tiennm99/litellm
1
0
Fork 0
mirror of https://github.com/tiennm99/litellm.git synced 2026-07-16 00:23:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #24173 from joereyna/fix/black-format-batch-3

Browse Source 
chore: apply black formatting to fix lint CI (batch 3)
This commit is contained in:
Krish Dholakia
2026-03-19 18:27:30 -07:00
committed by GitHub
parent 482b77e71d 87b5039aaa
commit d8b28141b9
10 changed files with 121 additions and 99 deletions
Download Patch File Download Diff File Expand all files Collapse all files
litellm/integrations/langsmith.py
+14 -4
View File
@@ -114,7 +114,9 @@ class LangsmithLogger(CustomBatchLogger):
self, metadata: dict, credentials: LangsmithCredentialsObject
):
return {
"project_name": metadata.get("project_name", credentials["LANGSMITH_PROJECT"]),
"project_name": metadata.get(
"project_name", credentials["LANGSMITH_PROJECT"]
),
"run_name": metadata.get("run_name", self.langsmith_default_run_name),
"run_id": metadata.get("id", metadata.get("run_id", None)),
"parent_run_id": metadata.get("parent_run_id", None),
@@ -132,7 +134,9 @@ class LangsmithLogger(CustomBatchLogger):
extra_metadata[key] = requester_metadata[key]
return extra_metadata
def _build_outputs_with_usage(self, payload: StandardLoggingPayload) -> Dict[str, Any]:
def _build_outputs_with_usage(
self, payload: StandardLoggingPayload
) -> Dict[str, Any]:
response = payload["response"]
outputs: Dict[str, Any]
if isinstance(response, dict):
@@ -171,7 +175,7 @@ class LangsmithLogger(CustomBatchLogger):
try:
_litellm_params = kwargs.get("litellm_params", {}) or {}
metadata = _litellm_params.get("metadata", {}) or {}
fields = self._extract_metadata_fields(metadata, credentials)
verbose_logger.debug(
f"Langsmith Logging - project_name: {fields['project_name']}, run_name {fields['run_name']}"
@@ -202,7 +206,13 @@ class LangsmithLogger(CustomBatchLogger):
if payload["error_str"] is not None and payload["status"] == "failure":
data["error"] = payload["error_str"]
for key in ("id", "parent_run_id", "trace_id", "session_id", "dotted_order"):
for key in (
"id",
"parent_run_id",
"trace_id",
"session_id",
"dotted_order",
):
field_key = "run_id" if key == "id" else key
if fields[field_key]:
data[key] = fields[field_key]
litellm/litellm_core_utils/litellm_logging.py
+4 -1
View File
@@ -3331,7 +3331,10 @@ class Logging(LiteLLMLoggingBaseClass):
return result
elif isinstance(result, TextCompletionResponse):
return result
elif isinstance(result, (ResponseCompletedEvent, ResponseIncompleteEvent, ResponseFailedEvent)):
elif isinstance(
result,
(ResponseCompletedEvent, ResponseIncompleteEvent, ResponseFailedEvent),
):
## return unified Usage object
if isinstance(result.response.usage, ResponseAPIUsage):
transformed_usage = (
litellm/litellm_core_utils/streaming_handler.py
+1 -1
View File
@@ -279,7 +279,7 @@ class CustomStreamWrapper:
model="",
llm_provider="",
)
def check_special_tokens(self, chunk: str, finish_reason: Optional[str]):
"""
Output parse <s> / </s> special tokens for sagemaker + hf streaming.
