dict,
timeout: Any,
model_response: ModelResponse,
azure_ad_token: Optional[str]=None,
client = None, # this is the AsyncAzureOpenAI
):
response = None
try:
max_retries = data.pop("max_retries", 2)
if not isinstance(max_retries, int):
raise AzureOpenAIError(status_code=422, message="max retries must be an int")
# init AzureOpenAI Client
azure_client_params = {
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"http_client": litellm.client_session,
"max_retries": max_retries,
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if client is None:
azure_client = AsyncAzureOpenAI(**azure_client_params)
else:
azure_client = client
response = await azure_client.chat.completions.create(**data)
return convert_to_model_response_object(response_object=json.loads(response.model_dump_json()), model_response_object=model_response)
except AzureOpenAIError as e:
exception_mapping_worked = True
raise e
except Exception as e:
raise e
def streaming(self,
logging_obj,
api_base: str,
api_key: str,
api_version: str,
data: dict,
model: str,
timeout: Any,
azure_ad_token: Optional[str]=None,
client=None,
):
max_retries = data.pop("max_retries", 2)
if not isinstance(max_retries, int):
raise AzureOpenAIError(status_code=422, message="max retries must be an int")
# init AzureOpenAI Client
azure_client_params = {
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"http_client": litellm.client_session,
"max_retries": max_retries,
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if client is None:
azure_client = AzureOpenAI(**azure_client_params)
else:
azure_client = client
response = azure_client.chat.completions.create(**data)
streamwrapper = CustomStreamWrapper(completion_stream=response, model=model, custom_llm_provider="azure",logging_obj=logging_obj)
return streamwrapper
async def async_streaming(self,
logging_obj,
api_base: str,
api_key: str,
api_version: str,
data: dict,
model: str,
timeout: Any,
azure_ad_token: Optional[str]=None,
client = None,
):
# init AzureOpenAI Client
azure_client_params = {
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"http_client": litellm.client_session,
"max_retries": data.pop("max_retries", 2),
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if client is None:
azure_client = AsyncAzureOpenAI(**azure_client_params)
else:
azure_client = client
response = await azure_client.chat.completions.create(**data)
streamwrapper = CustomStreamWrapper(completion_stream=response, model=model, custom_llm_provider="azure",logging_obj=logging_obj)
async for transformed_chunk in streamwrapper:
yield transformed_chunk
async def aembedding(
self,
data: dict,
model_response: ModelResponse,
azure_client_params: dict,
client=None,
):
response = None
try:
if client is None:
openai_aclient = AsyncAzureOpenAI(**azure_client_params)
else:
openai_aclient = client
response = await openai_aclient.embeddings.create(**data)
return convert_to_model_response_object(response_object=json.loads(response.model_dump_json()), model_response_object=model_response, response_type="embedding")
except Exception as e:
raise e
def embedding(self,
model: str,
input: list,
api_key: str,
api_base: str,
api_version: str,
timeout: float,
logging_obj=None,
model_response=None,
optional_params=None,
azure_ad_token: Optional[str]=None,
client = None,
aembedding=None,
):
super().embedding()
exception_mapping_worked = False
if self._client_session is None:
self._client_session = self.create_client_session()
try:
data = {
"model": model,
"input": input,
**optional_params
}
max_retries = data.pop("max_retries", 2)
if not isinstance(max_retries, int):
raise AzureOpenAIError(status_code=422, message="max retries must be an int")
# init AzureOpenAI Client
azure_client_params = {
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"http_client": litellm.client_session,
"max_retries": max_retries,
"timeout": timeout
}
if api_key is not None:
azure_client_params["api_key"] = api_key
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if aembedding == True:
response = self.aembedding(data=data, model_response=model_response, azure_client_params=azure_client_params)
return response
if client is None:
azure_client = AzureOpenAI(**azure_client_params) # type: ignore
else:
azure_client = client
## LOGGING
logging_obj.pre_call(
input=input,
api_key=api_key,
additional_args={
"complete_input_dict": data,
"headers": {
"api_key": api_key,
"azure_ad_token": azure_ad_token
}
},
)
## COMPLETION CALL
response = azure_client.embeddings.create(**data) # type: ignore
## LOGGING
logging_obj.post_call(
input=input,
api_key=api_key,
additional_args={"complete_input_dict": data, "api_base": api_base},
original_response=response,
)
return convert_to_model_response_object(response_object=json.loads(response.model_dump_json()), model_response_object=model_response, response_type="embedding") # type: ignore
except AzureOpenAIError as e:
exception_mapping_worked = True
raise e
except Exception as e:
if exception_mapping_worked:
raise e
else:
import traceback
raise AzureOpenAIError(status_code=500, message=traceback.format_exc()) |