Spaces:
Sleeping
Sleeping
File size: 25,198 Bytes
7db0ae4 |
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 |
### What this tests ####
## This test asserts the type of data passed into each method of the custom callback handler
import sys, os, time, inspect, asyncio, traceback
from datetime import datetime
import pytest
sys.path.insert(0, os.path.abspath("../.."))
from typing import Optional, Literal, List
from litellm import Router, Cache
import litellm
from litellm.integrations.custom_logger import CustomLogger
# Test Scenarios (test across completion, streaming, embedding)
## 1: Pre-API-Call
## 2: Post-API-Call
## 3: On LiteLLM Call success
## 4: On LiteLLM Call failure
## fallbacks
## retries
# Test cases
## 1. Simple Azure OpenAI acompletion + streaming call
## 2. Simple Azure OpenAI aembedding call
## 3. Azure OpenAI acompletion + streaming call with retries
## 4. Azure OpenAI aembedding call with retries
## 5. Azure OpenAI acompletion + streaming call with fallbacks
## 6. Azure OpenAI aembedding call with fallbacks
# Test interfaces
## 1. router.completion() + router.embeddings()
## 2. proxy.completions + proxy.embeddings
class CompletionCustomHandler(
CustomLogger
): # https://docs.litellm.ai/docs/observability/custom_callback#callback-class
"""
The set of expected inputs to a custom handler for a
"""
# Class variables or attributes
def __init__(self):
self.errors = []
self.states: Optional[
List[
Literal[
"sync_pre_api_call",
"async_pre_api_call",
"post_api_call",
"sync_stream",
"async_stream",
"sync_success",
"async_success",
"sync_failure",
"async_failure",
]
]
] = []
def log_pre_api_call(self, model, messages, kwargs):
try:
print(f"received kwargs in pre-input: {kwargs}")
self.states.append("sync_pre_api_call")
## MODEL
assert isinstance(model, str)
## MESSAGES
assert isinstance(messages, list)
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
### ROUTER-SPECIFIC KWARGS
assert isinstance(kwargs["litellm_params"]["metadata"], dict)
assert isinstance(kwargs["litellm_params"]["metadata"]["model_group"], str)
assert isinstance(kwargs["litellm_params"]["metadata"]["deployment"], str)
assert isinstance(kwargs["litellm_params"]["model_info"], dict)
assert isinstance(kwargs["litellm_params"]["model_info"]["id"], str)
assert isinstance(
kwargs["litellm_params"]["proxy_server_request"], (str, type(None))
)
assert isinstance(
kwargs["litellm_params"]["preset_cache_key"], (str, type(None))
)
assert isinstance(kwargs["litellm_params"]["stream_response"], dict)
except Exception as e:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
def log_post_api_call(self, kwargs, response_obj, start_time, end_time):
try:
self.states.append("post_api_call")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert end_time == None
## RESPONSE OBJECT
assert response_obj == None
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert isinstance(kwargs["input"], (list, dict, str))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert (
isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
or inspect.iscoroutine(kwargs["original_response"])
or inspect.isasyncgen(kwargs["original_response"])
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
### ROUTER-SPECIFIC KWARGS
assert isinstance(kwargs["litellm_params"]["metadata"], dict)
assert isinstance(kwargs["litellm_params"]["metadata"]["model_group"], str)
assert isinstance(kwargs["litellm_params"]["metadata"]["deployment"], str)
assert isinstance(kwargs["litellm_params"]["model_info"], dict)
assert isinstance(kwargs["litellm_params"]["model_info"]["id"], str)
assert isinstance(
kwargs["litellm_params"]["proxy_server_request"], (str, type(None))
)
assert isinstance(
kwargs["litellm_params"]["preset_cache_key"], (str, type(None))
)
assert isinstance(kwargs["litellm_params"]["stream_response"], dict)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
async def async_log_stream_event(self, kwargs, response_obj, start_time, end_time):
try:
self.states.append("async_stream")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert isinstance(end_time, datetime)
## RESPONSE OBJECT
assert isinstance(response_obj, litellm.ModelResponse)
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list) and isinstance(
kwargs["messages"][0], dict
)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert (
isinstance(kwargs["input"], list)
and isinstance(kwargs["input"][0], dict)
) or isinstance(kwargs["input"], (dict, str))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert (
isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
or inspect.isasyncgen(kwargs["original_response"])
or inspect.iscoroutine(kwargs["original_response"])
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
def log_success_event(self, kwargs, response_obj, start_time, end_time):
try:
self.states.append("sync_success")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert isinstance(end_time, datetime)
## RESPONSE OBJECT
assert isinstance(response_obj, litellm.ModelResponse)
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list) and isinstance(
kwargs["messages"][0], dict
)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert (
isinstance(kwargs["input"], list)
and isinstance(kwargs["input"][0], dict)
) or isinstance(kwargs["input"], (dict, str))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
assert kwargs["cache_hit"] is None or isinstance(kwargs["cache_hit"], bool)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
def log_failure_event(self, kwargs, response_obj, start_time, end_time):
try:
self.states.append("sync_failure")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert isinstance(end_time, datetime)
## RESPONSE OBJECT
assert response_obj == None
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list) and isinstance(
kwargs["messages"][0], dict
)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert (
isinstance(kwargs["input"], list)
and isinstance(kwargs["input"][0], dict)
) or isinstance(kwargs["input"], (dict, str))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert (
isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
or kwargs["original_response"] == None
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
async def async_log_pre_api_call(self, model, messages, kwargs):
try:
"""
No-op.