litellm/llms/anthropic/chat/handler.py
+21 -19
View File
@@ -578,7 +578,9 @@ class ModelResponseIterator:
speed=self.speed,
)
def _content_block_delta_helper(self, chunk: dict) -> Tuple[
def _content_block_delta_helper(
self, chunk: dict
) -> Tuple[
str,
Optional[ChatCompletionToolCallChunk],
List[Union[ChatCompletionThinkingBlock, ChatCompletionRedactedThinkingBlock]],
@@ -803,9 +805,9 @@ class ModelResponseIterator:
tool_input = content_block_start["content_block"].get(
"input", {}
)
self._server_tool_inputs[self._current_server_tool_id] = (
tool_input
)
self._server_tool_inputs[
self._current_server_tool_id
] = tool_input
# Include caller information if present (for programmatic tool calling)
if "caller" in content_block_start["content_block"]:
caller_data = content_block_start["content_block"]["caller"]
@@ -826,9 +828,9 @@ class ModelResponseIterator:
# Handle compaction blocks
# The full content comes in content_block_start
self.compaction_blocks.append(content_block_start["content_block"])
provider_specific_fields["compaction_blocks"] = (
self.compaction_blocks
)
provider_specific_fields[
"compaction_blocks"
] = self.compaction_blocks
provider_specific_fields["compaction_start"] = {
"type": "compaction",
"content": content_block_start["content_block"].get(
@@ -850,9 +852,9 @@ class ModelResponseIterator:
self.web_search_results.append(
content_block_start["content_block"]
)
provider_specific_fields["web_search_results"] = (
self.web_search_results
)
provider_specific_fields[
"web_search_results"
] = self.web_search_results
elif content_type == "web_fetch_tool_result":
# Capture web_fetch_tool_result for multi-turn reconstruction
# The full content comes in content_block_start, not in deltas
@@ -860,18 +862,18 @@ class ModelResponseIterator:
self.web_search_results.append(
content_block_start["content_block"]
)
provider_specific_fields["web_search_results"] = (
self.web_search_results
)
provider_specific_fields[
"web_search_results"
] = self.web_search_results
elif content_type != "tool_search_tool_result":
# Handle other tool results (code execution, etc.)
# Skip tool_search_tool_result as it's internal metadata
self.tool_results.append(content_block_start["content_block"])
provider_specific_fields["tool_results"] = self.tool_results
# Convert to provider-neutral code_interpreter_results
provider_specific_fields["code_interpreter_results"] = (
self._build_code_interpreter_results()
)
provider_specific_fields[
"code_interpreter_results"
] = self._build_code_interpreter_results()
elif type_chunk == "content_block_stop":
ContentBlockStop(**chunk) # type: ignore
@@ -928,9 +930,9 @@ class ModelResponseIterator:
)
if container_id and self.tool_results:
self._container_id = container_id
provider_specific_fields["code_interpreter_results"] = (
self._build_code_interpreter_results()
)
provider_specific_fields[
"code_interpreter_results"
] = self._build_code_interpreter_results()
elif type_chunk == "message_start":
"""
Anthropic
litellm/llms/anthropic/chat/transformation.py
+41 -26
View File
@@ -964,11 +964,11 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
if mcp_servers:
optional_params["mcp_servers"] = mcp_servers
elif param == "tool_choice" or param == "parallel_tool_calls":
_tool_choice: Optional[AnthropicMessagesToolChoice] = (
self._map_tool_choice(
tool_choice=non_default_params.get("tool_choice"),
parallel_tool_use=non_default_params.get("parallel_tool_calls"),
)
_tool_choice: Optional[
AnthropicMessagesToolChoice
] = self._map_tool_choice(
tool_choice=non_default_params.get("tool_choice"),
parallel_tool_use=non_default_params.get("parallel_tool_calls"),
)
if _tool_choice is not None:
@@ -1066,9 +1066,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
self.map_openai_context_management_to_anthropic(value)
)
if anthropic_context_management is not None:
optional_params["context_management"] = (
anthropic_context_management
)
optional_params[
"context_management"
] = anthropic_context_management
elif param == "speed" and isinstance(value, str):
# Pass through Anthropic-specific speed parameter for fast mode
optional_params["speed"] = value
@@ -1142,9 +1142,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
text=system_message_block["content"],
)
if "cache_control" in system_message_block:
anthropic_system_message_content["cache_control"] = (
system_message_block["cache_control"]
)
anthropic_system_message_content[
"cache_control"
] = system_message_block["cache_control"]
anthropic_system_message_list.append(
anthropic_system_message_content
)
@@ -1168,9 +1168,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
)
)
if "cache_control" in _content:
anthropic_system_message_content["cache_control"] = (
_content["cache_control"]
)
anthropic_system_message_content[
"cache_control"
] = _content["cache_control"]
anthropic_system_message_list.append(
anthropic_system_message_content
@@ -1467,7 +1467,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
)
return _message
def extract_response_content(self, completion_response: dict) -> Tuple[
def extract_response_content(
self, completion_response: dict
) -> Tuple[
str,
Optional[List[Any]],
Optional[
@@ -1684,7 +1686,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
)
return usage
def _build_code_by_id_map(self, tool_calls: List[ChatCompletionToolCallChunk]) -> Dict[str, str]:
