tiennm99/litellm
1
0
Fork 0
mirror of https://github.com/tiennm99/litellm.git synced 2026-07-13 21:06:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor: cleanup deprecated models

Browse Source
This commit is contained in:
Krrish Dholakia
2026-03-28 19:39:11 -07:00
parent bc829d51f2
commit cdcab8a243
4 changed files with 76 additions and 172 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/batches_tests/test_openai_batches_and_files.py
+25 -21
View File
@@ -234,9 +234,9 @@ def cleanup_azure_ft_models():
import requests
client = AzureOpenAI(
api_key=os.getenv("AZURE_FT_API_KEY"),
azure_endpoint=os.getenv("AZURE_FT_API_BASE"),
api_version=os.getenv("AZURE_API_VERSION"),
api_key=os.getenv("AZURE_AI_API_KEY"),
azure_endpoint=os.getenv("AZURE_AI_API_BASE"),
api_version=os.getenv("AZURE_AI_API_VERSION"),
)
_list_ft_jobs = client.fine_tuning.jobs.list()
@@ -577,7 +577,10 @@ async def test_vertex_list_batches(monkeypatch):
monkeypatch.setattr(
"litellm.llms.vertex_ai.batches.handler.VertexAIBatchPrediction._ensure_access_token",
lambda self, credentials, project_id, custom_llm_provider: ("mock-token", "litellm-test-project"),
lambda self, credentials, project_id, custom_llm_provider: (
"mock-token",
"litellm-test-project",
),
)
with patch(
@@ -648,7 +651,7 @@ async def test_vertex_async_create_batch_logs_error_body_on_http_error():
async def test_delete_batch_output_file():
"""
Test that deleting a batch output file works correctly.
This test verifies the fix for:
- When a batch is retrieved and has an output_file_id, the file object is properly stored
- The output file can be deleted without validation errors
@@ -656,11 +659,11 @@ async def test_delete_batch_output_file():
"""
litellm._turn_on_debug()
print("Testing delete batch output file")
file_name = "openai_batch_completions.jsonl"
_current_dir = os.path.dirname(os.path.abspath(__file__))
file_path = os.path.join(_current_dir, file_name)
# Create file for batch
file_obj = await litellm.acreate_file(
file=open(file_path, "rb"),
@@ -669,7 +672,7 @@ async def test_delete_batch_output_file():
)
print("Response from creating file=", file_obj)
batch_input_file_id = file_obj.id
# Create batch
create_batch_response = await litellm.acreate_batch(
completion_window="24h",
@@ -678,36 +681,37 @@ async def test_delete_batch_output_file():
custom_llm_provider="openai",
)
print("Batch created with ID=", create_batch_response.id)
# Retrieve batch to get output_file_id
retrieved_batch = await litellm.aretrieve_batch(
batch_id=create_batch_response.id,
custom_llm_provider="openai"
batch_id=create_batch_response.id, custom_llm_provider="openai"
)
print("Retrieved batch=", retrieved_batch)
# If batch has completed and has output file, test deleting it
if retrieved_batch.output_file_id:
print(f"Testing deletion of output file: {retrieved_batch.output_file_id}")
# This is the key test - deleting the output file should work
# without validation errors (file_object should not be None)
delete_output_file_response = await litellm.afile_delete(
file_id=retrieved_batch.output_file_id,
custom_llm_provider="openai"
file_id=retrieved_batch.output_file_id, custom_llm_provider="openai"
)
print("Delete output file response=", delete_output_file_response)
assert delete_output_file_response.id == retrieved_batch.output_file_id
assert delete_output_file_response.deleted is True or hasattr(delete_output_file_response, 'id')
assert delete_output_file_response.deleted is True or hasattr(
delete_output_file_response, "id"
)
print("✓ Successfully deleted batch output file")
else:
print("⚠ Batch has not completed yet or no output file available, skipping output file deletion test")
print(
