build: merge squashed commit

Squashed commit of the following:

commit 6678e15381
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 09:29:15 2025 -0800

    test_prompt_caching

commit bd86e0ac47
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 08:57:16 2025 -0800

    test_prompt_caching

commit 2fc21ad51e
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 08:13:45 2025 -0800

    test_aprompt_caching

commit d94cff55ff
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 08:13:12 2025 -0800

    test_prompt_caching

commit 49c5e7811e
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 07:43:53 2025 -0800

    ui new build

commit cb8d5e5917
Author: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Date:   Wed Feb 26 07:38:56 2025 -0800

    (UI) - Create Key flow for existing users (#8844)

    * working create user button

    * working create user for a key flow

    * allow searching users

    * working create user + key

    * use clear sections on create key

    * better search for users

    * fix create key

    * ui fix create key button - make it neater / cleaner

    * ui fix all keys table

commit 335ba30467
Author: Krrish Dholakia <krrishdholakia@gmail.com>
Date:   Wed Feb 26 08:53:17 2025 -0800

    fix: fix file name

commit b8c5b31a4e
Author: Krrish Dholakia <krrishdholakia@gmail.com>
Date:   Tue Feb 25 22:54:46 2025 -0800

    fix: fix utils

commit ac6e503461
Author: Krrish Dholakia <krrishdholakia@gmail.com>
Date:   Mon Feb 24 10:43:31 2025 -0800

    fix(main.py): fix openai message for assistant msg if role is missing - openai allows this

    Fixes https://github.com/BerriAI/litellm/issues/8661

commit de3989dbc5
Author: Krrish Dholakia <krrishdholakia@gmail.com>
Date:   Mon Feb 24 21:19:25 2025 -0800

    fix(get_litellm_params.py): handle no-log being passed in via kwargs

    Fixes https://github.com/BerriAI/litellm/issues/8380

litellm/litellm_core_utils/get_litellm_params.py

@@ -75,7 +75,7 @@ def get_litellm_params(
         "model_info": model_info,
         "proxy_server_request": proxy_server_request,
         "preset_cache_key": preset_cache_key,
-        "no-log": no_log,
+        "no-log": no_log or kwargs.get("no-log"),
         "stream_response": {},  # litellm_call_id: ModelResponse Dict
         "input_cost_per_token": input_cost_per_token,
         "input_cost_per_second": input_cost_per_second,

litellm/litellm_core_utils/litellm_logging.py

@@ -3,7 +3,6 @@
 # Logging function -> log the exact model details + what's being sent | Non-Blocking
 import copy
 import datetime
-from functools import lru_cache
 import json
 import os
 import re
@@ -13,6 +12,7 @@ import time
 import traceback
 import uuid
 from datetime import datetime as dt_object
+from functools import lru_cache
 from typing import Any, Callable, Dict, List, Literal, Optional, Tuple, Union, cast
 
 from pydantic import BaseModel
@@ -2513,7 +2513,9 @@ def _init_custom_logger_compatible_class(  # noqa: PLR0915
 
             # auth can be disabled on local deployments of arize phoenix
             if arize_phoenix_config.otlp_auth_headers is not None:
-                os.environ["OTEL_EXPORTER_OTLP_TRACES_HEADERS"] = arize_phoenix_config.otlp_auth_headers
+                os.environ["OTEL_EXPORTER_OTLP_TRACES_HEADERS"] = (
+                    arize_phoenix_config.otlp_auth_headers
+                )
 
             for callback in _in_memory_loggers:
                 if (
@@ -2521,7 +2523,9 @@ def _init_custom_logger_compatible_class(  # noqa: PLR0915
                     and callback.callback_name == "arize_phoenix"
                 ):
                     return callback  # type: ignore
-            _otel_logger = OpenTelemetry(config=otel_config, callback_name="arize_phoenix")
+            _otel_logger = OpenTelemetry(
+                config=otel_config, callback_name="arize_phoenix"
+            )
             _in_memory_loggers.append(_otel_logger)
             return _otel_logger  # type: ignore
         elif logging_integration == "otel":

litellm/main.py

@@ -94,7 +94,7 @@ from litellm.utils import (
     read_config_args,
     supports_httpx_timeout,
     token_counter,
-    validate_chat_completion_messages,
+    validate_and_fix_openai_messages,
     validate_chat_completion_tool_choice,
 )
 
@@ -851,7 +851,7 @@ def completion(  # type: ignore # noqa: PLR0915
     if model is None:
         raise ValueError("model param not passed in.")
     # validate messages
-    messages = validate_chat_completion_messages(messages=messages)
+    messages = validate_and_fix_openai_messages(messages=messages)
     # validate tool_choice
     tool_choice = validate_chat_completion_tool_choice(tool_choice=tool_choice)
     ######### unpacking kwargs #####################

litellm/proxy/_new_secret_config.yaml

@@ -1,5 +1,24 @@
 model_list:
   - model_name: claude-3.5
     litellm_params:
+<<<<<<< HEAD
       model: claude-3-5-sonnet-latest
       api_key: os.environ/ANTHROPIC_API_KEY
+=======
+      model: openai/gpt-3.5-turbo
+      api_key: os.environ/OPENAI_API_KEY
+      api_base: http://0.0.0.0:8090
+  - model_name: deepseek-r1
+    litellm_params:
+      model: bedrock/deepseek_r1/arn:aws:bedrock:us-west-2:888602223428:imported-model/bnnr6463ejgf
+  - model_name: deepseek-r1-api
+    litellm_params:
+      model: deepseek/deepseek-reasoner
+  - model_name: cohere.embed-english-v3
+    litellm_params:
+      model: bedrock/cohere.embed-english-v3
+      api_key: os.environ/COHERE_API_KEY
+
+litellm_settings:
+  callbacks: ["langfuse"]
+>>>>>>> f86a609ea (fix(get_litellm_params.py): handle no-log being passed in via kwargs)

litellm/utils.py

@@ -5932,6 +5932,18 @@ def convert_to_dict(message: Union[BaseModel, dict]) -> dict:
         )
 
