fix(oci): address PR review findings

- types/llms/oci.py: add "TOOL_CALL" to CohereChatResponse.finishReason
  Literal so Pydantic does not raise ValidationError on non-streaming
  Cohere tool-use calls (Greptile P1)
- test_oci_cohere_tool_calls.py: add test covering TOOL_CALL finish reason
- model_prices_and_context_window.json: remove 6 duplicate oci/cohere.embed-*
  keys that were silently overridden by the more complete entries already
  present in the file (Greptile P1)
- common_utils.py: move OCI_API_VERSION here from chat/transformation.py
  so embed/transformation.py does not need to import chat/transformation;
  change Protocol stub body from ... to pass (CodeQL "statement no effect");
  add comment to sha256_base64 clarifying it implements OCI HTTP signing
  spec, not password hashing (CodeQL false positive)
- chat/transformation.py: import CustomStreamWrapper from
  litellm_core_utils.streaming_handler instead of litellm.utils to reduce
  import cycle depth (CodeQL cyclic import)
- chat/cohere.py, chat/generic.py: import Usage and
  ChatCompletionMessageToolCall from litellm.types.utils instead of
  litellm.utils for the same reason
- embed/transformation.py: import OCI_API_VERSION from common_utils
  instead of chat/transformation (removes the embed→chat import edge)

litellm/llms/oci/chat/cohere.py

@@ -37,7 +37,7 @@ from litellm.types.utils import (
     ModelResponseStream,
     StreamingChoices,
 )
-from litellm.utils import Usage
+from litellm.types.utils import Usage
 
 
 def _extract_text_content(content: Any) -> str:

litellm/llms/oci/chat/generic.py

@@ -38,7 +38,7 @@ from litellm.types.utils import (
     ModelResponseStream,
     StreamingChoices,
 )
-from litellm.utils import ChatCompletionMessageToolCall, Usage
+from litellm.types.utils import ChatCompletionMessageToolCall, Usage
 
 # Maps OpenAI role names to OCI GENERIC role names.
 open_ai_to_generic_oci_role_map: Dict[str, OCIRoles] = {

litellm/llms/oci/chat/transformation.py

@@ -48,6 +48,7 @@ from litellm.llms.oci.chat.generic import (
     handle_generic_stream_chunk,
 )
 from litellm.llms.oci.common_utils import (
+    OCI_API_VERSION,
     OCIError,
     OCIRequestWrapper,  # re-exported for backwards compatibility
     get_oci_base_url,
@@ -68,7 +69,7 @@ from litellm.types.utils import (
     ModelResponse,
     ModelResponseStream,
 )
-from litellm.utils import CustomStreamWrapper
+from litellm.litellm_core_utils.streaming_handler import CustomStreamWrapper
 
 if TYPE_CHECKING:
     from litellm.litellm_core_utils.litellm_logging import Logging as _LiteLLMLoggingObj
@@ -78,9 +79,6 @@ else:
     LiteLLMLoggingObj = Any
 
 
-# OCI GenAI REST API version — stable since service launch, unlikely to change
-OCI_API_VERSION = "20231130"
-
 # Streaming timeout — generous because OCI models may need to warm up on first request
 STREAMING_TIMEOUT = 60 * 5
 

litellm/llms/oci/common_utils.py

@@ -25,6 +25,10 @@ except ImportError:
     _litellm_version = "0.0.0"
 
 
+# OCI GenAI REST API version — stable since service launch, unlikely to change
+OCI_API_VERSION = "20231130"
+
+
 def _require_cryptography() -> None:
     if not _CRYPTOGRAPHY_AVAILABLE:
         raise ImportError(
@@ -62,7 +66,8 @@ class OCISignerProtocol(Protocol):
 
     def do_request_sign(
         self, request: Any, *, enforce_content_headers: bool = False
-    ) -> None: ...
+    ) -> None:
+        pass
 
 
 @dataclass
@@ -87,6 +92,10 @@ class OCIRequestWrapper:
 
 
 def sha256_base64(data: bytes) -> str:
+    # SHA-256 is used here to compute the x-content-sha256 header required by the
+    # OCI HTTP signing specification (RSA-SHA256 request signing), not for password
+    # or secret hashing.  This is the correct and mandated algorithm for this purpose.
+    # See: https://docs.oracle.com/en-us/iaas/Content/API/Concepts/signingrequests.htm
     digest = hashlib.sha256(data).digest()
     return base64.b64encode(digest).decode()
 