def _build_code_by_id_map(
self, tool_calls: List[ChatCompletionToolCallChunk]
) -> Dict[str, str]:
code_by_id: Dict[str, str] = {}
for tc in tool_calls:
try:
@@ -1698,7 +1702,10 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
return code_by_id
def _build_code_interpreter_results(
self, tool_results: List[Any], code_by_id: Dict[str, str], container_id: Optional[str]
self,
tool_results: List[Any],
code_by_id: Dict[str, str],
container_id: Optional[str],
) -> List[OutputCodeInterpreterCall]:
code_interpreter_results = []
for tr in tool_results:
@@ -1723,7 +1730,11 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
self,
completion_response: dict,
citations: Optional[List[Any]],
thinking_blocks: Optional[List[Union[ChatCompletionThinkingBlock, ChatCompletionRedactedThinkingBlock]]],
thinking_blocks: Optional[
List[
Union[ChatCompletionThinkingBlock, ChatCompletionRedactedThinkingBlock]
]
],
web_search_results: Optional[List[Any]],
tool_results: Optional[List[Any]],
compaction_blocks: Optional[List[Any]],
@@ -1733,14 +1744,14 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
"citations": citations,
"thinking_blocks": thinking_blocks,
}
context_management = completion_response.get("context_management")
if context_management is not None:
provider_specific_fields["context_management"] = context_management
if web_search_results is not None:
provider_specific_fields["web_search_results"] = web_search_results
if tool_results is not None:
provider_specific_fields["tool_results"] = tool_results
container_id = (
@@ -1752,15 +1763,17 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
code_interpreter_results = self._build_code_interpreter_results(
tool_results, code_by_id, container_id
)
provider_specific_fields["code_interpreter_results"] = code_interpreter_results
provider_specific_fields[
"code_interpreter_results"
] = code_interpreter_results
container = completion_response.get("container")
if container is not None:
provider_specific_fields["container"] = container
if compaction_blocks is not None:
provider_specific_fields["compaction_blocks"] = compaction_blocks
return provider_specific_fields
def transform_parsed_response(
@@ -1830,7 +1843,9 @@ class AnthropicConfig(AnthropicModelInfo, BaseConfig):
_message = json_mode_message
model_response.choices[0].message = _message
model_response._hidden_params["original_response"] = completion_response["content"]
model_response._hidden_params["original_response"] = completion_response[
"content"
]
model_response.choices[0].finish_reason = cast(
OpenAIChatCompletionFinishReason,
map_finish_reason(completion_response["stop_reason"]),
litellm/proxy/pass_through_endpoints/llm_provider_handlers/anthropic_passthrough_logging_handler.py
+12 -16
View File
@@ -6,23 +6,21 @@ import httpx
import litellm
from litellm._logging import verbose_proxy_logger
from litellm.litellm_core_utils.litellm_logging import \
Logging as LiteLLMLoggingObj
from litellm.litellm_core_utils.litellm_logging import \
use_custom_pricing_for_model
from litellm.litellm_core_utils.litellm_logging import Logging as LiteLLMLoggingObj
from litellm.litellm_core_utils.litellm_logging import use_custom_pricing_for_model
from litellm.llms.anthropic import get_anthropic_config
from litellm.llms.anthropic.chat.handler import \
ModelResponseIterator as AnthropicModelResponseIterator
from litellm.llms.anthropic.chat.handler import (
ModelResponseIterator as AnthropicModelResponseIterator,
)
from litellm.proxy._types import PassThroughEndpointLoggingTypedDict
from litellm.proxy.auth.auth_utils import get_end_user_id_from_request_body
from litellm.types.passthrough_endpoints.pass_through_endpoints import \
PassthroughStandardLoggingPayload
from litellm.types.utils import (LiteLLMBatch, ModelResponse,
TextCompletionResponse)
from litellm.types.passthrough_endpoints.pass_through_endpoints import (
PassthroughStandardLoggingPayload,
)
from litellm.types.utils import LiteLLMBatch, ModelResponse, TextCompletionResponse
if TYPE_CHECKING:
from litellm.types.passthrough_endpoints.pass_through_endpoints import \
EndpointType
from litellm.types.passthrough_endpoints.pass_through_endpoints import EndpointType
from ..success_handler import PassThroughEndpointLogging
else:
@@ -333,8 +331,7 @@ class AnthropicPassthroughLoggingHandler:
import base64
from litellm._uuid import uuid
from litellm.llms.anthropic.batches.transformation import \
AnthropicBatchesConfig
from litellm.llms.anthropic.batches.transformation import AnthropicBatchesConfig
from litellm.types.utils import Choices, SpecialEnums
try:
@@ -550,8 +547,7 @@ class AnthropicPassthroughLoggingHandler:
managed_files_hook, "store_unified_object_id"
):
# Create a mock user API key dict for the managed object storage
from litellm.proxy._types import (LitellmUserRoles,
UserAPIKeyAuth)
from litellm.proxy._types import LitellmUserRoles, UserAPIKeyAuth
user_api_key_dict = UserAPIKeyAuth(
user_id=kwargs.get("user_id", "default-user"),
litellm/proxy/utils.py
+3 -3
View File
@@ -1898,9 +1898,9 @@ class ProxyLogging:
normalized_call_type = CallTypes.aembedding.value
if normalized_call_type is not None:
litellm_logging_obj.call_type = normalized_call_type
litellm_logging_obj.model_call_details["call_type"] = (
normalized_call_type
)
litellm_logging_obj.model_call_details[
"call_type"
