sort strings according to edit distance
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
Brian Hicks 166a8d8f56 update to nixpkgs-unstable 2 years ago
src don't make target required 2 years ago
.envrc add a .envrc 2 years ago
.gitignore add basic CLI 2 years ago
BENCHMARKING.md par_iter to go even faster! 2 years ago
Cargo.lock go back to clap 2 years ago
Cargo.toml go back to clap 2 years ago
LICENSE correct capitalization of Wikipedia 3 years ago
README.md revise non-nix installation instructions 2 years ago
default.nix add compatibility layer 2 years ago
flake.lock update to nixpkgs-unstable 2 years ago
flake.nix update to nixpkgs-unstable 2 years ago
shell.nix add compatibility layer 2 years ago

README.md

Similar Sort

This is a small Go program that will:

  1. take a reference string as the first argument
  2. and a list of candidate strings in stdin
  3. and output the candidates sorted according to their edit distance from the reference, lowest first.

"What use is this?" you may ask! Well! It turns out to be really useful to do fuzzy file finding a large project.

When I am in some filesystem hierarchy and I trigger my fuzzy-finder, I want to see sibling files before I see similarly-named files further away. I also want to match on test files pretty easily. Say I have this project structure:

example
└── src
    ├── Main.elm
    └── Page
        └── Learn
            └── Home
                ├── Main.elm
                └── View.elm

If I am in src/Page/Learn/Home/View.elm and I want to get to the sibling file Main.elm, the default fzf config shows me src/Main.elm first. That's not what I wanted!

But if I sort the files instead by piping them through similar-sort src/Page/Learn/Home/View.elm, the sibling file will show up first. This works surprisingly well; I really like it!

It could probably perform a little better by doing some heuristic based on equivalent file structure except for the addition/removal of "tests", "specs", etc, but I haven't bothered yet.

Installing

If you have nix installed, clone this and type:

nix-env -if .

After this, the similar-sort binary should be available on your PATH.

If you don't have nix, you'll need to install a Rust compiler toolchain yourself and run cargo build.

Adding to Vim

Add this to your vim config:

nnoremap <silent> <C-t> :call fzf#run(fzf#wrap({
  \ "source": "git ls-files --others --cached --exclude-standard \| similar-sort " . @% . " \| grep -v " . @%,
  \ "sink": "edit",
  \ "options": "--tiebreak index"
  \ }))<CR>

(You'll need fzf and fzf.vim installed.) This will bind ctrl-t to the fuzzy finder. When you select a match, it will open in the current pane.

If you want to split or vsplit, change "sink": "edit" to "sink": "split" or "sink": "vsplit". See the docs for fzf#run for more customization options.

Adding to Kakoune

Assuming you're using tmux as your window manager, integration looks something like this:

define-command -docstring 'open files named similarly to the current buffer' open-similar %{
    tmux-terminal-horizontal sh -c %{
        set -euo pipefail
        FILE="$(git ls-files --others --cached --exclude-standard | similar-sort $1 | grep -v $1 | fzf --tiebreak index)"
        printf "evaluate-commands -client %s edit %s\n" "$2" "$FILE" | kak -p $3
    } -- %val{bufname} %val{client} %val{session}
}

I have this bound to - with map global normal <minus> ': open-similar<ret>'.

License

CC BY-SA 4.0. See LICENSE in the root of the project for details.