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

LiteLLM Minor Fixes & Improvements (10/04/2024) (#6064)

Browse source 
* fix(litellm_logging.py): ensure cache hits are scrubbed if 'turn_off_message_logging' is enabled

* fix(sagemaker.py): fix streaming to raise error immediately

Fixes https://github.com/BerriAI/litellm/issues/6054

* (fixes)  gcs bucket key based logging  (#6044)

* fixes for gcs bucket logging

* fix StandardCallbackDynamicParams

* fix - gcs logging when payload is not serializable

* add test_add_callback_via_key_litellm_pre_call_utils_gcs_bucket

* working success callbacks

* linting fixes

* fix linting error

* add type hints to functions

* fixes for dynamic success and failure logging

* fix for test_async_chat_openai_stream

* fix handle case when key based logging vars are set as os.environ/ vars

* fix prometheus track cooldown events on custom logger (#6060)

* (docs) add 1k rps load test doc  (#6059)

* docs 1k rps load test

* docs load testing

* docs load testing litellm

* docs load testing

* clean up load test doc

* docs prom metrics for load testing

* docs using prometheus on load testing

* doc load testing with prometheus

* (fixes) docs + qa - gcs key based logging  (#6061)

* fixes for required values for gcs bucket

* docs gcs bucket logging

* bump: version 1.48.12 → 1.48.13

* ci/cd run again

* bump: version 1.48.13 → 1.48.14

* update load test doc

* (docs) router settings - on litellm config  (#6037)

* add yaml with all router settings

* add docs for router settings

* docs router settings litellm settings

* (feat)  OpenAI prompt caching models to model cost map (#6063)

* add prompt caching for latest models

* add cache_read_input_token_cost for prompt caching models

* fix(litellm_logging.py): check if param is iterable

Fixes https://github.com/BerriAI/litellm/issues/6025#issuecomment-2393929946

* fix(factory.py): support passing an 'assistant_continue_message' to prevent bedrock error

Fixes https://github.com/BerriAI/litellm/issues/6053

* fix(databricks/chat): handle streaming responses

* fix(factory.py): fix linting error

* fix(utils.py): unify anthropic + deepseek prompt caching information to openai format

Fixes https://github.com/BerriAI/litellm/issues/6069

* test: fix test

* fix(types/utils.py): support all openai roles

Fixes https://github.com/BerriAI/litellm/issues/6052

* test: fix test

---------

Co-authored-by: Ishaan Jaff <ishaanjaffer0324@gmail.com>
This commit is contained in:
Krish Dholakia 2024-10-04 21:28:53 -04:00 • committed by GitHub
parent fc6e0dd6cb
commit 2e5c46ef6d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
19 changed files with 1034 additions and 259 deletions
Download patch file Download diff file Expand all files Collapse all files

199
docs/my-website/docs/completion/prompt_caching.md Normal file
View file

@ -0,0 +1,199 @@
# Prompt Caching
For OpenAI + Anthropic + Deepseek, LiteLLM follows the OpenAI prompt caching usage object format:
```bash
"usage": {
"prompt_tokens": 2006,
"completion_tokens": 300,
"total_tokens": 2306,
"prompt_tokens_details": {
"cached_tokens": 1920
},
"completion_tokens_details": {
"reasoning_tokens": 0
}
# ANTHROPIC_ONLY #
"cache_creation_input_tokens": 0
}
```
- `prompt_tokens`: These are the non-cached prompt tokens (same as Anthropic, equivalent to Deepseek `prompt_cache_miss_tokens`).
- `completion_tokens`: These are the output tokens generated by the model.
- `total_tokens`: Sum of prompt_tokens + completion_tokens.
- `prompt_tokens_details`: Object containing cached_tokens.
- `cached_tokens`: Tokens that were a cache-hit for that call.
- `completion_tokens_details`: Object containing reasoning_tokens.
- **ANTHROPIC_ONLY**: `cache_creation_input_tokens` are the number of tokens that were written to cache. (Anthropic charges for this).
## Quick Start
Note: OpenAI caching is only available for prompts containing 1024 tokens or more
```python
from litellm import completion
import os
os.environ["OPENAI_API_KEY"] = ""
for _ in range(2):
response = completion(
model="gpt-4o",
messages=[
# System Message
{
"role": "system",
"content": [
{
"type": "text",
"text": "Here is the full text of a complex legal agreement"
* 400,
}
],
},
# marked for caching with the cache_control parameter, so that this checkpoint can read from the previous cache.
{
"role": "user",
"content": [
{
"type": "text",
"text": "What are the key terms and conditions in this agreement?",
}
],
},
{
"role": "assistant",
"content": "Certainly! the key terms and conditions are the following: the contract is 1 year long for $10/mo",
},
# The final turn is marked with cache-control, for continuing in followups.
{
"role": "user",
"content": [
{
"type": "text",
"text": "What are the key terms and conditions in this agreement?",
}
],
},
],
temperature=0.2,
max_tokens=10,
)
print("response=", response)
print("response.usage=", response.usage)
assert "prompt_tokens_details" in response.usage
assert response.usage.prompt_tokens_details.cached_tokens > 0
```
### Anthropic Example
Anthropic charges for cache writes.
Specify the content to cache with `"cache_control": {"type": "ephemeral"}`.
If you pass that in for any other llm provider, it will be ignored.
```python
from litellm import completion
import litellm
import os
litellm.set_verbose = True # 👈 SEE RAW REQUEST
os.environ["ANTHROPIC_API_KEY"] = ""
response = completion(
model="anthropic/claude-3-5-sonnet-20240620",
messages=[
{
"role": "system",
"content": [
{
"type": "text",
"text": "You are an AI assistant tasked with analyzing legal documents.",
},
{
"type": "text",
"text": "Here is the full text of a complex legal agreement" * 400,
"cache_control": {"type": "ephemeral"},
},
],
},
{
"role": "user",
"content": "what are the key terms and conditions in this agreement?",
},
]
)
print(response.usage)
```
### Deepeek Example
Works the same as OpenAI.
```python
from litellm import completion
import litellm
import os
os.environ["DEEPSEEK_API_KEY"] = ""
litellm.set_verbose = True # 👈 SEE RAW REQUEST
model_name = "deepseek/deepseek-chat"
messages_1 = [
{
"role": "system",
"content": "You are a history expert. The user will provide a series of questions, and your answers should be concise and start with `Answer:`",
},
{
"role": "user",
"content": "In what year did Qin Shi Huang unify the six states?",
},
{"role": "assistant", "content": "Answer: 221 BC"},
{"role": "user", "content": "Who was the founder of the Han Dynasty?"},
{"role": "assistant", "content": "Answer: Liu Bang"},
{"role": "user", "content": "Who was the last emperor of the Tang Dynasty?"},
{"role": "assistant", "content": "Answer: Li Zhu"},
{
"role": "user",
"content": "Who was the founding emperor of the Ming Dynasty?",
},
{"role": "assistant", "content": "Answer: Zhu Yuanzhang"},
{
"role": "user",
"content": "Who was the founding emperor of the Qing Dynasty?",
},
]
message_2 = [
{
"role": "system",
"content": "You are a history expert. The user will provide a series of questions, and your answers should be concise and start with `Answer:`",
},
{
"role": "user",
"content": "In what year did Qin Shi Huang unify the six states?",
},
{"role": "assistant", "content": "Answer: 221 BC"},
{"role": "user", "content": "Who was the founder of the Han Dynasty?"},
{"role": "assistant", "content": "Answer: Liu Bang"},
{"role": "user", "content": "Who was the last emperor of the Tang Dynasty?"},
{"role": "assistant", "content": "Answer: Li Zhu"},
{
"role": "user",
"content": "Who was the founding emperor of the Ming Dynasty?",
},
{"role": "assistant", "content": "Answer: Zhu Yuanzhang"},
{"role": "user", "content": "When did the Shang Dynasty fall?"},
]
response_1 = litellm.completion(model=model_name, messages=messages_1)
response_2 = litellm.completion(model=model_name, messages=message_2)
# Add any assertions here to check the response
print(response_2.usage)
```

