Track costs for streaming with OpenAI

[brenkao]

Is your feature request related to a problem? Please describe.
Prior versions of openai did not have usage stats when streaming.

Describe the solution you'd like
Add stream_options: {"include_usage": true}. Add total_cost as a property of OpenAICallResponseChunk.

Additional context
OpenAI Cookbook Reference

[willbakst]

We may also want to consider updating the generator to just return total cost separate from the response chunk so the generator would have type Generator[BaseCallResponseChunkT, None, Optional[float]] and then return total cost at the end of the generator if available, otherwise return None.

[brenkao]

The stream_options: {"include_usage": true} has been implemented with #239