"⚠ Batch has not completed yet or no output file available, skipping output file deletion test"
)
# Clean up - delete the input file
delete_input_file_response = await litellm.afile_delete(
file_id=batch_input_file_id,
custom_llm_provider="openai"
file_id=batch_input_file_id, custom_llm_provider="openai"
)
print("Delete input file response=", delete_input_file_response)
assert delete_input_file_response.id == batch_input_file_id
tests/litellm_utils_tests/test_health_check.py
-29
View File
@@ -246,35 +246,6 @@ async def test_audio_transcription_health_check():
print(response)
@pytest.mark.asyncio
@pytest.mark.parametrize(
"model", ["azure/gpt-4o-realtime-preview", "openai/gpt-4o-realtime-preview"]
)
async def test_async_realtime_health_check(model, mocker):
"""
Test Health Check with Valid models passes
"""
mock_websocket = AsyncMock()
mock_connect = AsyncMock().__aenter__.return_value = mock_websocket
mocker.patch("websockets.connect", return_value=mock_connect)
litellm.set_verbose = True
model_params = {
"model": model,
}
if model == "azure/gpt-4o-realtime-preview":
model_params["api_base"] = os.getenv("AZURE_REALTIME_API_BASE")
model_params["api_key"] = os.getenv("AZURE_REALTIME_API_KEY")
model_params["api_version"] = os.getenv("AZURE_REALTIME_API_VERSION")
response = await litellm.ahealth_check(
model_params=model_params,
mode="realtime",
)
print(response)
assert response == {}
def test_update_litellm_params_for_health_check():
"""
Test if _update_litellm_params_for_health_check correctly:
tests/llm_translation/test_azure_ai.py
+1 -30
View File
@@ -188,35 +188,6 @@ def test_azure_ai_services_with_api_version():
)
@pytest.mark.skip(reason="Skipping due to cohere ssl issues")
def test_completion_azure_ai_command_r():
try:
import os
litellm.set_verbose = True
os.environ["AZURE_AI_API_BASE"] = os.getenv("AZURE_COHERE_API_BASE", "")
os.environ["AZURE_AI_API_KEY"] = os.getenv("AZURE_COHERE_API_KEY", "")
response = completion(
model="azure_ai/command-r-plus",
messages=[
{
"role": "user",
"content": [
{"type": "text", "text": "What is the meaning of life?"}
],
}
],
) # type: ignore
assert "azure_ai" in response.model
except litellm.Timeout as e:
pass
except Exception as e:
pytest.fail(f"Error occurred: {e}")
def test_azure_deepseek_reasoning_content():
import json
@@ -387,7 +358,7 @@ async def test_azure_ai_model_router():
response = await litellm.acompletion(
model="azure_ai/model_router/azure-model-router",
messages=[{"role": "user", "content": "hi who is this"}],
api_base="https://ishaa-mh6uutut-swedencentral.cognitiveservices.azure.com/openai/v1/",
api_base=os.getenv("AZURE_MODEL_ROUTER_API_BASE"),
api_key=os.getenv("AZURE_MODEL_ROUTER_API_KEY"),
)
print("response: ", response)
tests/local_testing/test_router.py
+50 -92
View File
@@ -126,7 +126,9 @@ async def test_router_provider_wildcard_routing():
print("response 3 = ", response3)
response4 = await router.acompletion(
model=os.environ.get("CI_CD_DEFAULT_ANTHROPIC_MODEL", "claude-haiku-4-5-20251001"),
model=os.environ.get(
"CI_CD_DEFAULT_ANTHROPIC_MODEL", "claude-haiku-4-5-20251001"
),
messages=[{"role": "user", "content": "hello"}],
)
@@ -356,51 +358,6 @@ async def test_router_retries(sync_mode):
print(response.choices[0].message)
@pytest.mark.parametrize(
"mistral_api_base",
[
"os.environ/AZURE_MISTRAL_API_BASE",
"https://Mistral-large-nmefg-serverless.eastus2.inference.ai.azure.com/v1/",
"https://Mistral-large-nmefg-serverless.eastus2.inference.ai.azure.com/v1",
"https://Mistral-large-nmefg-serverless.eastus2.inference.ai.azure.com/",
"https://Mistral-large-nmefg-serverless.eastus2.inference.ai.azure.com",
],
)
@pytest.mark.skip(
reason="Router no longer creates clients, this is delegated to the provider integration."
