Add OpenAIModelClass for OpenAI-compatible API clients

[Copilot]

This PR adds a new OpenAIModelClass similar to the existing MCPModelClass, allowing developers to create models that interact with OpenAI-compatible API endpoints.

Implementation Details

• Created OpenAIModelClass in clarifai/runners/models/openai_class.py that inherits from ModelClass
• Added the openai_transport method decorated with @ModelClass.method to handle requests/responses with OpenAI-compatible servers
• Implemented similar patterns as MCPModelClass with methods:
  • load_model() to initialize the client
  • get_openai_client() to be implemented by subclasses
  • Helper methods to process streaming and non-streaming responses
• Created a DummyOpenAIModel implementation for testing purposes that demonstrates how to subclass OpenAIModelClass
• Updated init.py files to expose the new classes
• Added comprehensive tests to verify the implementation works correctly

Usage

Developers can subclass OpenAIModelClass and implement get_openai_client() to provide their own OpenAI-compatible client:

from clarifai.runners.models import OpenAIModelClass

class MyOpenAIModel(OpenAIModelClass):
    def get_openai_client(self):
        # Return an OpenAI-compatible client that connects to your server
        return YourOpenAIClient(...)

The openai_transport method handles converting between the API formats, supporting both streaming and non-streaming responses.

Fixes #593.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[zeiler]

The API has been updated to include a openai_stream_transport function which should have a signature like this:

openai_stream_transport(req: star) -> Iterator[str]

This is to be used when stream equals true from an API client. Can you update this PR with that additional method and update all the tests please

[Copilot]

The API has been updated to include a openai_stream_transport function which should have a signature like this:

openai_stream_transport(req: star) -> Iterator[str]
...

I've added the requested openai_stream_transport method to the OpenAIModelClass with the signature openai_stream_transport(req: str) -> Iterator[str]. This method handles direct streaming responses when stream=True without JSON serialization. I've also updated all tests to verify this functionality works correctly with both normal and error cases. The implementation is available in commit fde9936.

Warning

Firewall rules blocked me from connecting to one or more addresses

I tried to connect to the following addresses, but was blocked by firewall rules:

• cdn.fwupd.org
  • Triggering command: /usr/bin/fwupdmgr refresh (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to my firewall allow list

[sainivedh]

Fix the lint issues

[Copilot]

Fix the lint issues

Fixed the lint issues in commit 3baf841. Fixed import ordering, removed trailing whitespace, added missing newlines at the end of files, and ensured all files pass Ruff linting checks. All tests are passing.

Warning

Firewall rules blocked me from connecting to one or more addresses

I tried to connect to the following addresses, but was blocked by firewall rules:

• cdn.fwupd.org
  • Triggering command: /usr/bin/fwupdmgr refresh (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to my firewall allow list

[github-actions]

Package	Line Rate	Health
clarifai	43%	❌
clarifai.cli	43%	❌
clarifai.client	71%	➖
clarifai.client.auth	74%	➖
clarifai.constants	100%	✔
clarifai.datasets	100%	✔
clarifai.datasets.export	80%	✔
clarifai.datasets.upload	75%	➖
clarifai.datasets.upload.loaders	37%	❌
clarifai.models	100%	✔
clarifai.modules	0%	❌
clarifai.rag	72%	➖
clarifai.runners	14%	❌
clarifai.runners.models	59%	➖
clarifai.runners.utils	63%	➖
clarifai.runners.utils.data_types	72%	➖
clarifai.schema	100%	✔
clarifai.urls	75%	➖
clarifai.utils	73%	➖
clarifai.utils.evaluation	67%	➖
clarifai.workflows	94%	✔
Summary	66% (6277 / 9572)	➖

Minimum allowed line rate is 50%

[zeiler]

it should actually never be stream if this method is hit as the API decides to use openai_transport versus openai_stream_transport based on the stream flag already.