Add support for experimental UIs
What is the problem that this fixes or functionality that this introduces? Does it fix any open issues?
This plays with the idea of having alternative UIs for specific tasks. It specifically creates a "gif editor", which allows rapid iteration on a python script that generates an animated gif.
Give a summary of what the PR does, explaining any non-trivial design decisions
In addition to the new UI, there are two major changes to the core code:
- A React router, which allows us to have multiple pages
- Adding
state.prompt_context, which can be used to add additional information to the LLM prompts. SeeGifEditor.tsxfor an example
Other references
Would love to see a screenie of this, sounds exciting!
Would this also "catch" clicks on images in the workspace folder to display them?
@rbren will you be continuing with this PR? Just checking so the conflicts don't become too cumbersome.
Nah let's ditch it for now--we can always come back to it