)
def test_router_azure_ai_studio_init(mistral_api_base):
router = Router(
model_list=[
{
"model_name": "test-model",
"litellm_params": {
"model": "azure/mistral-large-latest",
"api_key": "os.environ/AZURE_MISTRAL_API_KEY",
"api_base": mistral_api_base,
},
"model_info": {"id": 1234},
}
]
)
# model_client = router._get_client(
# deployment={"model_info": {"id": 1234}}, client_type="sync_client", kwargs={}
# )
# url = getattr(model_client, "_base_url")
# uri_reference = str(getattr(url, "_uri_reference"))
# print(f"uri_reference: {uri_reference}")
# assert "/v1/" in uri_reference
# assert uri_reference.count("v1") == 1
response = router.completion(
model="azure/mistral-large-latest",
messages=[{"role": "user", "content": "Hey, how's it going?"}],
)
assert response is not None
def test_exception_raising():
# this tests if the router raises an exception when invalid params are set
# in this test both deployments have bad keys - Keep this test. It validates if the router raises the most recent exception
@@ -409,8 +366,8 @@ def test_exception_raising():
try:
print("testing if router raises an exception")
old_api_key = os.environ["AZURE_API_KEY"]
os.environ["AZURE_API_KEY"] = ""
old_api_key = os.environ["AZURE_AI_API_KEY"]
os.environ["AZURE_AI_API_KEY"] = ""
model_list = [
{
"model_name": "gpt-3.5-turbo", # openai model name
@@ -418,7 +375,7 @@ def test_exception_raising():
"model": "azure/gpt-4.1-mini",
"api_key": "bad-key",
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
@@ -446,16 +403,16 @@ def test_exception_raising():
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hello this request will fail"}],
)
os.environ["AZURE_API_KEY"] = old_api_key
os.environ["AZURE_AI_API_KEY"] = old_api_key
pytest.fail(f"Should have raised an Auth Error")
except openai.AuthenticationError:
print(
"Test Passed: Caught an OPENAI AUTH Error, Good job. This is what we needed!"
)
os.environ["AZURE_API_KEY"] = old_api_key
os.environ["AZURE_AI_API_KEY"] = old_api_key
router.reset()
except Exception as e:
os.environ["AZURE_API_KEY"] = old_api_key
os.environ["AZURE_AI_API_KEY"] = old_api_key
print("Got unexpected exception on router!", e)
@@ -530,7 +487,7 @@ def test_call_one_endpoint():
# this test makes a completion calls azure/gpt-4.1-mini, it should work
try:
print("Testing calling a specific deployment")
old_api_key = os.environ["AZURE_API_KEY"]
old_api_key = os.environ["AZURE_AI_API_KEY"]
model_list = [
{
@@ -539,7 +496,7 @@ def test_call_one_endpoint():
"model": "azure/gpt-4.1-mini",
"api_key": old_api_key,
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
},
"tpm": 240000,
"rpm": 1800,
@@ -548,8 +505,8 @@ def test_call_one_endpoint():
"model_name": "text-embedding-ada-002",
"litellm_params": {
"model": "azure/text-embedding-ada-002",
"api_key": os.environ["AZURE_API_KEY"],
"api_base": os.environ["AZURE_API_BASE"],
"api_key": os.environ["AZURE_AI_API_KEY"],
"api_base": os.environ["AZURE_AI_API_BASE"],
},
"tpm": 100000,
"rpm": 10000,
@@ -562,7 +519,7 @@ def test_call_one_endpoint():
set_verbose=True,
num_retries=1,
) # type: ignore
old_api_base = os.environ.pop("AZURE_API_BASE", None)
old_api_base = os.environ.pop("AZURE_AI_API_BASE", None)
async def call_azure_completion():
response = await router.acompletion(
@@ -584,8 +541,8 @@ def test_call_one_endpoint():
asyncio.run(call_azure_completion())
asyncio.run(call_azure_embedding())
os.environ["AZURE_API_BASE"] = old_api_base
os.environ["AZURE_API_KEY"] = old_api_key
os.environ["AZURE_AI_API_BASE"] = old_api_base
os.environ["AZURE_AI_API_KEY"] = old_api_key
except Exception as e:
print(f"FAILED TEST")
pytest.fail(f"Got unexpected exception on router! - {e}")
@@ -594,7 +551,6 @@ def test_call_one_endpoint():
# test_call_one_endpoint()
@pytest.mark.asyncio
@pytest.mark.parametrize("sync_mode", [True, False])