124
docs/my-website/docs/completion/usage.md
View file

@ -49,127 +49,3 @@ for chunk in completion:
print(chunk.choices[0].delta) print(chunk.choices[0].delta)
``` ```
## Prompt Caching
For Anthropic + Deepseek, LiteLLM follows the Anthropic prompt caching usage object format:
```bash
"usage": {
"prompt_tokens": int,
"completion_tokens": int,
"total_tokens": int,
"_cache_creation_input_tokens": int, # hidden param for prompt caching. Might change, once openai introduces their equivalent.
"_cache_read_input_tokens": int # hidden param for prompt caching. Might change, once openai introduces their equivalent.
}
```
- `prompt_tokens`: These are the non-cached prompt tokens (same as Anthropic, equivalent to Deepseek `prompt_cache_miss_tokens`).
- `completion_tokens`: These are the output tokens generated by the model.
- `total_tokens`: Sum of prompt_tokens + completion_tokens.
- `_cache_creation_input_tokens`: Input tokens that were written to cache. (Anthropic only).
- `_cache_read_input_tokens`: Input tokens that were read from cache for that call. (equivalent to Deepseek `prompt_cache_hit_tokens`).
### Anthropic Example
```python
from litellm import completion
import litellm
import os
litellm.set_verbose = True # 👈 SEE RAW REQUEST
os.environ["ANTHROPIC_API_KEY"] = ""
response = completion(
model="anthropic/claude-3-5-sonnet-20240620",
messages=[
{
"role": "system",
"content": [
{
"type": "text",
"text": "You are an AI assistant tasked with analyzing legal documents.",
},
{
"type": "text",
"text": "Here is the full text of a complex legal agreement" * 400,
"cache_control": {"type": "ephemeral"},
},
],
},
{
"role": "user",
"content": "what are the key terms and conditions in this agreement?",
},
]
)
print(response.usage)
```
### Deepeek Example
```python
from litellm import completion
import litellm
import os
os.environ["DEEPSEEK_API_KEY"] = ""
litellm.set_verbose = True # 👈 SEE RAW REQUEST
model_name = "deepseek/deepseek-chat"
messages_1 = [
{
"role": "system",
"content": "You are a history expert. The user will provide a series of questions, and your answers should be concise and start with `Answer:`",
},
{
"role": "user",
"content": "In what year did Qin Shi Huang unify the six states?",
},
{"role": "assistant", "content": "Answer: 221 BC"},
{"role": "user", "content": "Who was the founder of the Han Dynasty?"},
{"role": "assistant", "content": "Answer: Liu Bang"},
{"role": "user", "content": "Who was the last emperor of the Tang Dynasty?"},
{"role": "assistant", "content": "Answer: Li Zhu"},
{
"role": "user",
"content": "Who was the founding emperor of the Ming Dynasty?",
},
{"role": "assistant", "content": "Answer: Zhu Yuanzhang"},
{
"role": "user",
"content": "Who was the founding emperor of the Qing Dynasty?",
},
]
message_2 = [
{
"role": "system",
"content": "You are a history expert. The user will provide a series of questions, and your answers should be concise and start with `Answer:`",
},
{
"role": "user",
"content": "In what year did Qin Shi Huang unify the six states?",
},
{"role": "assistant", "content": "Answer: 221 BC"},
{"role": "user", "content": "Who was the founder of the Han Dynasty?"},
{"role": "assistant", "content": "Answer: Liu Bang"},
{"role": "user", "content": "Who was the last emperor of the Tang Dynasty?"},
{"role": "assistant", "content": "Answer: Li Zhu"},
{
"role": "user",
"content": "Who was the founding emperor of the Ming Dynasty?",
},
{"role": "assistant", "content": "Answer: Zhu Yuanzhang"},
{"role": "user", "content": "When did the Shang Dynasty fall?"},
]
response_1 = litellm.completion(model=model_name, messages=messages_1)
response_2 = litellm.completion(model=model_name, messages=message_2)
# Add any assertions here to check the response
print(response_2.usage)
```

1
docs/my-website/sidebars.js
View file

@ -198,6 +198,7 @@ const sidebars = {
"completion/drop_params", "completion/drop_params",
"completion/prompt_formatting", "completion/prompt_formatting",
"completion/output", "completion/output",
"completion/prompt_caching",
"completion/usage", "completion/usage",
"exception_mapping", "exception_mapping",
"completion/stream", "completion/stream",

100
litellm/litellm_core_utils/exception_mapping_utils.py
View file

@ -750,6 +750,17 @@ def exception_type( # type: ignore
model=model, model=model,
llm_provider="bedrock", llm_provider="bedrock",
) )
elif (
"Conversation blocks and tool result blocks cannot be provided in the same turn."
in error_str
):
exception_mapping_worked = True
raise BadRequestError(
message=f"BedrockException - {error_str}\n. Enable 'litellm.modify_params=True' (for PROXY do: `litellm_settings::modify_params: True`) to insert a dummy assistant message and fix this error.",
model=model,
llm_provider="bedrock",
response=original_exception.response,
)
elif "Malformed input request" in error_str: elif "Malformed input request" in error_str:
exception_mapping_worked = True exception_mapping_worked = True
raise BadRequestError( raise BadRequestError(
@ -895,7 +906,10 @@ def exception_type( # type: ignore
llm_provider=custom_llm_provider, llm_provider=custom_llm_provider,
litellm_debug_info=extra_information, litellm_debug_info=extra_information,
) )
elif custom_llm_provider == "sagemaker": elif (
custom_llm_provider == "sagemaker"
or custom_llm_provider == "sagemaker_chat"
):
if "Unable to locate credentials" in error_str: if "Unable to locate credentials" in error_str:
exception_mapping_worked = True exception_mapping_worked = True
raise BadRequestError( raise BadRequestError(
@ -926,6 +940,90 @@ def exception_type( # type: ignore
llm_provider="sagemaker", llm_provider="sagemaker",
response=original_exception.response, response=original_exception.response,
) )
elif hasattr(original_exception, "status_code"):
if original_exception.status_code == 500:
exception_mapping_worked = True
raise ServiceUnavailableError(
message=f"SagemakerException - {original_exception.message}",
llm_provider=custom_llm_provider,
model=model,
response=httpx.Response(
status_code=500,
request=httpx.Request(
method="POST", url="https://api.openai.com/v1/"
),
),
)
elif original_exception.status_code == 401:
exception_mapping_worked = True
raise AuthenticationError(
message=f"SagemakerException - {original_exception.message}",
llm_provider=custom_llm_provider,
model=model,
response=original_exception.response,
)
elif original_exception.status_code == 400:
exception_mapping_worked = True
raise BadRequestError(
message=f"SagemakerException - {original_exception.message}",
llm_provider=custom_llm_provider,
model=model,
response=original_exception.response,
)
elif original_exception.status_code == 404:
exception_mapping_worked = True
raise NotFoundError(
message=f"SagemakerException - {original_exception.message}",
llm_provider=custom_llm_provider,
model=model,
response=original_exception.response,
)
elif original_exception.status_code == 408:
exception_mapping_worked = True
raise Timeout(
message=f"SagemakerException - {original_exception.message}",
model=model,
llm_provider=custom_llm_provider,
litellm_debug_info=extra_information,
)
elif (
original_exception.status_code == 422
or original_exception.status_code == 424
):
exception_mapping_worked = True
raise BadRequestError(
message=f"SagemakerException - {original_exception.message}",
model=model,
llm_provider=custom_llm_provider,
response=original_exception.response,
litellm_debug_info=extra_information,
)
elif original_exception.status_code == 429:
exception_mapping_worked = True
raise RateLimitError(
message=f"SagemakerException - {original_exception.message}",
model=model,
llm_provider=custom_llm_provider,
response=original_exception.response,
litellm_debug_info=extra_information,
)
elif original_exception.status_code == 503:
exception_mapping_worked = True
raise ServiceUnavailableError(
message=f"SagemakerException - {original_exception.message}",
model=model,
llm_provider=custom_llm_provider,
response=original_exception.response,
litellm_debug_info=extra_information,
)
elif original_exception.status_code == 504: # gateway timeout error
exception_mapping_worked = True
raise Timeout(
message=f"SagemakerException - {original_exception.message}",
model=model,
llm_provider=custom_llm_provider,
litellm_debug_info=extra_information,
)
elif ( elif (
custom_llm_provider == "vertex_ai" custom_llm_provider == "vertex_ai"
or custom_llm_provider == "vertex_ai_beta" or custom_llm_provider == "vertex_ai_beta"

41
litellm/litellm_core_utils/litellm_logging.py
View file

@ -363,7 +363,7 @@ class Logging:
for param in _supported_callback_params: for param in _supported_callback_params:
if param in kwargs: if param in kwargs:
_param_value = kwargs.pop(param) _param_value = kwargs.pop(param)
if "os.environ/" in _param_value: if _param_value is not None and "os.environ/" in _param_value:
_param_value = get_secret_str(secret_name=_param_value) _param_value = get_secret_str(secret_name=_param_value)
standard_callback_dynamic_params[param] = _param_value # type: ignore standard_callback_dynamic_params[param] = _param_value # type: ignore
return standard_callback_dynamic_params return standard_callback_dynamic_params
