feat(litellm): handle streams for litellm code completions

What does this merge request do and why?

Calling api/v4/code_suggestions/completions with stream: true fails with AttributeError: 'async_generator' object has no attribute 'model' and the error is reproducible by the test provided in this merge request. This MR handles the stream response