- AI-powered text & code completion client.
Currently supported backends: Ollama, Hugging Face’s Inference API.
- Demo:
demo.webm
(Maybe not up to date) Mirror: https://github.com/daanturo/starhugger.el
Starhugger can be installed from MELPA - Starhugger, using the command package-install.
Of if you want to install from source, add one of the following to your configuration:
;; package-vc.el (built-in from Emacs 29 and above)
(unless (package-installed-p 'starhugger)
(package-vc-install '(starhugger :url "https://gitlab.com/daanturo/starhugger.el")))
;; straight.el
(straight-use-package '(starhugger :files (:defaults "*.py")))
;; Doom
(package! starhugger :recipe (:files (:defaults "*.py")))
;; elpaca.el
(elpaca (starhugger :repo "https://gitlab.com/daanturo/starhugger.el" :files (:defaults "*.py")))
Or any package manager of your choice.
- Previewing overlay:
starhugger-trigger-suggestionto display the suggestion.
M-x starhugger-show-next-suggestion and starhugger-show-prev-suggestion to cycle suggestions.
M-x starhugger-accept-suggestion to insert current suggestion.
starhugger-dismiss-suggestion (bound to C-g by default when showing) to cancel.
There is also starhugger-auto-mode (non-global minor mode), but take note of its usage for the Hugging Face text inference backend because of the limit rate.
;; Optionally use another model, it's best to be set before loading this package, or use Emacs 29+ `setopt'
(setq starhugger-model-id "codellama")
(global-set-key (kbd "M-\\") #'starhugger-trigger-suggestion)
(with-eval-after-load 'starhugger
;; `starhugger-inline-menu-item' makes a conditional binding that is only active at the inline suggestion start
(define-key starhugger-inlining-mode-map (kbd "TAB") (starhugger-inline-menu-item #'starhugger-accept-suggestion))
(define-key starhugger-inlining-mode-map (kbd "M-[") (starhugger-inline-menu-item #'starhugger-show-prev-suggestion))
(define-key starhugger-inlining-mode-map (kbd "M-]") (starhugger-inline-menu-item #'starhugger-show-next-suggestion))
(define-key starhugger-inlining-mode-map (kbd "M-f") (starhugger-inline-menu-item #'starhugger-accept-suggestion-by-word)))Use the Hugging Face backend:
(setq starhugger-completion-backend-function #'starhugger-hugging-face-inference-api)
(setq starhugger-hugging-face-api-token "hf_ your token here")
(setopt starhugger-model-id "organization/some-model")
;; Appropriate tokens for the model
(setq starhugger-fill-tokens '("<PRE>" "<SUF>" "<MID>"))
(setq starhugger-stop-tokens '("<|endoftext|>" "<EOT>"))Additional settings:
;; For evil users, dismiss after pressing ESC twice
(defvar my-evil-force-normal-state-hook '())
(defun my-evil-run-force-normal-state-hook-after-a (&rest _)
(run-hooks 'my-evil-force-normal-state-hook))
(advice-add #'evil-force-normal-state
:after #'my-evil-run-force-normal-state-hook-after-a)
;; Workaround conflict with `blamer.el'
;; (https://github.com/Artawower/blamer.el): when at the end of line, blamer's
;; overlay's `after-string' property will display before starhugger's
;; `display' property, which will result in starhugger's part of suggestion on
;; current line (1) being pushed out of the display
;; <before point>| commit info<right edge of the window><suggestion after point, before newline>
;; <the rest of suggestion>
;; workaround: disable `blamer-mode' while `starhugger-inlining-mode'
(defvar-local my-starhugger-inlining-mode--blamer-mode-state nil)
(defvar-local blamer-mode nil)
(defun my-starhugger-inlining-mode-h ()
(if starhugger-inlining-mode
(progn
(add-hook 'my-evil-force-normal-state-hook
(lambda () (starhugger-dismiss-suggestion t))
nil t)
(setq my-starhugger-inlining-mode--blamer-mode-state blamer-mode)
(when my-starhugger-inlining-mode--blamer-mode-state
(blamer-mode 0)))
(progn
(when (and my-starhugger-inlining-mode--blamer-mode-state
(not blamer-mode))
(blamer-mode 1)))))
(add-hook 'starhugger-inlining-mode-hook #'my-starhugger-inlining-mode-h)When using the Hugging Face text inference backend, remember to set starhugger-hugging-face-api-token (from https://huggingface.co/settings/tokens), otherwise you may easily get hit by the limit rate.
From the model (https://huggingface.co/bigcode/starcoder):
- Doesn’t use
num_return_sequences(detailed_parameters) to return multiple responses, workaround by making multiple requests. - Doesn’t use
use_cache, current workaround is forcing a different response via randomizingtemperature.
Emacs overlays are used under the hood to display inline suggestion, there are some shortcomings with this approach:
- Not possible to display PRE|<ov>SUF without using 2 different types of overlay properties when SUF isn’t emtpy (in the middle of the buffer) and empty (at buffer end)
- At the end of the buffer (overlaystart = overlay-end), the overlay’s
keymapproperty doesn’t work - Conflict with https://github.com/Artawower/blamer.el, mentioned in “Example configuration”
- [x] Support setting parameters at https://huggingface.co/docs/api-inference/detailed_parameters#text-generation-task.
- [x] VSCode-like previewing overlays: take after https://github.com/zerolfx/copilot.el.
- [x] Let
starhugger-trigger-suggestionfetch about 3 suggestions to quickly cycle. - [x] Fill-in-the-middle support https://github.com/huggingface/huggingface-vscode: <fim_prefix>〈code before〉<fim_suffix>〈code after〉<fim_middle>.
- [ ] More robust and reliable method to show a different suggestion.
- [-] Batch-previewing multiple suggestions, maybe with syntax highlighting.
- [-] Support for auto-completing when typing: investigate Emacs’s built-in
completion-at-point-functions’s asynchronous capabilities, or another framework? Current implementation:starhugger-auto-modeusing overlays. - [ ] Find a way to take other files into account Copilot Internals | thakkarparth007.github.io, or a RAG system.
- [x] Separate frontend (inline suggestion interface) and backend (the text completion provider), and allow multiple backends, not just Hugging Face’s inference API.
- [ ] With prompt engineering, allow making use of conversational models, not just limited to FIM-supported coding ones.