@ -632,7 +632,12 @@ class Logging:
) )
) )
original_response = redact_message_input_output_from_logging( original_response = redact_message_input_output_from_logging(
litellm_logging_obj=self, result=original_response model_call_details=(
self.model_call_details
if hasattr(self, "model_call_details")
else {}
),
result=original_response,
) )
# Input Integration Logging -> If you want to log the fact that an attempt to call the model was made # Input Integration Logging -> If you want to log the fact that an attempt to call the model was made
@ -933,7 +938,12 @@ class Logging:
callbacks = litellm.success_callback callbacks = litellm.success_callback
result = redact_message_input_output_from_logging( result = redact_message_input_output_from_logging(
result=result, litellm_logging_obj=self model_call_details=(
self.model_call_details
if hasattr(self, "model_call_details")
else {}
),
result=result,
) )
## LOGGING HOOK ## ## LOGGING HOOK ##
@ -1591,7 +1601,10 @@ class Logging:
callbacks = litellm._async_success_callback callbacks = litellm._async_success_callback
result = redact_message_input_output_from_logging( result = redact_message_input_output_from_logging(
result=result, litellm_logging_obj=self model_call_details=(
self.model_call_details if hasattr(self, "model_call_details") else {}
),
result=result,
) )
## LOGGING HOOK ## ## LOGGING HOOK ##
@ -1899,7 +1912,12 @@ class Logging:
result = None # result sent to all loggers, init this to None incase it's not created result = None # result sent to all loggers, init this to None incase it's not created
result = redact_message_input_output_from_logging( result = redact_message_input_output_from_logging(
result=result, litellm_logging_obj=self model_call_details=(
self.model_call_details
if hasattr(self, "model_call_details")
else {}
),
result=result,
) )
for callback in callbacks: for callback in callbacks:
try: try:
@ -2747,8 +2765,17 @@ def get_standard_logging_object_payload(
else: else:
final_response_obj = None final_response_obj = None
if litellm.turn_off_message_logging: modified_final_response_obj = redact_message_input_output_from_logging(
final_response_obj = "redacted-by-litellm" model_call_details=kwargs,
result=final_response_obj,
)
if modified_final_response_obj is not None and isinstance(
modified_final_response_obj, BaseModel
):
final_response_obj = modified_final_response_obj.model_dump()
else:
final_response_obj = modified_final_response_obj
payload: StandardLoggingPayload = StandardLoggingPayload( payload: StandardLoggingPayload = StandardLoggingPayload(
id=str(id), id=str(id),

33
litellm/litellm_core_utils/redact_messages.py
View file

@ -8,7 +8,7 @@
# Thank you users! We ❤️ you! - Krrish & Ishaan # Thank you users! We ❤️ you! - Krrish & Ishaan
import copy import copy
from typing import TYPE_CHECKING, Any from typing import TYPE_CHECKING, Any, Optional
import litellm import litellm
from litellm.integrations.custom_logger import CustomLogger from litellm.integrations.custom_logger import CustomLogger
@ -30,29 +30,27 @@ def redact_message_input_output_from_custom_logger(
hasattr(custom_logger, "message_logging") hasattr(custom_logger, "message_logging")
and custom_logger.message_logging is not True and custom_logger.message_logging is not True
): ):
return perform_redaction(litellm_logging_obj, result) return perform_redaction(litellm_logging_obj.model_call_details, result)
return result return result
def perform_redaction(litellm_logging_obj: LiteLLMLoggingObject, result): def perform_redaction(model_call_details: dict, result):
""" """
Performs the actual redaction on the logging object and result. Performs the actual redaction on the logging object and result.
""" """
# Redact model_call_details # Redact model_call_details
litellm_logging_obj.model_call_details["messages"] = [ model_call_details["messages"] = [
{"role": "user", "content": "redacted-by-litellm"} {"role": "user", "content": "redacted-by-litellm"}
] ]
litellm_logging_obj.model_call_details["prompt"] = "" model_call_details["prompt"] = ""
litellm_logging_obj.model_call_details["input"] = "" model_call_details["input"] = ""
# Redact streaming response # Redact streaming response
if ( if (
litellm_logging_obj.stream is True model_call_details.get("stream", False) is True
and "complete_streaming_response" in litellm_logging_obj.model_call_details and "complete_streaming_response" in model_call_details
): ):
_streaming_response = litellm_logging_obj.model_call_details[ _streaming_response = model_call_details["complete_streaming_response"]
"complete_streaming_response"
]
for choice in _streaming_response.choices: for choice in _streaming_response.choices:
if isinstance(choice, litellm.Choices): if isinstance(choice, litellm.Choices):
choice.message.content = "redacted-by-litellm" choice.message.content = "redacted-by-litellm"
@ -69,22 +67,19 @@ def perform_redaction(litellm_logging_obj: LiteLLMLoggingObject, result):
elif isinstance(choice, litellm.utils.StreamingChoices): elif isinstance(choice, litellm.utils.StreamingChoices):
choice.delta.content = "redacted-by-litellm" choice.delta.content = "redacted-by-litellm"
return _result return _result
else:
return result return "redacted-by-litellm"
def redact_message_input_output_from_logging( def redact_message_input_output_from_logging(
litellm_logging_obj: LiteLLMLoggingObject, result model_call_details: dict, result, input: Optional[Any] = None
): ):
""" """
Removes messages, prompts, input, response from logging. This modifies the data in-place Removes messages, prompts, input, response from logging. This modifies the data in-place
only redacts when litellm.turn_off_message_logging == True only redacts when litellm.turn_off_message_logging == True
""" """
_request_headers = ( _request_headers = (
litellm_logging_obj.model_call_details.get("litellm_params", {}).get( model_call_details.get("litellm_params", {}).get("metadata", {}) or {}
"metadata", {}
)
or {}
) )
request_headers = _request_headers.get("headers", {}) request_headers = _request_headers.get("headers", {})
@ -101,7 +96,7 @@ def redact_message_input_output_from_logging(
): ):
return result return result
return perform_redaction(litellm_logging_obj, result) return perform_redaction(model_call_details, result)
def redact_user_api_key_info(metadata: dict) -> dict: def redact_user_api_key_info(metadata: dict) -> dict:

19
litellm/llms/anthropic/chat/handler.py
View file

@ -43,7 +43,7 @@ from litellm.types.llms.openai import (
ChatCompletionToolCallFunctionChunk, ChatCompletionToolCallFunctionChunk,
ChatCompletionUsageBlock, ChatCompletionUsageBlock,
) )
from litellm.types.utils import GenericStreamingChunk from litellm.types.utils import GenericStreamingChunk, PromptTokensDetails
from litellm.utils import CustomStreamWrapper, ModelResponse, Usage from litellm.utils import CustomStreamWrapper, ModelResponse, Usage
from ...base import BaseLLM from ...base import BaseLLM
@ -283,19 +283,28 @@ class AnthropicChatCompletion(BaseLLM):
completion_tokens = completion_response["usage"]["output_tokens"] completion_tokens = completion_response["usage"]["output_tokens"]
_usage = completion_response["usage"] _usage = completion_response["usage"]
total_tokens = prompt_tokens + completion_tokens total_tokens = prompt_tokens + completion_tokens
cache_creation_input_tokens: int = 0
cache_read_input_tokens: int = 0
model_response.created = int(time.time()) model_response.created = int(time.time())
model_response.model = model model_response.model = model
if "cache_creation_input_tokens" in _usage:
cache_creation_input_tokens = _usage["cache_creation_input_tokens"]
if "cache_read_input_tokens" in _usage:
cache_read_input_tokens = _usage["cache_read_input_tokens"]
prompt_tokens_details = PromptTokensDetails(
cached_tokens=cache_read_input_tokens
)
usage = Usage( usage = Usage(
prompt_tokens=prompt_tokens, prompt_tokens=prompt_tokens,
completion_tokens=completion_tokens, completion_tokens=completion_tokens,
total_tokens=total_tokens, total_tokens=total_tokens,
prompt_tokens_details=prompt_tokens_details,
cache_creation_input_tokens=cache_creation_input_tokens,
cache_read_input_tokens=cache_read_input_tokens,
) )
if "cache_creation_input_tokens" in _usage:
usage["cache_creation_input_tokens"] = _usage["cache_creation_input_tokens"]
if "cache_read_input_tokens" in _usage:
usage["cache_read_input_tokens"] = _usage["cache_read_input_tokens"]
setattr(model_response, "usage", usage) # type: ignore setattr(model_response, "usage", usage) # type: ignore
model_response._hidden_params = _hidden_params model_response._hidden_params = _hidden_params

68
litellm/llms/databricks/chat.py
View file

@ -14,7 +14,11 @@ import requests # type: ignore
import litellm import litellm
from litellm.litellm_core_utils.core_helpers import map_finish_reason from litellm.litellm_core_utils.core_helpers import map_finish_reason
from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler, HTTPHandler from litellm.llms.custom_httpx.http_handler import (
AsyncHTTPHandler,
HTTPHandler,
get_async_httpx_client,
)
from litellm.llms.databricks.exceptions import DatabricksError from litellm.llms.databricks.exceptions import DatabricksError