Not implemented yet.
"""
pass
except Exception as e:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
async def async_log_success_event(self, kwargs, response_obj, start_time, end_time):
try:
self.states.append("async_success")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert isinstance(end_time, datetime)
## RESPONSE OBJECT
assert isinstance(
response_obj, (litellm.ModelResponse, litellm.EmbeddingResponse)
)
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert isinstance(kwargs["input"], (list, dict, str))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert (
isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
or inspect.isasyncgen(kwargs["original_response"])
or inspect.iscoroutine(kwargs["original_response"])
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
assert kwargs["cache_hit"] is None or isinstance(kwargs["cache_hit"], bool)
### ROUTER-SPECIFIC KWARGS
assert isinstance(kwargs["litellm_params"]["metadata"], dict)
assert isinstance(kwargs["litellm_params"]["metadata"]["model_group"], str)
assert isinstance(kwargs["litellm_params"]["metadata"]["deployment"], str)
assert isinstance(kwargs["litellm_params"]["model_info"], dict)
assert isinstance(kwargs["litellm_params"]["model_info"]["id"], str)
assert isinstance(
kwargs["litellm_params"]["proxy_server_request"], (str, type(None))
)
assert isinstance(
kwargs["litellm_params"]["preset_cache_key"], (str, type(None))
)
assert isinstance(kwargs["litellm_params"]["stream_response"], dict)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
async def async_log_failure_event(self, kwargs, response_obj, start_time, end_time):
try:
print(f"received original response: {kwargs['original_response']}")
self.states.append("async_failure")
## START TIME
assert isinstance(start_time, datetime)
## END TIME
assert isinstance(end_time, datetime)
## RESPONSE OBJECT
assert response_obj == None
## KWARGS
assert isinstance(kwargs["model"], str)
assert isinstance(kwargs["messages"], list)
assert isinstance(kwargs["optional_params"], dict)
assert isinstance(kwargs["litellm_params"], dict)
assert isinstance(kwargs["start_time"], (datetime, type(None)))
assert isinstance(kwargs["stream"], bool)
assert isinstance(kwargs["user"], (str, type(None)))
assert isinstance(kwargs["input"], (list, str, dict))
assert isinstance(kwargs["api_key"], (str, type(None)))
assert (
isinstance(
kwargs["original_response"], (str, litellm.CustomStreamWrapper)
)
or inspect.isasyncgen(kwargs["original_response"])
or inspect.iscoroutine(kwargs["original_response"])
or kwargs["original_response"] == None
)
assert isinstance(kwargs["additional_args"], (dict, type(None)))
assert isinstance(kwargs["log_event_type"], str)
except:
print(f"Assertion Error: {traceback.format_exc()}")
self.errors.append(traceback.format_exc())
# Simple Azure OpenAI call
## COMPLETION
@pytest.mark.asyncio
async def test_async_chat_azure():
try:
customHandler_completion_azure_router = CompletionCustomHandler()
customHandler_streaming_azure_router = CompletionCustomHandler()
customHandler_failure = CompletionCustomHandler()
litellm.callbacks = [customHandler_completion_azure_router]
model_list = [
{
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": os.getenv("AZURE_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
]
router = Router(model_list=model_list) # type: ignore
response = await router.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}],
)
await asyncio.sleep(2)
assert len(customHandler_completion_azure_router.errors) == 0
assert (
len(customHandler_completion_azure_router.states) == 3
) # pre, post, success
# streaming
litellm.callbacks = [customHandler_streaming_azure_router]
router2 = Router(model_list=model_list) # type: ignore
response = await router2.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}],
stream=True,
)
async for chunk in response:
print(f"async azure router chunk: {chunk}")
continue
await asyncio.sleep(1)
print(f"customHandler.states: {customHandler_streaming_azure_router.states}")
assert len(customHandler_streaming_azure_router.errors) == 0
assert (
len(customHandler_streaming_azure_router.states) >= 4
) # pre, post, stream (multiple times), success
# failure
model_list = [
{
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": "my-bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
]
litellm.callbacks = [customHandler_failure]
router3 = Router(model_list=model_list) # type: ignore
try:
response = await router3.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}],
)
print(f"response in router3 acompletion: {response}")
except:
pass
await asyncio.sleep(1)
print(f"customHandler.states: {customHandler_failure.states}")
assert len(customHandler_failure.errors) == 0
assert len(customHandler_failure.states) == 3 # pre, post, failure
assert "async_failure" in customHandler_failure.states
except Exception as e:
print(f"Assertion Error: {traceback.format_exc()}")
pytest.fail(f"An exception occurred - {str(e)}")
# asyncio.run(test_async_chat_azure())
## EMBEDDING
@pytest.mark.asyncio
async def test_async_embedding_azure():
try:
customHandler = CompletionCustomHandler()
customHandler_failure = CompletionCustomHandler()
litellm.callbacks = [customHandler]
model_list = [
{
"model_name": "azure-embedding-model", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/azure-embedding-model",
"api_key": os.getenv("AZURE_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
]
router = Router(model_list=model_list) # type: ignore
response = await router.aembedding(
model="azure-embedding-model", input=["hello from litellm!"]