async def test_async_router_context_window_fallback(sync_mode):
@@ -708,9 +664,9 @@ def test_router_context_window_check_pre_call_check_in_group_custom_model_info()
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/gpt-4.1-mini",
"api_key": os.getenv("AZURE_API_KEY"),
"api_key": os.getenv("AZURE_AI_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
"base_model": "azure/gpt-35-turbo",
"mock_response": "Hello world 1!",
},
@@ -762,9 +718,9 @@ def test_router_context_window_check_pre_call_check():
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/gpt-4.1-mini",
"api_key": os.getenv("AZURE_API_KEY"),
"api_key": os.getenv("AZURE_AI_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
"base_model": "azure/gpt-35-turbo",
"mock_response": "Hello world 1!",
},
@@ -816,9 +772,9 @@ def test_router_context_window_check_pre_call_check_out_group():
"model_name": "gpt-3.5-turbo-small", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/gpt-4.1-mini",
"api_key": os.getenv("AZURE_API_KEY"),
"api_key": os.getenv("AZURE_AI_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
"base_model": "azure/gpt-35-turbo",
},
},
@@ -896,9 +852,9 @@ def test_router_region_pre_call_check(allowed_model_region):
"model_name": "gpt-3.5-turbo", # openai model name
"litellm_params": { # params for litellm completion/embedding call
"model": "azure/gpt-4.1-mini",
"api_key": os.getenv("AZURE_API_KEY"),
"api_key": os.getenv("AZURE_AI_API_KEY"),
"api_version": os.getenv("AZURE_API_VERSION"),
"api_base": os.getenv("AZURE_API_BASE"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
"base_model": "azure/gpt-35-turbo",
"region_name": allowed_model_region,
},
@@ -1173,8 +1129,8 @@ def test_azure_embedding_on_router():
"model_name": "text-embedding-ada-002",
"litellm_params": {
"model": "azure/text-embedding-ada-002",
"api_key": os.environ["AZURE_API_KEY"],
"api_base": os.environ["AZURE_API_BASE"],
"api_key": os.environ["AZURE_AI_API_KEY"],
"api_base": os.environ["AZURE_AI_API_BASE"],
},
"tpm": 100000,
"rpm": 10000,
@@ -1381,8 +1337,8 @@ def test_reading_keys_os_environ():
"model_name": "gpt-3.5-turbo",
"litellm_params": {
"model": "gpt-3.5-turbo",
"api_key": "os.environ/AZURE_API_KEY",
"api_base": "os.environ/AZURE_API_BASE",
"api_key": "os.environ/AZURE_AI_API_KEY",
"api_base": "os.environ/AZURE_AI_API_BASE",
"api_version": "os.environ/AZURE_API_VERSION",
"timeout": "os.environ/AZURE_TIMEOUT",
"stream_timeout": "os.environ/AZURE_STREAM_TIMEOUT",
@@ -1394,11 +1350,11 @@ def test_reading_keys_os_environ():
router = Router(model_list=model_list)
for model in router.model_list:
assert (
model["litellm_params"]["api_key"] == os.environ["AZURE_API_KEY"]
), f"{model['litellm_params']['api_key']} vs {os.environ['AZURE_API_KEY']}"
model["litellm_params"]["api_key"] == os.environ["AZURE_AI_API_KEY"]
), f"{model['litellm_params']['api_key']} vs {os.environ['AZURE_AI_API_KEY']}"
assert (
model["litellm_params"]["api_base"] == os.environ["AZURE_API_BASE"]
), f"{model['litellm_params']['api_base']} vs {os.environ['AZURE_API_BASE']}"
model["litellm_params"]["api_base"] == os.environ["AZURE_AI_API_BASE"]
), f"{model['litellm_params']['api_base']} vs {os.environ['AZURE_AI_API_BASE']}"
assert (
model["litellm_params"]["api_version"]
== os.environ["AZURE_API_VERSION"]
@@ -1415,8 +1371,8 @@ def test_reading_keys_os_environ():
print("passed testing of reading keys from os.environ")
model_id = model["model_info"]["id"]
async_client: openai.AsyncAzureOpenAI = router.cache.get_cache(f"{model_id}_async_client") # type: ignore
assert async_client.api_key == os.environ["AZURE_API_KEY"]
assert async_client.base_url == os.environ["AZURE_API_BASE"]
assert async_client.api_key == os.environ["AZURE_AI_API_KEY"]
assert async_client.base_url == os.environ["AZURE_AI_API_BASE"]