from litellm.llms.databricks.streaming_utils import ModelResponseIterator from litellm.llms.databricks.streaming_utils import ModelResponseIterator
from litellm.types.llms.openai import ( from litellm.types.llms.openai import (
@ -167,7 +171,7 @@ class DatabricksEmbeddingConfig:
async def make_call( async def make_call(
client: AsyncHTTPHandler, client: Optional[AsyncHTTPHandler],
api_base: str, api_base: str,
headers: dict, headers: dict,
data: str, data: str,
@ -176,6 +180,10 @@ async def make_call(
logging_obj, logging_obj,
streaming_decoder: Optional[CustomStreamingDecoder] = None, streaming_decoder: Optional[CustomStreamingDecoder] = None,
): ):
if client is None:
client = get_async_httpx_client(
llm_provider=litellm.LlmProviders.DATABRICKS
) # Create a new client if none provided
response = await client.post(api_base, headers=headers, data=data, stream=True) response = await client.post(api_base, headers=headers, data=data, stream=True)
if response.status_code != 200: if response.status_code != 200:
@ -211,7 +219,7 @@ def make_sync_call(
streaming_decoder: Optional[CustomStreamingDecoder] = None, streaming_decoder: Optional[CustomStreamingDecoder] = None,
): ):
if client is None: if client is None:
client = HTTPHandler() # Create a new client if none provided client = litellm.module_level_client # Create a new client if none provided
response = client.post(api_base, headers=headers, data=data, stream=True) response = client.post(api_base, headers=headers, data=data, stream=True)
@ -343,18 +351,18 @@ class DatabricksChatCompletion(BaseLLM):
) -> CustomStreamWrapper: ) -> CustomStreamWrapper:
data["stream"] = True data["stream"] = True
completion_stream = await make_call(
client=client,
api_base=api_base,
headers=headers,
data=json.dumps(data),
model=model,
messages=messages,
logging_obj=logging_obj,
streaming_decoder=streaming_decoder,
)
streamwrapper = CustomStreamWrapper( streamwrapper = CustomStreamWrapper(
completion_stream=None, completion_stream=completion_stream,
make_call=partial(
make_call,
api_base=api_base,
headers=headers,
data=json.dumps(data),
model=model,
messages=messages,
logging_obj=logging_obj,
streaming_decoder=streaming_decoder,
),
model=model, model=model,
custom_llm_provider=custom_llm_provider, custom_llm_provider=custom_llm_provider,
logging_obj=logging_obj, logging_obj=logging_obj,
@ -530,28 +538,32 @@ class DatabricksChatCompletion(BaseLLM):
base_model=base_model, base_model=base_model,
) )
else: else:
if client is None or not isinstance(client, HTTPHandler):
client = HTTPHandler(timeout=timeout) # type: ignore
## COMPLETION CALL ## COMPLETION CALL
if stream is True: if stream is True:
return CustomStreamWrapper( completion_stream = make_sync_call(
completion_stream=None, client=(
make_call=partial( client
make_sync_call, if client is not None and isinstance(client, HTTPHandler)
client=None, else None
api_base=api_base,
headers=headers, # type: ignore
data=json.dumps(data),
model=model,
messages=messages,
logging_obj=logging_obj,
streaming_decoder=streaming_decoder,
), ),
api_base=api_base,
headers=headers,
data=json.dumps(data),
model=model,
messages=messages,
logging_obj=logging_obj,
streaming_decoder=streaming_decoder,
)
# completion_stream.__iter__()
return CustomStreamWrapper(
completion_stream=completion_stream,
model=model, model=model,
custom_llm_provider=custom_llm_provider, custom_llm_provider=custom_llm_provider,
logging_obj=logging_obj, logging_obj=logging_obj,
) )
else: else:
if client is None or not isinstance(client, HTTPHandler):
client = HTTPHandler(timeout=timeout) # type: ignore
try: try:
response = client.post( response = client.post(
api_base, headers=headers, data=json.dumps(data) api_base, headers=headers, data=json.dumps(data)

10
litellm/llms/databricks/streaming_utils.py
View file

@ -54,10 +54,10 @@ class ModelResponseIterator:
is_finished = True is_finished = True
finish_reason = processed_chunk.choices[0].finish_reason finish_reason = processed_chunk.choices[0].finish_reason
if hasattr(processed_chunk, "usage") and isinstance( usage_chunk: Optional[litellm.Usage] = getattr(
processed_chunk.usage, litellm.Usage processed_chunk, "usage", None
): )
usage_chunk: litellm.Usage = processed_chunk.usage if usage_chunk is not None:
usage = ChatCompletionUsageBlock( usage = ChatCompletionUsageBlock(
prompt_tokens=usage_chunk.prompt_tokens, prompt_tokens=usage_chunk.prompt_tokens,
@ -82,6 +82,8 @@ class ModelResponseIterator:
return self return self
def __next__(self): def __next__(self):
if not hasattr(self, "response_iterator"):
self.response_iterator = self.streaming_response
try: try:
chunk = self.response_iterator.__next__() chunk = self.response_iterator.__next__()
except StopIteration: except StopIteration:

135
litellm/llms/prompt_templates/factory.py
View file

@ -62,7 +62,11 @@ DEFAULT_USER_CONTINUE_MESSAGE = {
# used to interweave assistant messages, to ensure user/assistant alternating # used to interweave assistant messages, to ensure user/assistant alternating
DEFAULT_ASSISTANT_CONTINUE_MESSAGE = { DEFAULT_ASSISTANT_CONTINUE_MESSAGE = {
"role": "assistant", "role": "assistant",
"content": "Please continue.", "content": [
{
"text": "Please continue.",
}
],
} # similar to autogen. Only used if `litellm.modify_params=True`. } # similar to autogen. Only used if `litellm.modify_params=True`.
@ -2354,17 +2358,40 @@ def _convert_to_bedrock_tool_call_result(
return content_block return content_block
def _insert_assistant_continue_message(
messages: List[BedrockMessageBlock],
assistant_continue_message: Optional[str] = None,
) -> List[BedrockMessageBlock]:
"""
Add dummy message between user/tool result blocks.
Conversation blocks and tool result blocks cannot be provided in the same turn. Issue: https://github.com/BerriAI/litellm/issues/6053
"""
if assistant_continue_message is not None:
messages.append(
BedrockMessageBlock(
role="assistant",
content=[BedrockContentBlock(text=assistant_continue_message)],
)
)
elif litellm.modify_params:
messages.append(BedrockMessageBlock(**DEFAULT_ASSISTANT_CONTINUE_MESSAGE)) # type: ignore
return messages
def _bedrock_converse_messages_pt( def _bedrock_converse_messages_pt(
messages: List, messages: List,
model: str, model: str,
llm_provider: str, llm_provider: str,
user_continue_message: Optional[dict] = None, user_continue_message: Optional[dict] = None,
assistant_continue_message: Optional[str] = None,
) -> List[BedrockMessageBlock]: ) -> List[BedrockMessageBlock]:
""" """
Converts given messages from OpenAI format to Bedrock format Converts given messages from OpenAI format to Bedrock format
- Roles must alternate b/w 'user' and 'model' (same as anthropic -> merge consecutive roles) - Roles must alternate b/w 'user' and 'model' (same as anthropic -> merge consecutive roles)
- Please ensure that function response turn comes immediately after a function call turn - Please ensure that function response turn comes immediately after a function call turn
- Conversation blocks and tool result blocks cannot be provided in the same turn. Issue: https://github.com/BerriAI/litellm/issues/6053
""" """
contents: List[BedrockMessageBlock] = [] contents: List[BedrockMessageBlock] = []
@ -2387,37 +2414,78 @@ def _bedrock_converse_messages_pt(
while msg_i < len(messages): while msg_i < len(messages):
user_content: List[BedrockContentBlock] = [] user_content: List[BedrockContentBlock] = []
init_msg_i = msg_i init_msg_i = msg_i
valid_user_roles = ["user", "tool", "function"] ## MERGE CONSECUTIVE USER CONTENT ##
## MERGE CONSECUTIVE USER + TOOL CONTENT ## while msg_i < len(messages) and messages[msg_i]["role"] == "user":
while msg_i < len(messages) and messages[msg_i]["role"] in valid_user_roles: if isinstance(messages[msg_i]["content"], list):
if messages[msg_i]["role"] == "user": _parts: List[BedrockContentBlock] = []
if isinstance(messages[msg_i]["content"], list): for element in messages[msg_i]["content"]:
_parts: List[BedrockContentBlock] = [] if isinstance(element, dict):
for element in messages[msg_i]["content"]: if element["type"] == "text":
if isinstance(element, dict): _part = BedrockContentBlock(text=element["text"])
if element["type"] == "text": _parts.append(_part)
_part = BedrockContentBlock(text=element["text"]) elif element["type"] == "image_url":
_parts.append(_part) image_url = element["image_url"]["url"]
elif element["type"] == "image_url": _part = _process_bedrock_converse_image_block( # type: ignore
image_url = element["image_url"]["url"] image_url=image_url
_part = _process_bedrock_converse_image_block( # type: ignore )
image_url=image_url _parts.append(BedrockContentBlock(image=_part)) # type: ignore
) user_content.extend(_parts)
_parts.append(BedrockContentBlock(image=_part)) # type: ignore else:
user_content.extend(_parts) _part = BedrockContentBlock(text=messages[msg_i]["content"])
elif isinstance(messages[msg_i]["content"], str): user_content.append(_part)
_part = BedrockContentBlock(text=messages[msg_i]["content"])
user_content.append(_part)
elif (
messages[msg_i]["role"] == "tool"
or messages[msg_i]["role"] == "function"
):
tool_call_result = _convert_to_bedrock_tool_call_result(messages[msg_i])
