Split textgen and embedding models in UI

[estohlmann]

Description of changes:

• Updated UI to pull Models from config vs API, this allows us to get supplemental info
• UI now only separates textgen vs embedding models in dropdowns, also only enables streaming for models that support it
• Fixed longstanding bug where switching to a model with streaming active to a model that doesn't support streaming "locks" streaming as enabled

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[estohlmann]

update to call the describe API and uses the results to map against the config. I was a little forward looking with my approach and this accounts for the following:

• If a model is returned in the API that isn't in the config it will add it to both drop-downs and allow streaming to be toggled on
• If a model isn't returned that IS listed in the config, it will NOT show up in the dropdowns