Roadmap with planned future features about promptvision

[Automaticism]

• Refresh button that will update metadata and such without restarting the gallery
• Save favourites to a favourites directory
• NLP analyze prompts
• Export prompts to (such as dynamic prompts)
• SD WebUI API send-to [controlnet | img2img | ... ]
• Improve thumbnails so that browsing becomes more pleasant

[Automaticism]

I've started working on porting this over to using Streamlit as the UI.
Currently I have a working solution that loads up a directory (and all sub directories), with browsing the images and seeing the information.

I've unfortunately been under the weather and a lack of motivation has delayed further developments.

Currently done parts:

• Thumbnails
• Browsing
• Directory support
• Prompt analysis - clustering, levenstein distance, intent, and a few more. This is basically only limited by the imagination (and the NLP frameworks that exist)

To be implemented:

• Form for setting rating, favorite, tag and categorizes (I have the rough implementation idea, additionally with multi-select support)
• Deletion, moving, copying -- utility functions
• Option to load in the exif & metadata csv instead of inserting a directory