lam: Local Abbrev Manager

Buffer local abbrevs at the speed of light!

The aim is to lower the barrier to define, use and discard context specific shortcuts.

Why?

1. Current method to define abbrevs involves some number of keystrokes under the C-x a prefix that is not easy to remember. We want to reduce the friction in using abbrevs as much as possible.
2. Out of the box, you can only define global (for all buffers) or mode specific (for all buffers of this major mode). There is no way to create an abbrev for a single buffer with the confidence that it will not disturb any other buffers.

Installation

Not on melpa yet. For now:

$ git clone https://github.com/ChanderG/lam

and then in your init file:

(load "<path to lam>/lam")

Bind it to a convenient key, for eg:

(global-set-key "\C-xl" 'lam/control)

Now, you can use C-x l as a single keybinding to do everything - enable lam for current buffer, bring up the lam panel, close it, etc.

Usuage

Basic

Just type abbrevations into the *lam* buffer for them to become instantly available. Abbrevs are reloaded when you switch window focus - no saving or extra commands required!

Abbrev definitions look like the following:

("eia" "Emacs is Awesome!")
("lam" "Local Abbrev Manager")

You can comment out lines using ";". Since the *lam* buffer is reloaded on window focus change with no extra effort from your side: changing abbrev values, adding, removing, commenting, uncommenting all work seamlessly.

Executing Elisp

But, wait. Abbrevs can do a lot more. Did you know that we can call arbitrary elisp from abbrevs?

If there is a third arg, it is wrapped into a lambda and called after expansion. The second arg is expanded first - you may not want to use it in this case.

("sayhi" "" (message "Hi"))

Will echo "Hi" to the minibuffer.

Just think of the possibilities!!!

Binding Key Sequences

If you don't remember/can't be bothered to lookup exact elisp functions, you can use the following approach to bind key sequences.

For example in org-mode, we can insert timestamps using C-c .. But, to add a timestamp for current time with the time portion requires 2 universal prefixes. The key sequence is thus: C-u C-u C-c .. In some workflows (not all - this is important, otherwise I would just bind a global abbrev), I insert a lot of timestamps as I take notes.

The "right" way to do this would be:

• Look up help using C-h k to figure out the right function being called: org-time-stamp
• Understand that 2 C-u's translates to an arg of '(16). That took me some time to figure out! See this: http://xahlee.info/emacs/emacs/elisp_universal_argument.html

Finally, we can bind this as:

("Now" "" (org-time-stamp '(16)))

Instead of all this effort, you can do:

("Now" "" (execute-kbd-macro (read-kbd-macro "C-u C-u C-c .")))

But, even this is too much, so we provide an easy to use macro that does exactly this:

("Now" "" (lam/kbd "C-u C-u C-c ."))

Using Clipboard data

You can use external data directly from the elisp execution mode for custom abbrevations. Consider the following example:

("fn" "" (insert "\\footnote{" (substring-no-properties (x-get-clipboard)) "}"))

which inserts a footnote into a Latex file at point using whatever is in your clipboard - a url of the website for example that you just copied.

To simplify this use case, we provide a convenience macro lam/cb that does just this:

("fn" "" (insert "\\footnote{" (lam/cb) "}"))

Possible Features/TODOs

1. Add a command to easily add abbrev for word under cursor.
2. Command to directly add sexp at point into lam and delete it out: allowing seamless adding of entried from primary buffer.
3. Highlight candidates for abbrev by word length and frequency
4. Support multiple lam buffers.
5. Save/load lam buffers.
6. Use a single lam buffer for an entire project

LICENSE

MIT