chore: Add OpenAI compatibility for Ollama embeddings (#2440)

# What does this PR do?
This PR adds OpenAI compatibility for Ollama embeddings. Closes
https://github.com/meta-llama/llama-stack/issues/2428

Summary of changes:
- `llama_stack/providers/remote/inference/ollama/ollama.py`
- Implements the OpenAI embeddings endpoint for Ollama, replacing the
NotImplementedError with a full function that validates the model,
prepares parameters, calls the client, encodes embedding data
(optionally in base64), and returns a correctly structured response.
- Updates import statements to include the new embedding response
utilities.

- `llama_stack/providers/utils/inference/litellm_openai_mixin.py`
- Refactors the embedding data encoding logic to use a new shared
utility (`b64_encode_openai_embeddings_response`) instead of inline
base64 encoding and packing logic.
   - Cleans up imports accordingly.

- `llama_stack/providers/utils/inference/openai_compat.py`
- Adds `b64_encode_openai_embeddings_response` to handle encoding OpenAI
embedding outputs (including base64 support) in a reusable way.
- Adds `prepare_openai_embeddings_params` utility for standardizing
embedding parameter preparation.
   - Updates imports to include the new embedding data class.

- `tests/integration/inference/test_openai_embeddings.py`
- Removes `"remote::ollama"` from the list of providers that skip OpenAI
embeddings tests, since support is now implemented.

## Note

There was one minor issue, which required me to override the
`OpenAIEmbeddingsResponse.model` name with
`self._get_model(model).identifier` name, which is very unsatisfying.

## Test Plan
Unit Tests and integration tests

---------

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

llama_stack/providers/remote/inference/ollama/ollama.py

@@ -33,7 +33,6 @@ from llama_stack.apis.inference import (
     JsonSchemaResponseFormat,
     LogProbConfig,
     Message,
-    OpenAIEmbeddingsResponse,
     ResponseFormat,
     SamplingParams,
     TextTruncation,
@@ -46,6 +45,8 @@ from llama_stack.apis.inference.inference import (
     OpenAIChatCompletion,
     OpenAIChatCompletionChunk,
     OpenAICompletion,
+    OpenAIEmbeddingsResponse,
+    OpenAIEmbeddingUsage,
     OpenAIMessageParam,
     OpenAIResponseFormatParam,
 )
@@ -62,8 +63,10 @@ from llama_stack.providers.utils.inference.model_registry import (
 from llama_stack.providers.utils.inference.openai_compat import (
     OpenAICompatCompletionChoice,
     OpenAICompatCompletionResponse,
+    b64_encode_openai_embeddings_response,
     get_sampling_options,
     prepare_openai_completion_params,
+    prepare_openai_embeddings_params,
     process_chat_completion_response,
     process_chat_completion_stream_response,
     process_completion_response,
@@ -386,7 +389,35 @@ class OllamaInferenceAdapter(
         dimensions: int | None = None,
         user: str | None = None,
     ) -> OpenAIEmbeddingsResponse:
-        raise NotImplementedError()
+        model_obj = await self._get_model(model)
+        if model_obj.model_type != ModelType.embedding:
+            raise ValueError(f"Model {model} is not an embedding model")
+
+        if model_obj.provider_resource_id is None:
+            raise ValueError(f"Model {model} has no provider_resource_id set")
+
+        # Note, at the moment Ollama does not support encoding_format, dimensions, and user parameters
+        params = prepare_openai_embeddings_params(
+            model=model_obj.provider_resource_id,
+            input=input,
+            encoding_format=encoding_format,
+            dimensions=dimensions,
+            user=user,
+        )
+
+        response = await self.openai_client.embeddings.create(**params)
+        data = b64_encode_openai_embeddings_response(response.data, encoding_format)
+
+        usage = OpenAIEmbeddingUsage(
+            prompt_tokens=response.usage.prompt_tokens,
+            total_tokens=response.usage.total_tokens,
+        )
+        # TODO: Investigate why model_obj.identifier is used instead of response.model
+        return OpenAIEmbeddingsResponse(
+            data=data,
+            model=model_obj.identifier,
+            usage=usage,
+        )
 
     async def openai_completion(
         self,