assert async_client.max_retries == int(
os.environ["AZURE_MAX_RETRIES"]
), f"{async_client.max_retries} vs {os.environ['AZURE_MAX_RETRIES']}"
@@ -1428,8 +1384,8 @@ def test_reading_keys_os_environ():
print("\n Testing async streaming client")
stream_async_client: openai.AsyncAzureOpenAI = router.cache.get_cache(f"{model_id}_stream_async_client") # type: ignore
assert stream_async_client.api_key == os.environ["AZURE_API_KEY"]
assert stream_async_client.base_url == os.environ["AZURE_API_BASE"]
assert stream_async_client.api_key == os.environ["AZURE_AI_API_KEY"]
assert stream_async_client.base_url == os.environ["AZURE_AI_API_BASE"]
assert stream_async_client.max_retries == int(
os.environ["AZURE_MAX_RETRIES"]
), f"{stream_async_client.max_retries} vs {os.environ['AZURE_MAX_RETRIES']}"
@@ -1440,8 +1396,8 @@ def test_reading_keys_os_environ():
print("\n Testing sync client")
client: openai.AzureOpenAI = router.cache.get_cache(f"{model_id}_client") # type: ignore
assert client.api_key == os.environ["AZURE_API_KEY"]
assert client.base_url == os.environ["AZURE_API_BASE"]
assert client.api_key == os.environ["AZURE_AI_API_KEY"]
assert client.base_url == os.environ["AZURE_AI_API_BASE"]
assert client.max_retries == int(
os.environ["AZURE_MAX_RETRIES"]
), f"{client.max_retries} vs {os.environ['AZURE_MAX_RETRIES']}"
@@ -1452,8 +1408,8 @@ def test_reading_keys_os_environ():
print("\n Testing sync stream client")
stream_client: openai.AzureOpenAI = router.cache.get_cache(f"{model_id}_stream_client") # type: ignore
assert stream_client.api_key == os.environ["AZURE_API_KEY"]
assert stream_client.base_url == os.environ["AZURE_API_BASE"]
assert stream_client.api_key == os.environ["AZURE_AI_API_KEY"]
assert stream_client.base_url == os.environ["AZURE_AI_API_BASE"]
assert stream_client.max_retries == int(
os.environ["AZURE_MAX_RETRIES"]
), f"{stream_client.max_retries} vs {os.environ['AZURE_MAX_RETRIES']}"
@@ -1503,7 +1459,7 @@ def test_reading_openai_keys_os_environ():
for model in router.model_list:
assert (
model["litellm_params"]["api_key"] == os.environ["OPENAI_API_KEY"]
), f"{model['litellm_params']['api_key']} vs {os.environ['AZURE_API_KEY']}"
), f"{model['litellm_params']['api_key']} vs {os.environ['AZURE_AI_API_KEY']}"
assert float(model["litellm_params"]["timeout"]) == float(
os.environ["AZURE_TIMEOUT"]
), f"{model['litellm_params']['timeout']} vs {os.environ['AZURE_TIMEOUT']}"
@@ -1574,7 +1530,9 @@ def test_router_anthropic_key_dynamic():
{
"model_name": "anthropic-claude",
"litellm_params": {
"model": os.environ.get("CI_CD_DEFAULT_ANTHROPIC_MODEL", "claude-haiku-4-5-20251001"),
"model": os.environ.get(
"CI_CD_DEFAULT_ANTHROPIC_MODEL", "claude-haiku-4-5-20251001"
),
"api_key": anthropic_api_key,
},
}
@@ -2273,8 +2231,8 @@ async def test_router_batch_endpoints(provider):
"model_name": "my-custom-name",
"litellm_params": {
"model": "azure/gpt-4o-mini",
"api_base": os.getenv("AZURE_API_BASE"),
"api_key": os.getenv("AZURE_API_KEY"),
"api_base": os.getenv("AZURE_AI_API_BASE"),
"api_key": os.getenv("AZURE_AI_API_KEY"),
},
},
]
@@ -2452,8 +2410,8 @@ def test_is_team_specific_model():
# "model_name": "gpt-3.5-turbo",
# "litellm_params": {
# "model": "azure/gpt-4.1-mini",
# "api_key": os.getenv("AZURE_API_KEY"),
# "api_base": os.getenv("AZURE_API_BASE"),
# "api_key": os.getenv("AZURE_AI_API_KEY"),
# "api_base": os.getenv("AZURE_AI_API_BASE"),
# "tpm": 100000,
# "rpm": 100000,
# },
@@ -2462,8 +2420,8 @@ def test_is_team_specific_model():
# "model_name": "gpt-3.5-turbo",
# "litellm_params": {
# "model": "azure/gpt-4.1-mini",
# "api_key": os.getenv("AZURE_API_KEY"),
# "api_base": os.getenv("AZURE_API_BASE"),
# "api_key": os.getenv("AZURE_AI_API_KEY"),
# "api_base": os.getenv("AZURE_AI_API_BASE"),
# "tpm": 500,
# "rpm": 500,
# },