phoenix/litellm
1
0
Fork 0
forked from phoenix/litellm-mirror
Code Pull requests Activity

LiteLLM Minor Fixes & Improvements (10/09/2024) (#6139)

Browse source 
* fix(utils.py): don't return 'none' response headers

Fixes https://github.com/BerriAI/litellm/issues/6123

* fix(vertex_and_google_ai_studio_gemini.py): support parsing out additional properties and strict value for tool calls

Fixes https://github.com/BerriAI/litellm/issues/6136

* fix(cost_calculator.py): set default character value to none

Fixes https://github.com/BerriAI/litellm/issues/6133#issuecomment-2403290196

* fix(google.py): fix cost per token / cost per char conversion

Fixes https://github.com/BerriAI/litellm/issues/6133#issuecomment-2403370287

* build(model_prices_and_context_window.json): update gemini pricing

Fixes https://github.com/BerriAI/litellm/issues/6133

* build(model_prices_and_context_window.json): update gemini pricing

* fix(litellm_logging.py): fix streaming caching logging when 'turn_off_message_logging' enabled

Stores unredacted response in cache

* build(model_prices_and_context_window.json): update gemini-1.5-flash pricing

* fix(cost_calculator.py): fix default prompt_character count logic

Fixes error in gemini cost calculation

* fix(cost_calculator.py): fix cost calc for tts models
This commit is contained in:
Krish Dholakia 2024-10-10 00:42:11 -07:00 • committed by GitHub
parent 60baa65e0e
commit 6005450c8f
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
16 changed files with 788 additions and 534 deletions
Download patch file Download diff file Expand all files Collapse all files

25
tests/local_testing/test_caching.py
View file

@ -2209,3 +2209,28 @@ async def test_redis_proxy_batch_redis_get_cache():
print(response._hidden_params)
assert "cache_key" in response._hidden_params
def test_logging_turn_off_message_logging_streaming():
litellm.turn_off_message_logging = True
mock_obj = Cache(type="local")
litellm.cache = mock_obj
with patch.object(mock_obj, "add_cache", new=MagicMock()) as mock_client:
print(f"mock_obj.add_cache: {mock_obj.add_cache}")
resp = litellm.completion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hi"}],
mock_response="hello",
stream=True,
)
for chunk in resp:
continue
time.sleep(1)
mock_client.assert_called_once()
assert mock_client.call_args.args[0].choices[0].message.content == "hello"