chore: turn OpenAIMixin into a pydantic.BaseModel (#3671)

# What does this PR do?

- implement get_api_key instead of relying on
LiteLLMOpenAIMixin.get_api_key
 - remove use of LiteLLMOpenAIMixin
 - add default initialize/shutdown methods to OpenAIMixin
 - remove __init__s to allow proper pydantic construction
- remove dead code from vllm adapter and associated / duplicate unit
tests
 - update vllm adapter to use openaimixin for model registration
 - remove ModelRegistryHelper from fireworks & together adapters
 - remove Inference from nvidia adapter
 - complete type hints on embedding_model_metadata
- allow extra fields on OpenAIMixin, for model_store, __provider_id__,
etc
 - new recordings for ollama
 - enhance the list models error handling
- update cerebras (remove cerebras-cloud-sdk) and anthropic (custom
model listing) inference adapters
 - parametrized test_inference_client_caching
- remove cerebras, databricks, fireworks, together from blanket mypy
exclude
 - removed unnecessary litellm deps

## Test Plan

ci

llama_stack/providers/remote/inference/vertexai/__init__.py

@@ -10,6 +10,6 @@ from .config import VertexAIConfig
 async def get_adapter_impl(config: VertexAIConfig, _deps):
     from .vertexai import VertexAIInferenceAdapter
 
-    impl = VertexAIInferenceAdapter(config)
+    impl = VertexAIInferenceAdapter(config=config)
     await impl.initialize()
     return impl

llama_stack/providers/remote/inference/vertexai/vertexai.py

@@ -4,29 +4,19 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-from typing import Any
 
 import google.auth.transport.requests
 from google.auth import default
 
-from llama_stack.apis.inference import ChatCompletionRequest
-from llama_stack.providers.utils.inference.litellm_openai_mixin import (
-    LiteLLMOpenAIMixin,
-)
 from llama_stack.providers.utils.inference.openai_mixin import OpenAIMixin
 
 from .config import VertexAIConfig
 
 
-class VertexAIInferenceAdapter(OpenAIMixin, LiteLLMOpenAIMixin):
-    def __init__(self, config: VertexAIConfig) -> None:
-        LiteLLMOpenAIMixin.__init__(
-            self,
-            litellm_provider_name="vertex_ai",
-            api_key_from_config=None,  # Vertex AI uses ADC, not API keys
-            provider_data_api_key_field="vertex_project",  # Use project for validation
-        )
-        self.config = config
+class VertexAIInferenceAdapter(OpenAIMixin):
+    config: VertexAIConfig
+
+    provider_data_api_key_field: str = "vertex_project"
 
     def get_api_key(self) -> str:
         """
@@ -41,8 +31,7 @@ class VertexAIInferenceAdapter(OpenAIMixin, LiteLLMOpenAIMixin):
             credentials.refresh(google.auth.transport.requests.Request())
             return str(credentials.token)
         except Exception:
-            # If we can't get credentials, return empty string to let LiteLLM handle it
-            # This allows the LiteLLM mixin to work with ADC directly
+            # If we can't get credentials, return empty string to let the env work with ADC directly
             return ""
 
     def get_base_url(self) -> str:
@@ -53,23 +42,3 @@ class VertexAIInferenceAdapter(OpenAIMixin, LiteLLMOpenAIMixin):
         Source: https://cloud.google.com/vertex-ai/generative-ai/docs/start/openai
         """
         return f"https://{self.config.location}-aiplatform.googleapis.com/v1/projects/{self.config.project}/locations/{self.config.location}/endpoints/openapi"
-
-    async def _get_params(self, request: ChatCompletionRequest) -> dict[str, Any]:
-        # Get base parameters from parent
-        params = await super()._get_params(request)
-
-        # Add Vertex AI specific parameters
-        provider_data = self.get_request_provider_data()
-        if provider_data:
-            if getattr(provider_data, "vertex_project", None):
-                params["vertex_project"] = provider_data.vertex_project
-            if getattr(provider_data, "vertex_location", None):
-                params["vertex_location"] = provider_data.vertex_location
-        else:
-            params["vertex_project"] = self.config.project
-            params["vertex_location"] = self.config.location
-
-        # Remove api_key since Vertex AI uses ADC
-        params.pop("api_key", None)
-
-        return params