s-kostyaev / ellama

Ellama is a tool for interacting with large language models from Emacs.
GNU General Public License v3.0
378 stars 27 forks source link

Org-mode support? #17

Closed tusharhero closed 7 months ago

tusharhero commented 7 months ago

I wanted to know if we can use Org-mode instead of markdown-mode.

s-kostyaev commented 7 months ago

Not sure if models have enough data in org format in training set. I don't think it will works. Maybe we can do it with translation layer inside emacs, but it will make things harder and I can't see what advantages we will get.

Maybe better idea is to support some commands for org mode, but I have no ideas here.

tusharhero commented 7 months ago

oh, does the LLM need to know the markup language before we can use it?

s-kostyaev commented 7 months ago

For example if there is code in model response, probably it will be inside markdown code block. Same with lists and tables. What will we see if we try to enable org-mode for this type of content? I think something strange.

tusharhero commented 7 months ago

https://github.com/karthink/gptel

How do they handle it?

s-kostyaev commented 7 months ago

I will check it, thank you.

s-kostyaev commented 7 months ago

@tusharhero try new command ellama-complete in org-mode buffer. Will be soon on Melpa.

s-kostyaev commented 6 months ago

@tusharhero Implemented. Feedback id welcome.