Improve handling of triggers words for LoRAs

[yondonfu]

A LoRA typically is associated with word(s) that can be included in the prompt to trigger/strengthen the effect of the LoRA in the output. In some cases, the effect is visible just by loading/fusing the LoRA, but oftentimes using the trigger word in the prompt is more effective.

A cumbersome part of using LoRAs in Scope right now is that you have to manually lookup the trigger word and add it to the prompt for the specific LoRA that you're using. And you do this every time you use a new LoRA or switch to a different LoRA. It would be great to have a lower friction way for handling this.

Here is a strawman to encourage more ideas for improvement on this front:

Let's assume that we have a way to lookup the the trigger words for a LoRA - this could involve querying the hub where the LoRA is from eg Civitai, HuggingFace, etc. Perhaps we have a "grammar" for prompts where there are different components to the prompt that get combined together into the final prompt. One of the components could be the trigger words. If a LoRA is loaded then the trigger word component could automatically be populated (but still modifiable if needed).