user_content.append(tool_call_result)
msg_i += 1
msg_i += 1
if user_content: if user_content:
contents.append(BedrockMessageBlock(role="user", content=user_content)) if len(contents) > 0 and contents[-1]["role"] == "user":
if (
assistant_continue_message is not None
or litellm.modify_params is True
):
# if last message was a 'user' message, then add a dummy assistant message (bedrock requires alternating roles)
contents = _insert_assistant_continue_message(
messages=contents,
assistant_continue_message=assistant_continue_message,
)
contents.append(
BedrockMessageBlock(role="user", content=user_content)
)
else:
verbose_logger.warning(
"Potential consecutive user/tool blocks. Trying to merge. If error occurs, please set a 'assistant_continue_message' or set 'modify_params=True' to insert a dummy assistant message for bedrock calls."
)
contents[-1]["content"].extend(user_content)
else:
contents.append(BedrockMessageBlock(role="user", content=user_content))
## MERGE CONSECUTIVE TOOL CALL MESSAGES ##
tool_content: List[BedrockContentBlock] = []
while msg_i < len(messages) and messages[msg_i]["role"] == "tool":
tool_call_result = _convert_to_bedrock_tool_call_result(messages[msg_i])
tool_content.append(tool_call_result)
msg_i += 1
if tool_content:
# if last message was a 'user' message, then add a blank assistant message (bedrock requires alternating roles)
if len(contents) > 0 and contents[-1]["role"] == "user":
if (
assistant_continue_message is not None
or litellm.modify_params is True
):
# if last message was a 'user' message, then add a dummy assistant message (bedrock requires alternating roles)
contents = _insert_assistant_continue_message(
messages=contents,
assistant_continue_message=assistant_continue_message,
)
contents.append(
BedrockMessageBlock(role="user", content=tool_content)
)
else:
verbose_logger.warning(
"Potential consecutive user/tool blocks. Trying to merge. If error occurs, please set a 'assistant_continue_message' or set 'modify_params=True' to insert a dummy assistant message for bedrock calls."
)
contents[-1]["content"].extend(tool_content)
else:
contents.append(BedrockMessageBlock(role="user", content=tool_content))
assistant_content: List[BedrockContentBlock] = [] assistant_content: List[BedrockContentBlock] = []
## MERGE CONSECUTIVE ASSISTANT CONTENT ## ## MERGE CONSECUTIVE ASSISTANT CONTENT ##
while msg_i < len(messages) and messages[msg_i]["role"] == "assistant": while msg_i < len(messages) and messages[msg_i]["role"] == "assistant":
@ -2555,10 +2623,9 @@ def _bedrock_tools_pt(tools: List) -> List[BedrockToolBlock]:
""" """
tool_block_list: List[BedrockToolBlock] = [] tool_block_list: List[BedrockToolBlock] = []
for tool in tools: for tool in tools:
parameters = tool.get("function", {}).get("parameters", { parameters = tool.get("function", {}).get(
"type": "object", "parameters", {"type": "object", "properties": {}}
"properties": {} )
})
name = tool.get("function", {}).get("name", "") name = tool.get("function", {}).get("name", "")
# related issue: https://github.com/BerriAI/litellm/issues/5007 # related issue: https://github.com/BerriAI/litellm/issues/5007

15
litellm/llms/sagemaker/sagemaker.py
View file

@ -633,15 +633,14 @@ class SagemakerLLM(BaseAWSLLM):
"aws_region_name": aws_region_name, "aws_region_name": aws_region_name,
} }
prepared_request = await asyncified_prepare_request(**prepared_request_args) prepared_request = await asyncified_prepare_request(**prepared_request_args)
completion_stream = await self.make_async_call(
api_base=prepared_request.url,
headers=prepared_request.headers, # type: ignore
data=data,
logging_obj=logging_obj,
)
streaming_response = CustomStreamWrapper( streaming_response = CustomStreamWrapper(
completion_stream=None, completion_stream=completion_stream,
make_call=partial(
self.make_async_call,
api_base=prepared_request.url,
headers=prepared_request.headers, # type: ignore
data=data,
logging_obj=logging_obj,
),
model=model, model=model,
custom_llm_provider="sagemaker", custom_llm_provider="sagemaker",
logging_obj=logging_obj, logging_obj=logging_obj,

79
litellm/proxy/_new_secret_config.yaml
View file

@ -1,12 +1,73 @@
model_list: model_list:
- model_name: gpt-4o - model_name: fake-claude-endpoint
litellm_params: litellm_params:
model: azure/gpt-4o-realtime-preview model: anthropic.claude-3-sonnet-20240229-v1:0
api_key: os.environ/AZURE_SWEDEN_API_KEY api_base: https://exampleopenaiendpoint-production.up.railway.app
api_base: os.environ/AZURE_SWEDEN_API_BASE aws_secret_access_key: os.environ/AWS_SECRET_ACCESS_KEY
aws_access_key_id: os.environ/AWS_ACCESS_KEY_ID
- model_name: gemini-vision
litellm_params:
model: vertex_ai/gemini-1.0-pro-vision-001
api_base: https://exampleopenaiendpoint-production.up.railway.app/v1/projects/adroit-crow-413218/locations/us-central1/publishers/google/models/gemini-1.0-pro-vision-001
vertex_project: "adroit-crow-413218"
vertex_location: "us-central1"
- model_name: fake-azure-endpoint
litellm_params:
model: openai/429
api_key: fake-key
api_base: https://exampleopenaiendpoint-production.up.railway.app
- model_name: fake-openai-endpoint
litellm_params:
model: gpt-3.5-turbo
api_base: https://exampleopenaiendpoint-production.up.railway.app
- model_name: o1-preview
litellm_params:
model: o1-preview
- model_name: rerank-english-v3.0
litellm_params:
model: cohere/rerank-english-v3.0
api_key: os.environ/COHERE_API_KEY
- model_name: azure-rerank-english-v3.0
litellm_params:
model: azure_ai/rerank-english-v3.0
api_base: os.environ/AZURE_AI_COHERE_API_BASE
api_key: os.environ/AZURE_AI_COHERE_API_KEY
- model_name: "databricks/*"
litellm_params:
model: "databricks/*"
api_key: os.environ/DATABRICKS_API_KEY
api_base: os.environ/DATABRICKS_API_BASE
- model_name: "anthropic/*"
litellm_params:
model: "anthropic/*"
- model_name: "*"
litellm_params:
model: "openai/*"
- model_name: "fireworks_ai/*"
litellm_params:
model: "fireworks_ai/*"
configurable_clientside_auth_params: ["api_base"]
- model_name: "gemini-flash-experimental"
litellm_params:
model: "vertex_ai/gemini-flash-experimental"
- model_name: openai-gpt-4o-realtime-audio litellm_settings:
litellm_params: turn_off_message_logging: true
model: openai/gpt-4o-realtime-preview-2024-10-01 # callbacks:
api_key: os.environ/OPENAI_API_KEY # - prometheus
api_base: http://localhost:8080 # - otel
failure_callback:
- sentry
- prometheus
success_callback:
- prometheus
- s3
s3_callback_params:
s3_bucket_name: mytestbucketlitellm # AWS Bucket Name for S3
s3_region_name: us-west-2 # AWS Region Name for S3
s3_aws_access_key_id: os.environ/AWS_ACCESS_KEY_ID # us os.environ/<variable name> to pass environment variables. This is AWS Access Key ID for S3
s3_aws_secret_access_key: os.environ/AWS_SECRET_ACCESS_KEY # AWS Secret Access Key for S3
general_settings:
db_url: os.environ/DATABASE_URL
# disable_prisma_schema_update: true

28
litellm/types/utils.py
View file

@ -319,7 +319,7 @@ ChatCompletionMessage(content='This is a test', role='assistant', function_call=
class Message(OpenAIObject): class Message(OpenAIObject):
content: Optional[str] content: Optional[str]
role: Literal["assistant"] role: Literal["assistant", "user", "system", "tool", "function"]
tool_calls: Optional[List[ChatCompletionMessageToolCall]] tool_calls: Optional[List[ChatCompletionMessageToolCall]]
function_call: Optional[FunctionCall] function_call: Optional[FunctionCall]
@ -488,14 +488,6 @@ class Usage(CompletionUsage):
] = None, ] = None,
**params, **params,
): ):
## DEEPSEEK PROMPT TOKEN HANDLING ## - follow the anthropic format, of having prompt tokens be just the non-cached token input. Enables accurate cost-tracking - Relevant issue: https://github.com/BerriAI/litellm/issues/5285
if (
"prompt_cache_miss_tokens" in params
and isinstance(params["prompt_cache_miss_tokens"], int)
and prompt_tokens is not None
):
prompt_tokens = params["prompt_cache_miss_tokens"]
# handle reasoning_tokens # handle reasoning_tokens
_completion_tokens_details: Optional[CompletionTokensDetails] = None _completion_tokens_details: Optional[CompletionTokensDetails] = None
if reasoning_tokens: if reasoning_tokens:
@ -512,6 +504,24 @@ class Usage(CompletionUsage):
elif isinstance(completion_tokens_details, CompletionTokensDetails): elif isinstance(completion_tokens_details, CompletionTokensDetails):
_completion_tokens_details = completion_tokens_details _completion_tokens_details = completion_tokens_details
## DEEPSEEK MAPPING ##
if "prompt_cache_hit_tokens" in params and isinstance(
params["prompt_cache_hit_tokens"], int
):
if prompt_tokens_details is None:
prompt_tokens_details = PromptTokensDetails(
cached_tokens=params["prompt_cache_hit_tokens"]
)
## ANTHROPIC MAPPING ##
if "cache_read_input_tokens" in params and isinstance(
params["cache_read_input_tokens"], int
):
if prompt_tokens_details is None:
prompt_tokens_details = PromptTokensDetails(
cached_tokens=params["cache_read_input_tokens"]
)
# handle prompt_tokens_details # handle prompt_tokens_details
_prompt_tokens_details: Optional[PromptTokensDetails] = None _prompt_tokens_details: Optional[PromptTokensDetails] = None
if prompt_tokens_details: if prompt_tokens_details:

