Python implementation always doubles up

[Maclean-D]

Is your feature request related to a problem? Please describe.

No response

Describe the solution you'd like

Open-Interpreter streams the messages first, and then sends the full thing again after start_of_code which is fine because you can at least filter it out because it doesn't start with {"code": ""}. However, when it starts a message after, it spits out the code again and then the rest of the message. because the messages don't start with {"message": ""} there's no easy way to distinguish/filter out between messages and when it's giving you the code again. I propose we remove sending the code again, or put it in something like {"executed_code": ""}

Describe alternatives you've considered

No response

Additional context

No response