] = normalized_call_type
# Pass-through endpoints are logged via the callback loop's
# async_post_call_failure_hook — skip pre_call and failure handlers.
if litellm_logging_obj.call_type == CallTypes.pass_through.value:
litellm/responses/litellm_completion_transformation/transformation.py
+3 -3
View File
@@ -2113,9 +2113,9 @@ class LiteLLMCompletionResponsesConfig:
hasattr(completion_details, "reasoning_tokens")
and completion_details.reasoning_tokens is not None
):
output_details_dict["reasoning_tokens"] = (
completion_details.reasoning_tokens
)
output_details_dict[
"reasoning_tokens"
] = completion_details.reasoning_tokens
else:
output_details_dict["reasoning_tokens"] = 0
litellm/responses/streaming_iterator.py
+4 -8
View File
@@ -168,14 +168,10 @@ class BaseResponsesAPIStreamingIterator:
# Store the completed response (also for incomplete/failed so logging still fires)
_chunk_type = getattr(openai_responses_api_chunk, "type", None)
if (
openai_responses_api_chunk
and _chunk_type
in (
ResponsesAPIStreamEvents.RESPONSE_COMPLETED,
ResponsesAPIStreamEvents.RESPONSE_INCOMPLETE,
ResponsesAPIStreamEvents.RESPONSE_FAILED,
)
if openai_responses_api_chunk and _chunk_type in (
ResponsesAPIStreamEvents.RESPONSE_COMPLETED,
ResponsesAPIStreamEvents.RESPONSE_INCOMPLETE,
ResponsesAPIStreamEvents.RESPONSE_FAILED,
):
self.completed_response = openai_responses_api_chunk
# Add cost to usage object if include_cost_in_streaming_usage is True
litellm/types/llms/openai.py
+18 -18
View File
@@ -970,12 +970,12 @@ class OpenAIChatCompletionChunk(ChatCompletionChunk):
class Hyperparameters(BaseModel):
batch_size: Optional[Union[str, int]] = None # "Number of examples in each batch."
learning_rate_multiplier: Optional[Union[str, float]] = (
None # Scaling factor for the learning rate
)
n_epochs: Optional[Union[str, int]] = (
None # "The number of epochs to train the model for"
)
learning_rate_multiplier: Optional[
Union[str, float]
] = None # Scaling factor for the learning rate
n_epochs: Optional[
Union[str, int]
] = None # "The number of epochs to train the model for"
model_config = {"extra": "allow"}
@@ -1004,18 +1004,18 @@ class FineTuningJobCreate(BaseModel):
model: str # "The name of the model to fine-tune."
training_file: str # "The ID of an uploaded file that contains training data."
hyperparameters: Optional[Hyperparameters] = (
None # "The hyperparameters used for the fine-tuning job."
)
suffix: Optional[str] = (
None # "A string of up to 18 characters that will be added to your fine-tuned model name."
)
validation_file: Optional[str] = (
None # "The ID of an uploaded file that contains validation data."
)
integrations: Optional[List[str]] = (
None # "A list of integrations to enable for your fine-tuning job."
)
hyperparameters: Optional[
Hyperparameters
] = None # "The hyperparameters used for the fine-tuning job."
suffix: Optional[
str
] = None # "A string of up to 18 characters that will be added to your fine-tuned model name."
validation_file: Optional[
str
] = None # "The ID of an uploaded file that contains validation data."
integrations: Optional[
List[str]
] = None # "A list of integrations to enable for your fine-tuning job."
seed: Optional[int] = None # "The seed controls the reproducibility of the job."