 
+def validate_and_fix_openai_messages(messages: List):
+    """
+    Ensures all messages are valid OpenAI chat completion messages.
+
+    Handles missing role for assistant messages.
+    """
+    for message in messages:
+        if not message.get("role"):
+            message["role"] = "assistant"
+    return validate_chat_completion_messages(messages=messages)
+
+
 def validate_chat_completion_messages(messages: List[AllMessageValues]):
     """
     Ensures all messages are valid OpenAI chat completion messages.
@@ -6282,11 +6294,18 @@ def get_end_user_id_for_cost_tracking(
         return None
     return end_user_id
 
-def should_use_cohere_v1_client(api_base: Optional[str], present_version_params: List[str]):
+
+def should_use_cohere_v1_client(
+    api_base: Optional[str], present_version_params: List[str]
+):
     if not api_base:
         return False
-    uses_v1_params = ("max_chunks_per_doc" in present_version_params) and ('max_tokens_per_doc' not in present_version_params) 
+    uses_v1_params = ("max_chunks_per_doc" in present_version_params) and (
-    return api_base.endswith("/v1/rerank") or (uses_v1_params and not api_base.endswith("/v2/rerank"))
+        "max_tokens_per_doc" not in present_version_params
+    )
+    return api_base.endswith("/v1/rerank") or (
+        uses_v1_params and not api_base.endswith("/v2/rerank")
+    )
 
 
 def is_prompt_caching_valid_prompt(

tests/litellm/test_main.py

@@ -0,0 +1,119 @@
+import json
+import os
+import sys
+
+import pytest
+from fastapi.testclient import TestClient
+
+sys.path.insert(
+    0, os.path.abspath("../..")
+)  # Adds the parent directory to the system path
+
+from unittest.mock import MagicMock, patch
+
+import litellm
+
+
+@pytest.fixture
+def openai_api_response():
+    mock_response_data = {
+        "id": "chatcmpl-B0W3vmiM78Xkgx7kI7dr7PC949DMS",
+        "choices": [
+            {
+                "finish_reason": "stop",
+                "index": 0,
+                "logprobs": None,
+                "message": {
+                    "content": "",
+                    "refusal": None,
+                    "role": "assistant",
+                    "audio": None,
+                    "function_call": None,
+                    "tool_calls": None,
+                },
+            }
+        ],
+        "created": 1739462947,
+        "model": "gpt-4o-mini-2024-07-18",
+        "object": "chat.completion",
+        "service_tier": "default",
+        "system_fingerprint": "fp_bd83329f63",
+        "usage": {
+            "completion_tokens": 1,
+            "prompt_tokens": 121,
+            "total_tokens": 122,
+            "completion_tokens_details": {
+                "accepted_prediction_tokens": 0,
+                "audio_tokens": 0,
+                "reasoning_tokens": 0,
+                "rejected_prediction_tokens": 0,
+            },
+            "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 0},
+        },
+    }
+
+    return mock_response_data
+
+
+def test_completion_missing_role(openai_api_response):
+    from openai import OpenAI
+
+    from litellm.types.utils import ModelResponse
+
+    client = OpenAI(api_key="test_api_key")
+
+    mock_raw_response = MagicMock()
+    mock_raw_response.headers = {
+        "x-request-id": "123",
+        "openai-organization": "org-123",
+        "x-ratelimit-limit-requests": "100",
+        "x-ratelimit-remaining-requests": "99",
+    }
+    mock_raw_response.parse.return_value = ModelResponse(**openai_api_response)
+
+    print(f"openai_api_response: {openai_api_response}")
+
+    with patch.object(
+        client.chat.completions.with_raw_response, "create", mock_raw_response
+    ) as mock_create:
+        litellm.completion(
+            model="gpt-4o-mini",
+            messages=[
+                {"role": "user", "content": "Hey"},
+                {
+                    "content": "",
+                    "tool_calls": [
+                        {
+                            "id": "call_m0vFJjQmTH1McvaHBPR2YFwY",
+                            "function": {
+                                "arguments": '{"input": "dksjsdkjdhskdjshdskhjkhlk"}',
+                                "name": "tool_name",
+                            },
+                            "type": "function",
+                            "index": 0,
+                        },
+                        {
+                            "id": "call_Vw6RaqV2n5aaANXEdp5pYxo2",
+                            "function": {
+                                "arguments": '{"input": "jkljlkjlkjlkjlk"}',
+                                "name": "tool_name",
+                            },
+                            "type": "function",
+                            "index": 1,
+                        },
+                        {
+                            "id": "call_hBIKwldUEGlNh6NlSXil62K4",
+                            "function": {
+                                "arguments": '{"input": "jkjlkjlkjlkj;lj"}',
+                                "name": "tool_name",
+                            },
+                            "type": "function",
+                            "index": 2,
+                        },
+                    ],
+                },
+            ],
+            client=client,
+        )
+
+        mock_create.assert_called_once()