This is an experimental tool which generates autocompletion suggestions for Elm code based on available type information. It lets you do things like this:
or this:
There are basically two steps to this:
- Given a syntactically correct Elm file, it will try to infer the type of the expression at the current cursor position.
- If this was successfull, it will generate new expressions which are of this type using locally available values. (Take a look at elm/projects for more information on this idea.)
Please keep in mind, that the type inference algorithm is not completely implemented, yet. So you might run into weird bugs or may not get any completions at all.
If you want to try it out locally, you can run
$ npm run elm
$ npm linkThen, inside the folder which contains your elm.json file, run
$ elm-autocomplete --serverTo get suggestions at a certain position of some Elm file, you have to run the
following command in a different terminal (again inside the folder with your
elm.json file)
$ cat src/Main.elm | elm-autocomplete --file src/Main.elm --column 5 --row 10If successfull, this should output a list of possible expressions.
There is a very basic vim integration available using
coc.nvim. To use this you have to copy
vim/elm.js
to ~/.vim/coc-extensions/elm.js. Then, if you have the server running, you
should get the autocompletion suggestions within vim.
There are some tests available, which can be run via
$ npx elm-test --watchIf you are not satisfied with the suggestions, you can try to adjust the
completion generator yourself. The CLI tool uses the default Generator within
the Generator module. So, adjusting this and rebuilding the CLI might be
a good place to start!
- The CLI-wrapper is an adaptation of zwilias/elm-xref
- The parsing of the Elm sources is done using stil4m/elm-syntax
- When implementing the type inference algorithm, the article Write You a Haskell was very helpfull, as well as the actual Elm compiler itself.