search

Robitx / gp.nvim

Gp.nvim (GPT prompt) Neovim AI plugin: ChatGPT sessions & Instructable text/code operations & Speech to text [OpenAI]
MIT License
537 stars 50 forks source link

idea: create dall-e request and download the generated images #58

Closed divramod closed 7 months ago

divramod commented 7 months ago

i really like the newly integrated dall-e image generation in the chat and i use it a lot.

do you have an idea, if it would be possible, to call the model generating the images?

after the request finished, it would be nice to have a download image dialog, where one could choose the location.

Robitx commented 7 months ago

@divramod Hey, the latest version supports :GpImage to handle the generation flow and :GpImageAgent image

divramod commented 7 months ago

awesome stuff! simply awesome!

divramod commented 7 months ago

sooo nice!

would it be possible to configure a default folder for the images like with chat_dir, whisper_dir, state_dir?

i just switched to a very small config with only the dirs configured, after the hint after the latest update.

maybe a second command like GpImageDefault?

btw, GpNextAgent only changes the command agent. Is this on purpose?

i changed the chat agend in the state.json

Robitx commented 7 months ago

btw, GpNextAgent only changes the command agent. Is this on purpose?

it changes the agent based on where you are - if you're in a chat buffer it cycles between chat agents, anywhere else it cycles between command agents

maybe a second command like GpImageDefault?

what should it do? just as chat and command agents, the "image agent" (configuration) is persisted in state_dir across neovim instances, so it you set it for example to DALL-E-3-1792x1024-vivid it will be used for images until you change it.

would it be possible to configure a default folder for the images like with chat_dir, whisper_dir, state_dir?

this should be trivial, will do before crashing to bed for some sleep