285
tests/local_testing/test_bedrock_completion.py
View file

@ -1633,3 +1633,288 @@ def test_bedrock_completion_test_3():
} }
} }
] ]
@pytest.mark.parametrize("modify_params", [True, False])
def test_bedrock_completion_test_4(modify_params):
litellm.set_verbose = True
litellm.modify_params = modify_params
data = {
"model": "anthropic.claude-3-opus-20240229-v1:0",
"messages": [
{
"role": "user",
"content": [
{"type": "text", "text": "<task>\nWhat is this file?\n</task>"},
{
"type": "text",
"text": "<environment_details>\n# VSCode Visible Files\ncomputer-vision/hm-open3d/src/main.py\n\n# VSCode Open Tabs\ncomputer-vision/hm-open3d/src/main.py\n\n# Current Working Directory (/Users/hongbo-miao/Clouds/Git/hongbomiao.com) Files\n.ansible-lint\n.clang-format\n.cmakelintrc\n.dockerignore\n.editorconfig\n.gitignore\n.gitmodules\n.hadolint.yaml\n.isort.cfg\n.markdownlint-cli2.jsonc\n.mergify.yml\n.npmrc\n.nvmrc\n.prettierignore\n.rubocop.yml\n.ruby-version\n.ruff.toml\n.shellcheckrc\n.solhint.json\n.solhintignore\n.sqlfluff\n.sqlfluffignore\n.stylelintignore\n.yamllint.yaml\nCODE_OF_CONDUCT.md\ncommitlint.config.js\nGemfile\nGemfile.lock\nLICENSE\nlint-staged.config.js\nMakefile\nmiss_hit.cfg\nmypy.ini\npackage-lock.json\npackage.json\npoetry.lock\npoetry.toml\nprettier.config.js\npyproject.toml\nREADME.md\nrelease.config.js\nrenovate.json\nSECURITY.md\nstylelint.config.js\naerospace/\naerospace/air-defense-system/\naerospace/hm-aerosandbox/\naerospace/hm-openaerostruct/\naerospace/px4/\naerospace/quadcopter-pd-controller/\naerospace/simulate-satellite/\naerospace/simulated-and-actual-flights/\naerospace/toroidal-propeller/\nansible/\nansible/inventory.yaml\nansible/Makefile\nansible/requirements.yml\nansible/hm_macos_group/\nansible/hm_ubuntu_group/\nansible/hm_windows_group/\napi-go/\napi-go/buf.yaml\napi-go/go.mod\napi-go/go.sum\napi-go/Makefile\napi-go/api/\napi-go/build/\napi-go/cmd/\napi-go/config/\napi-go/internal/\napi-node/\napi-node/.env.development\napi-node/.env.development.local.example\napi-node/.env.development.local.example.docker\napi-node/.env.production\napi-node/.env.production.local.example\napi-node/.env.test\napi-node/.eslintignore\napi-node/.eslintrc.js\napi-node/.npmrc\napi-node/.nvmrc\napi-node/babel.config.js\napi-node/docker-compose.cypress.yaml\napi-node/docker-compose.development.yaml\napi-node/Dockerfile\napi-node/Dockerfile.development\napi-node/jest.config.js\napi-node/Makefile\napi-node/package-lock.json\napi-node/package.json\napi-node/Procfile\napi-node/stryker.conf.js\napi-node/tsconfig.json\napi-node/bin/\napi-node/postgres/\napi-node/scripts/\napi-node/src/\napi-python/\napi-python/.flaskenv\napi-python/docker-entrypoint.sh\napi-python/Dockerfile\napi-python/Makefile\napi-python/poetry.lock\napi-python/poetry.toml\napi-python/pyproject.toml\napi-python/flaskr/\nasterios/\nasterios/led-blinker/\nauthorization/\nauthorization/hm-opal-client/\nauthorization/ory-hydra/\nautomobile/\nautomobile/build-map-by-lidar-point-cloud/\nautomobile/detect-lane-by-lidar-point-cloud/\nbin/\nbin/clean.sh\nbin/count_code_lines.sh\nbin/lint_javascript_fix.sh\nbin/lint_javascript.sh\nbin/set_up.sh\nbiology/\nbiology/compare-nucleotide-sequences/\nbusybox/\nbusybox/Makefile\ncaddy/\ncaddy/Caddyfile\ncaddy/Makefile\ncaddy/bin/\ncloud-computing/\ncloud-computing/hm-ray/\ncloud-computing/hm-skypilot/\ncloud-cost/\ncloud-cost/komiser/\ncloud-infrastructure/\ncloud-infrastructure/hm-pulumi/\ncloud-infrastructure/karpenter/\ncloud-infrastructure/terraform/\ncloud-platform/\ncloud-platform/aws/\ncloud-platform/google-cloud/\ncloud-security/\ncloud-security/hm-prowler/\ncomputational-fluid-dynamics/\ncomputational-fluid-dynamics/matlab/\ncomputational-fluid-dynamics/openfoam/\ncomputer-vision/\ncomputer-vision/hm-open3d/\ncomputer-vision/hm-pyvista/\ndata-analytics/\ndata-analytics/hm-geopandas/\ndata-distribution-service/\ndata-distribution-service/dummy_test.py\ndata-distribution-service/hm_message.idl\ndata-distribution-service/hm_message.xml\ndata-distribution-service/Makefile\ndata-distribution-service/poetry.lock\ndata-distribution-service/poetry.toml\ndata-distribution-service/publish.py\ndata-ingestion/\ndata-orchestration/\ndata-processing/\ndata-storage/\ndata-transformation/\ndata-visualization/\ndesktop-qt/\nembedded/\nethereum/\ngit/\ngolang-migrate/\nhardware-in-the-loop/\nhasura-graphql-engine/\nhigh-performance-computing/\nhm-alpine/\nhm-kafka/\nhm-locust/\nhm-rust/\nhm-traefik/\nhm-xxhash/\nkubernetes/\nmachine-learning/\nmatlab/\nmobile/\nnetwork-programmability/\noperating-system/\nparallel-computing/\nphysics/\nquantum-computing/\nrclone/\nrestic/\nreverse-engineering/\nrobotics/\nsubmodules/\ntrino/\nvagrant/\nvalgrind/\nvhdl/\nvim/\nweb/\nweb-cypress/\nwireless-network/\n\n(File list truncated. Use list_files on specific subdirectories if you need to explore further.)\n</environment_details>",
},
],
},
{
"role": "assistant",
"content": '<thinking>\nThe user is asking about a specific file: main.py. Based on the environment details provided, this file is located in the computer-vision/hm-open3d/src/ directory and is currently open in a VSCode tab.\n\nTo answer the question of what this file is, the most relevant tool would be the read_file tool. This will allow me to examine the contents of main.py to determine its purpose.\n\nThe read_file tool requires the "path" parameter. I can infer this path based on the environment details:\npath: "computer-vision/hm-open3d/src/main.py"\n\nSince I have the necessary parameter, I can proceed with calling the read_file tool.\n</thinking>',
"tool_calls": [
{
"id": "tooluse_qCt-KEyWQlWiyHl26spQVA",
"type": "function",
"function": {
"name": "read_file",
"arguments": '{"path":"computer-vision/hm-open3d/src/main.py"}',
},
}
],
},
{
"role": "tool",
"tool_call_id": "tooluse_qCt-KEyWQlWiyHl26spQVA",
"content": 'import numpy as np\nimport open3d as o3d\n\n\ndef main():\n ply_point_cloud = o3d.data.PLYPointCloud()\n pcd = o3d.io.read_point_cloud(ply_point_cloud.path)\n print(pcd)\n print(np.asarray(pcd.points))\n\n demo_crop_data = o3d.data.DemoCropPointCloud()\n vol = o3d.visualization.read_selection_polygon_volume(\n demo_crop_data.cropped_json_path\n )\n chair = vol.crop_point_cloud(pcd)\n\n dists = pcd.compute_point_cloud_distance(chair)\n dists = np.asarray(dists)\n idx = np.where(dists > 0.01)[0]\n pcd_without_chair = pcd.select_by_index(idx)\n\n axis_aligned_bounding_box = chair.get_axis_aligned_bounding_box()\n axis_aligned_bounding_box.color = (1, 0, 0)\n\n oriented_bounding_box = chair.get_oriented_bounding_box()\n oriented_bounding_box.color = (0, 1, 0)\n\n o3d.visualization.draw_geometries(\n [pcd_without_chair, chair, axis_aligned_bounding_box, oriented_bounding_box],\n zoom=0.3412,\n front=[0.4, -0.2, -0.9],\n lookat=[2.6, 2.0, 1.5],\n up=[-0.10, -1.0, 0.2],\n )\n\n\nif __name__ == "__main__":\n main()\n',
},
{
"role": "user",
"content": [
{
"type": "text",
"text": "<environment_details>\n# VSCode Visible Files\ncomputer-vision/hm-open3d/src/main.py\n\n# VSCode Open Tabs\ncomputer-vision/hm-open3d/src/main.py\n</environment_details>",
}
],
},
],
"temperature": 0.2,
"tools": [
{
"type": "function",
"function": {
"name": "execute_command",
"description": "Execute a CLI command on the system. Use this when you need to perform system operations or run specific commands to accomplish any step in the user's task. You must tailor your command to the user's system and provide a clear explanation of what the command does. Prefer to execute complex CLI commands over creating executable scripts, as they are more flexible and easier to run. Commands will be executed in the current working directory: /Users/hongbo-miao/Clouds/Git/hongbomiao.com",
"parameters": {
"type": "object",
"properties": {
"command": {
"type": "string",
"description": "The CLI command to execute. This should be valid for the current operating system. Ensure the command is properly formatted and does not contain any harmful instructions.",
}
},
"required": ["command"],
},
},
},
{
"type": "function",
"function": {
"name": "read_file",
"description": "Read the contents of a file at the specified path. Use this when you need to examine the contents of an existing file, for example to analyze code, review text files, or extract information from configuration files. Automatically extracts raw text from PDF and DOCX files. May not be suitable for other types of binary files, as it returns the raw content as a string.",
"parameters": {
"type": "object",
"properties": {
"path": {
"type": "string",
"description": "The path of the file to read (relative to the current working directory /Users/hongbo-miao/Clouds/Git/hongbomiao.com)",
}
},
"required": ["path"],
},
},
},
{
"type": "function",
"function": {
"name": "write_to_file",
"description": "Write content to a file at the specified path. If the file exists, it will be overwritten with the provided content. If the file doesn't exist, it will be created. Always provide the full intended content of the file, without any truncation. This tool will automatically create any directories needed to write the file.",