litellm/llms/oci/embed/transformation.py

@@ -29,8 +29,8 @@ import httpx
 import litellm
 from litellm.llms.base_llm.chat.transformation import BaseLLMException
 from litellm.llms.base_llm.embedding.transformation import BaseEmbeddingConfig
-from litellm.llms.oci.chat.transformation import OCI_API_VERSION
 from litellm.llms.oci.common_utils import (
+    OCI_API_VERSION,
     OCIError,
     get_oci_base_url,
     resolve_oci_credentials,

litellm/types/llms/oci.py

@@ -379,7 +379,13 @@ class CohereChatResponse(BaseModel):
     text: str
     apiFormat: Literal["COHERE"] = "COHERE"
     finishReason: Literal[
-        "COMPLETE", "ERROR_TOXIC", "ERROR_LIMIT", "ERROR", "USER_CANCEL", "MAX_TOKENS"
+        "COMPLETE",
+        "ERROR_TOXIC",
+        "ERROR_LIMIT",
+        "ERROR",
+        "USER_CANCEL",
+        "MAX_TOKENS",
+        "TOOL_CALL",
     ]
 
     # Optional fields

model_prices_and_context_window.json

@@ -25959,47 +25959,6 @@
         "supports_response_schema": false,
         "supports_streaming": true
     },
-    "oci/cohere.embed-english-v3.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 512,
-        "mode": "embedding",
-        "output_vector_size": 1024,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/"
-    },
-    "oci/cohere.embed-multilingual-v3.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 512,
-        "mode": "embedding",
-        "output_vector_size": 1024,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/"
-    },
-    "oci/cohere.embed-english-light-v3.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 512,
-        "mode": "embedding",
-        "output_vector_size": 384,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/"
-    },
-    "oci/cohere.embed-multilingual-light-v3.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 512,
-        "mode": "embedding",
-        "output_vector_size": 384,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/"
-    },
-    "oci/cohere.embed-english-image-v3.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 512,
-        "mode": "embedding",
-        "output_vector_size": 1024,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/",
-        "supports_vision": true
-    },
     "oci/cohere.embed-multilingual-image-v3.0": {
         "input_cost_per_token": 1e-07,
         "litellm_provider": "oci",
@@ -26009,15 +25968,6 @@
         "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/",
         "supports_vision": true
     },
-    "oci/cohere.embed-v4.0": {
-        "input_cost_per_token": 1e-07,
-        "litellm_provider": "oci",
-        "max_input_tokens": 128000,
-        "mode": "embedding",
-        "output_vector_size": 1536,
-        "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/",
-        "supports_vision": true
-    },
     "oci/cohere.command-a-reasoning-08-2025": {
         "input_cost_per_token": 1.56e-06,
         "litellm_provider": "oci",

tests/test_litellm/llms/oci/chat/test_oci_cohere_tool_calls.py

@@ -507,6 +507,52 @@ class TestOCICohereToolCalls:
         assert "topP" not in chat_request
         assert "frequencyPenalty" not in chat_request
 
+    def test_cohere_response_finish_reason_tool_call(self):
+        """Test that finishReason='TOOL_CALL' is accepted by Pydantic and mapped to 'tool_calls'."""
+        config = OCIChatConfig()
+
+        mock_cohere_response = {
+            "modelId": "cohere.command-latest",
+            "modelVersion": "1.0",
+            "chatResponse": {
+                "apiFormat": "COHERE",
+                "text": "",
+                "finishReason": "TOOL_CALL",
+                "toolCalls": [
+                    {"name": "get_weather", "parameters": {"location": "London"}}
+                ],
+                "usage": {
+                    "promptTokens": 20,
+                    "completionTokens": 10,
+                    "totalTokens": 30,
+                },
+            },
+        }
+
+        response = httpx.Response(
+            status_code=200,
+            json=mock_cohere_response,
+            headers={"Content-Type": "application/json"},
+        )
+
+        result = config.transform_response(
+            model="cohere.command-latest",
+            raw_response=response,
+            model_response=ModelResponse(),
+            logging_obj={},  # type: ignore
+            request_data={},
+            messages=[],
+            optional_params={},
+            litellm_params={},
+            encoding={},
+        )
+
+        assert isinstance(result, ModelResponse)
+        assert result.choices[0].finish_reason == "tool_calls"
+        assert result.choices[0].message.tool_calls is not None
+        assert len(result.choices[0].message.tool_calls) == 1
+        assert result.choices[0].message.tool_calls[0].function.name == "get_weather"
+
     def test_cohere_vendor_detection(self):
         """Test that Cohere models are correctly identified"""
         assert get_vendor_from_model("cohere.command-latest") == OCIVendors.COHERE