)
await asyncio.sleep(2)
assert len(customHandler.errors) == 0
assert len(customHandler.states) == 3 # pre, post, success
# failure
model_list = [
{
"model_name": "azure-embedding-model", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/azure-embedding-model",
"api_key": "my-bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
]
litellm.callbacks = [customHandler_failure]
router3 = Router(model_list=model_list) # type: ignore
try:
response = await router3.aembedding(
model="azure-embedding-model", input=["hello from litellm!"]
)
print(f"response in router3 aembedding: {response}")
except:
pass
await asyncio.sleep(1)
print(f"customHandler.states: {customHandler_failure.states}")
assert len(customHandler_failure.errors) == 0
assert len(customHandler_failure.states) == 3 # pre, post, failure
assert "async_failure" in customHandler_failure.states
except Exception as e:
print(f"Assertion Error: {traceback.format_exc()}")
pytest.fail(f"An exception occurred - {str(e)}")
# asyncio.run(test_async_embedding_azure())
# Azure OpenAI call w/ Fallbacks
## COMPLETION
@pytest.mark.asyncio
async def test_async_chat_azure_with_fallbacks():
try:
customHandler_fallbacks = CompletionCustomHandler()
litellm.callbacks = [customHandler_fallbacks]
# with fallbacks
model_list = [
{
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": "my-bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
{
"model_name": "gpt-3.5-turbo-16k",
"litellm_params": {
"model": "gpt-3.5-turbo-16k",
},
"tpm": 240000,
"rpm": 1800,
},
]
router = Router(model_list=model_list, fallbacks=[{"gpt-3.5-turbo": ["gpt-3.5-turbo-16k"]}]) # type: ignore
response = await router.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}],
)
await asyncio.sleep(2)
print(f"customHandler_fallbacks.states: {customHandler_fallbacks.states}")
assert len(customHandler_fallbacks.errors) == 0
assert (
len(customHandler_fallbacks.states) == 6
) # pre, post, failure, pre, post, success
litellm.callbacks = []
except Exception as e:
print(f"Assertion Error: {traceback.format_exc()}")
pytest.fail(f"An exception occurred - {str(e)}")
# asyncio.run(test_async_chat_azure_with_fallbacks())
# CACHING
## Test Azure - completion, embedding
@pytest.mark.asyncio
async def test_async_completion_azure_caching():
customHandler_caching = CompletionCustomHandler()
litellm.cache = Cache(
type="redis",
host=os.environ["REDIS_HOST"],
port=os.environ["REDIS_PORT"],
password=os.environ["REDIS_PASSWORD"],
)
litellm.callbacks = [customHandler_caching]
unique_time = time.time()
model_list = [
{
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/chatgpt-v-2",
"api_key": os.getenv("AZURE_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
},
{
"model_name": "gpt-3.5-turbo-16k",
"litellm_params": {
"model": "gpt-3.5-turbo-16k",
},
"tpm": 240000,
"rpm": 1800,
},
]
router = Router(model_list=model_list) # type: ignore
response1 = await router.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": f"Hi 👋 - i'm async azure {unique_time}"}],
caching=True,
)
await asyncio.sleep(1)
print(f"customHandler_caching.states pre-cache hit: {customHandler_caching.states}")
response2 = await router.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": f"Hi 👋 - i'm async azure {unique_time}"}],
caching=True,
)
await asyncio.sleep(1) # success callbacks are done in parallel
print(
f"customHandler_caching.states post-cache hit: {customHandler_caching.states}"
)
assert len(customHandler_caching.errors) == 0
assert len(customHandler_caching.states) == 4 # pre, post, success, success
|