"parameters": {
"type": "object",
"properties": {
"path": {
"type": "string",
"description": "The path of the file to write to (relative to the current working directory /Users/hongbo-miao/Clouds/Git/hongbomiao.com)",
},
"content": {
"type": "string",
"description": "The full content to write to the file.",
},
},
"required": ["path", "content"],
},
},
},
{
"type": "function",
"function": {
"name": "search_files",
"description": "Perform a regex search across files in a specified directory, providing context-rich results. This tool searches for patterns or specific content across multiple files, displaying each match with encapsulating context.",
"parameters": {
"type": "object",
"properties": {
"path": {
"type": "string",
"description": "The path of the directory to search in (relative to the current working directory /Users/hongbo-miao/Clouds/Git/hongbomiao.com). This directory will be recursively searched.",
},
"regex": {
"type": "string",
"description": "The regular expression pattern to search for. Uses Rust regex syntax.",
},
"filePattern": {
"type": "string",
"description": "Optional glob pattern to filter files (e.g., '*.ts' for TypeScript files). If not provided, it will search all files (*).",
},
},
"required": ["path", "regex"],
},
},
},
{
"type": "function",
"function": {
"name": "list_files",
"description": "List files and directories within the specified directory. If recursive is true, it will list all files and directories recursively. If recursive is false or not provided, it will only list the top-level contents.",
"parameters": {
"type": "object",
"properties": {
"path": {
"type": "string",
"description": "The path of the directory to list contents for (relative to the current working directory /Users/hongbo-miao/Clouds/Git/hongbomiao.com)",
},
"recursive": {
"type": "string",
"enum": ["true", "false"],
"description": "Whether to list files recursively. Use 'true' for recursive listing, 'false' or omit for top-level only.",
},
},
"required": ["path"],
},
},
},
{
"type": "function",
"function": {
"name": "list_code_definition_names",
"description": "Lists definition names (classes, functions, methods, etc.) used in source code files at the top level of the specified directory. This tool provides insights into the codebase structure and important constructs, encapsulating high-level concepts and relationships that are crucial for understanding the overall architecture.",
"parameters": {
"type": "object",
"properties": {
"path": {
"type": "string",
"description": "The path of the directory (relative to the current working directory /Users/hongbo-miao/Clouds/Git/hongbomiao.com) to list top level source code definitions for",
}
},
"required": ["path"],
},
},
},
{
"type": "function",
"function": {
"name": "inspect_site",
"description": "Captures a screenshot and console logs of the initial state of a website. This tool navigates to the specified URL, takes a screenshot of the entire page as it appears immediately after loading, and collects any console logs or errors that occur during page load. It does not interact with the page or capture any state changes after the initial load.",
"parameters": {
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "The URL of the site to inspect. This should be a valid URL including the protocol (e.g. http://localhost:3000/page, file:///path/to/file.html, etc.)",
}
},
"required": ["url"],
},
},
},
{
"type": "function",
"function": {
"name": "ask_followup_question",
"description": "Ask the user a question to gather additional information needed to complete the task. This tool should be used when you encounter ambiguities, need clarification, or require more details to proceed effectively. It allows for interactive problem-solving by enabling direct communication with the user. Use this tool judiciously to maintain a balance between gathering necessary information and avoiding excessive back-and-forth.",
"parameters": {
"type": "object",
"properties": {
"question": {
"type": "string",
"description": "The question to ask the user. This should be a clear, specific question that addresses the information you need.",
}
},
"required": ["question"],
},
},
},
{
"type": "function",
"function": {
"name": "attempt_completion",
"description": "Once you've completed the task, use this tool to present the result to the user. Optionally you may provide a CLI command to showcase the result of your work, but avoid using commands like 'echo' or 'cat' that merely print text. They may respond with feedback if they are not satisfied with the result, which you can use to make improvements and try again.",
"parameters": {
"type": "object",
"properties": {
"command": {
"type": "string",
"description": "A CLI command to execute to show a live demo of the result to the user. For example, use 'open index.html' to display a created website. This command should be valid for the current operating system. Ensure the command is properly formatted and does not contain any harmful instructions.",
},
"result": {
"type": "string",
"description": "The result of the task. Formulate this result in a way that is final and does not require further input from the user. Don't end your result with questions or offers for further assistance.",
},
},
"required": ["result"],
},
},
},
],
"tool_choice": "auto",
}
if modify_params:
transformed_messages = _bedrock_converse_messages_pt(
messages=data["messages"], model="", llm_provider=""
)
expected_messages = [
{
"role": "user",
"content": [
{"text": "<task>\nWhat is this file?\n</task>"},
{
"text": "<environment_details>\n# VSCode Visible Files\ncomputer-vision/hm-open3d/src/main.py\n\n# VSCode Open Tabs\ncomputer-vision/hm-open3d/src/main.py\n\n# Current Working Directory (/Users/hongbo-miao/Clouds/Git/hongbomiao.com) Files\n.ansible-lint\n.clang-format\n.cmakelintrc\n.dockerignore\n.editorconfig\n.gitignore\n.gitmodules\n.hadolint.yaml\n.isort.cfg\n.markdownlint-cli2.jsonc\n.mergify.yml\n.npmrc\n.nvmrc\n.prettierignore\n.rubocop.yml\n.ruby-version\n.ruff.toml\n.shellcheckrc\n.solhint.json\n.solhintignore\n.sqlfluff\n.sqlfluffignore\n.stylelintignore\n.yamllint.yaml\nCODE_OF_CONDUCT.md\ncommitlint.config.js\nGemfile\nGemfile.lock\nLICENSE\nlint-staged.config.js\nMakefile\nmiss_hit.cfg\nmypy.ini\npackage-lock.json\npackage.json\npoetry.lock\npoetry.toml\nprettier.config.js\npyproject.toml\nREADME.md\nrelease.config.js\nrenovate.json\nSECURITY.md\nstylelint.config.js\naerospace/\naerospace/air-defense-system/\naerospace/hm-aerosandbox/\naerospace/hm-openaerostruct/\naerospace/px4/\naerospace/quadcopter-pd-controller/\naerospace/simulate-satellite/\naerospace/simulated-and-actual-flights/\naerospace/toroidal-propeller/\nansible/\nansible/inventory.yaml\nansible/Makefile\nansible/requirements.yml\nansible/hm_macos_group/\nansible/hm_ubuntu_group/\nansible/hm_windows_group/\napi-go/\napi-go/buf.yaml\napi-go/go.mod\napi-go/go.sum\napi-go/Makefile\napi-go/api/\napi-go/build/\napi-go/cmd/\napi-go/config/\napi-go/internal/\napi-node/\napi-node/.env.development\napi-node/.env.development.local.example\napi-node/.env.development.local.example.docker\napi-node/.env.production\napi-node/.env.production.local.example\napi-node/.env.test\napi-node/.eslintignore\napi-node/.eslintrc.js\napi-node/.npmrc\napi-node/.nvmrc\napi-node/babel.config.js\napi-node/docker-compose.cypress.yaml\napi-node/docker-compose.development.yaml\napi-node/Dockerfile\napi-node/Dockerfile.development\napi-node/jest.config.js\napi-node/Makefile\napi-node/package-lock.json\napi-node/package.json\napi-node/Procfile\napi-node/stryker.conf.js\napi-node/tsconfig.json\napi-node/bin/\napi-node/postgres/\napi-node/scripts/\napi-node/src/\napi-python/\napi-python/.flaskenv\napi-python/docker-entrypoint.sh\napi-python/Dockerfile\napi-python/Makefile\napi-python/poetry.lock\napi-python/poetry.toml\napi-python/pyproject.toml\napi-python/flaskr/\nasterios/\nasterios/led-blinker/\nauthorization/\nauthorization/hm-opal-client/\nauthorization/ory-hydra/\nautomobile/\nautomobile/build-map-by-lidar-point-cloud/\nautomobile/detect-lane-by-lidar-point-cloud/\nbin/\nbin/clean.sh\nbin/count_code_lines.sh\nbin/lint_javascript_fix.sh\nbin/lint_javascript.sh\nbin/set_up.sh\nbiology/\nbiology/compare-nucleotide-sequences/\nbusybox/\nbusybox/Makefile\ncaddy/\ncaddy/Caddyfile\ncaddy/Makefile\ncaddy/bin/\ncloud-computing/\ncloud-computing/hm-ray/\ncloud-computing/hm-skypilot/\ncloud-cost/\ncloud-cost/komiser/\ncloud-infrastructure/\ncloud-infrastructure/hm-pulumi/\ncloud-infrastructure/karpenter/\ncloud-infrastructure/terraform/\ncloud-platform/\ncloud-platform/aws/\ncloud-platform/google-cloud/\ncloud-security/\ncloud-security/hm-prowler/\ncomputational-fluid-dynamics/\ncomputational-fluid-dynamics/matlab/\ncomputational-fluid-dynamics/openfoam/\ncomputer-vision/\ncomputer-vision/hm-open3d/\ncomputer-vision/hm-pyvista/\ndata-analytics/\ndata-analytics/hm-geopandas/\ndata-distribution-service/\ndata-distribution-service/dummy_test.py\ndata-distribution-service/hm_message.idl\ndata-distribution-service/hm_message.xml\ndata-distribution-service/Makefile\ndata-distribution-service/poetry.lock\ndata-distribution-service/poetry.toml\ndata-distribution-service/publish.py\ndata-ingestion/\ndata-orchestration/\ndata-processing/\ndata-storage/\ndata-transformation/\ndata-visualization/\ndesktop-qt/\nembedded/\nethereum/\ngit/\ngolang-migrate/\nhardware-in-the-loop/\nhasura-graphql-engine/\nhigh-performance-computing/\nhm-alpine/\nhm-kafka/\nhm-locust/\nhm-rust/\nhm-traefik/\nhm-xxhash/\nkubernetes/\nmachine-learning/\nmatlab/\nmobile/\nnetwork-programmability/\noperating-system/\nparallel-computing/\nphysics/\nquantum-computing/\nrclone/\nrestic/\nreverse-engineering/\nrobotics/\nsubmodules/\ntrino/\nvagrant/\nvalgrind/\nvhdl/\nvim/\nweb/\nweb-cypress/\nwireless-network/\n\n(File list truncated. Use list_files on specific subdirectories if you need to explore further.)\n</environment_details>"
},
],
},
{
"role": "assistant",
"content": [
{
"toolUse": {
"input": {"path": "computer-vision/hm-open3d/src/main.py"},
"name": "read_file",
"toolUseId": "tooluse_qCt-KEyWQlWiyHl26spQVA",
}
}
],
},
{
"role": "user",
"content": [
{
"toolResult": {
"content": [
{
"text": 'import numpy as np\nimport open3d as o3d\n\n\ndef main():\n ply_point_cloud = o3d.data.PLYPointCloud()\n pcd = o3d.io.read_point_cloud(ply_point_cloud.path)\n print(pcd)\n print(np.asarray(pcd.points))\n\n demo_crop_data = o3d.data.DemoCropPointCloud()\n vol = o3d.visualization.read_selection_polygon_volume(\n demo_crop_data.cropped_json_path\n )\n chair = vol.crop_point_cloud(pcd)\n\n dists = pcd.compute_point_cloud_distance(chair)\n dists = np.asarray(dists)\n idx = np.where(dists > 0.01)[0]\n pcd_without_chair = pcd.select_by_index(idx)\n\n axis_aligned_bounding_box = chair.get_axis_aligned_bounding_box()\n axis_aligned_bounding_box.color = (1, 0, 0)\n\n oriented_bounding_box = chair.get_oriented_bounding_box()\n oriented_bounding_box.color = (0, 1, 0)\n\n o3d.visualization.draw_geometries(\n [pcd_without_chair, chair, axis_aligned_bounding_box, oriented_bounding_box],\n zoom=0.3412,\n front=[0.4, -0.2, -0.9],\n lookat=[2.6, 2.0, 1.5],\n up=[-0.10, -1.0, 0.2],\n )\n\n\nif __name__ == "__main__":\n main()\n'
}
],
"toolUseId": "tooluse_qCt-KEyWQlWiyHl26spQVA",
}
}
],
},
{"role": "assistant", "content": [{"text": "Please continue."}]},
{
"role": "user",
"content": [
{
"text": "<environment_details>\n# VSCode Visible Files\ncomputer-vision/hm-open3d/src/main.py\n\n# VSCode Open Tabs\ncomputer-vision/hm-open3d/src/main.py\n</environment_details>"
}
],
},
]
assert transformed_messages == expected_messages
else:
with pytest.raises(Exception) as e:
litellm.completion(**data)
assert "litellm.modify_params" in str(e.value)

4
tests/local_testing/test_completion_cost.py
View file

@ -1077,7 +1077,9 @@ def test_completion_cost_deepseek():
assert response_2.usage.prompt_cache_hit_tokens is not None assert response_2.usage.prompt_cache_hit_tokens is not None
assert response_2.usage.prompt_cache_miss_tokens is not None assert response_2.usage.prompt_cache_miss_tokens is not None
assert ( assert (
response_2.usage.prompt_tokens == response_2.usage.prompt_cache_miss_tokens response_2.usage.prompt_tokens
== response_2.usage.prompt_cache_miss_tokens
+ response_2.usage.prompt_cache_hit_tokens
) )
assert ( assert (
response_2.usage._cache_read_input_tokens response_2.usage._cache_read_input_tokens

17
tests/local_testing/test_custom_callback_input.py
View file

@ -1258,6 +1258,7 @@ def test_standard_logging_payload(model, turn_off_message_logging):
"standard_logging_object" "standard_logging_object"
] ]
if turn_off_message_logging: if turn_off_message_logging:
print("checks redacted-by-litellm")
assert "redacted-by-litellm" == slobject["messages"][0]["content"] assert "redacted-by-litellm" == slobject["messages"][0]["content"]
assert "redacted-by-litellm" == slobject["response"] assert "redacted-by-litellm" == slobject["response"]
@ -1307,9 +1308,15 @@ def test_aaastandard_logging_payload_cache_hit():
assert standard_logging_object["saved_cache_cost"] > 0 assert standard_logging_object["saved_cache_cost"] > 0
def test_logging_async_cache_hit_sync_call(): @pytest.mark.parametrize(
"turn_off_message_logging",
[False, True],
) # False
def test_logging_async_cache_hit_sync_call(turn_off_message_logging):
from litellm.types.utils import StandardLoggingPayload from litellm.types.utils import StandardLoggingPayload
litellm.turn_off_message_logging = turn_off_message_logging
litellm.cache = Cache() litellm.cache = Cache()
response = litellm.completion( response = litellm.completion(
@ -1356,6 +1363,14 @@ def test_logging_async_cache_hit_sync_call():
assert standard_logging_object["response_cost"] == 0 assert standard_logging_object["response_cost"] == 0
assert standard_logging_object["saved_cache_cost"] > 0 assert standard_logging_object["saved_cache_cost"] > 0
if turn_off_message_logging:
print("checks redacted-by-litellm")
assert (
"redacted-by-litellm"
== standard_logging_object["messages"][0]["content"]
)
assert "redacted-by-litellm" == standard_logging_object["response"]
def test_logging_standard_payload_failure_call(): def test_logging_standard_payload_failure_call():
from litellm.types.utils import StandardLoggingPayload from litellm.types.utils import StandardLoggingPayload

81
tests/local_testing/test_prompt_caching.py Normal file
View file

@ -0,0 +1,81 @@
"""Asserts that prompt caching information is correctly returned for Anthropic, OpenAI, and Deepseek"""
import io
import os
import sys
sys.path.insert(0, os.path.abspath("../.."))
import litellm
import pytest
@pytest.mark.parametrize(
"model",
[
"anthropic/claude-3-5-sonnet-20240620",
"openai/gpt-4o",
"deepseek/deepseek-chat",
],
)
def test_prompt_caching_model(model):
for _ in range(2):
response = litellm.completion(
model=model,
messages=[
# System Message
{
"role": "system",
"content": [
{
"type": "text",
"text": "Here is the full text of a complex legal agreement"
* 400,
"cache_control": {"type": "ephemeral"},
}
],
},
# marked for caching with the cache_control parameter, so that this checkpoint can read from the previous cache.
{
"role": "user",
"content": [
{
"type": "text",
"text": "What are the key terms and conditions in this agreement?",
"cache_control": {"type": "ephemeral"},
}
],
},
{
"role": "assistant",
"content": "Certainly! the key terms and conditions are the following: the contract is 1 year long for $10/mo",
},
# The final turn is marked with cache-control, for continuing in followups.
{
"role": "user",
"content": [
{
"type": "text",
"text": "What are the key terms and conditions in this agreement?",
"cache_control": {"type": "ephemeral"},
}
],
},
],
temperature=0.2,
max_tokens=10,
)
print("response=", response)
print("response.usage=", response.usage)
assert "prompt_tokens_details" in response.usage
assert response.usage.prompt_tokens_details.cached_tokens > 0
# assert "cache_read_input_tokens" in response.usage
# assert "cache_creation_input_tokens" in response.usage
# # Assert either a cache entry was created or cache was read - changes depending on the anthropic api ttl
# assert (response.usage.cache_read_input_tokens > 0) or (
# response.usage.cache_creation_input_tokens > 0
# )

34
tests/local_testing/test_sagemaker.py
View file

@ -180,6 +180,40 @@ async def test_completion_sagemaker_stream(sync_mode, model):
pytest.fail(f"Error occurred: {e}") pytest.fail(f"Error occurred: {e}")
@pytest.mark.asyncio()
@pytest.mark.parametrize("sync_mode", [False, True])
@pytest.mark.parametrize(
"model",
[
"sagemaker_chat/huggingface-pytorch-tgi-inference-2024-08-23-15-48-59-245",
"sagemaker/jumpstart-dft-hf-textgeneration1-mp-20240815-185614",
],
)
async def test_completion_sagemaker_streaming_bad_request(sync_mode, model):
litellm.set_verbose = True
print("testing sagemaker")
if sync_mode is True:
with pytest.raises(litellm.BadRequestError):
response = litellm.completion(
model=model,
messages=[
{"role": "user", "content": "hi"},
],
stream=True,
max_tokens=8000000000000000,
)
else:
with pytest.raises(litellm.BadRequestError):
response = await litellm.acompletion(
model=model,
messages=[
{"role": "user", "content": "hi"},
],
stream=True,
max_tokens=8000000000000000,
)
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_acompletion_sagemaker_non_stream(): async def test_acompletion_sagemaker_non_stream():
mock_response = AsyncMock() mock_response = AsyncMock()

20
tests/local_testing/test_utils.py
View file

@ -517,17 +517,19 @@ def test_redact_msgs_from_logs():
] ]
) )
litellm_logging_obj = Logging(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hi"}],
stream=False,
call_type="acompletion",
litellm_call_id="1234",
start_time=datetime.now(),
function_id="1234",
)
_redacted_response_obj = redact_message_input_output_from_logging( _redacted_response_obj = redact_message_input_output_from_logging(
result=response_obj, result=response_obj,
litellm_logging_obj=Logging( model_call_details=litellm_logging_obj.model_call_details,
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hi"}],
stream=False,
call_type="acompletion",
litellm_call_id="1234",
start_time=datetime.now(),
function_id="1234",
),
) )
# Assert the response_obj content is NOT modified # Assert the response_obj content is NOT modified