ruff-lsp

A Language Server Protocol implementation for Ruff, an extremely fast Python linter and code formatter, written in Rust.

Ruff can be used to replace Flake8 (plus dozens of plugins), Black, isort, pyupgrade, and more, all while executing tens or hundreds of times faster than any individual tool.

ruff-lsp enables Ruff to be used in any editor that supports the LSP, including Neovim, Sublime Text, Emacs and more. For Visual Studio Code, check out the Ruff VS Code extension.

ruff-lsp supports surfacing Ruff diagnostics and providing Code Actions to fix them, but is intended to be used alongside another Python LSP in order to support features like navigation and autocompletion.

Highlights

"Quick Fix" actions for auto-fixable violations (like unused imports)

"Fix all": automatically fix all auto-fixable violations

"Format Document": Black-compatible code formatting

"Organize Imports": isort-compatible import sorting

Installation

ruff-lsp is available as ruff-lsp on PyPI:

pip install ruff-lsp

Community packages

An Alpine Linux package is available in the testing repository:

apk add ruff-lsp

An Arch Linux package is available in the Extra repository:

pacman -S ruff-lsp

Setup

Once installed, ruff-lsp can be used with any editor that supports the Language Server Protocol, including Neovim, Emacs, Sublime Text, and more.

Example: Neovim

To use ruff-lsp with Neovim, follow these steps:

1. Install ruff-lsp from PyPI along with nvim-lspconfig.
2. Set up the Neovim LSP client using the suggested configuration (:h lspconfig-keybindings).
3. Finally, configure ruff-lsp in your init.lua:

-- Configure `ruff-lsp`.
-- See: https://github.com/neovim/nvim-lspconfig/blob/master/doc/server_configurations.md#ruff_lsp
-- For the default config, along with instructions on how to customize the settings
require('lspconfig').ruff_lsp.setup {
  init_options = {
    settings = {
      -- Any extra CLI arguments for `ruff` go here.
      args = {},
    }
  }
}

Upon successful installation, you should see Ruff's diagnostics surfaced directly in your editor:

Note that if you're using Ruff alongside another LSP (like Pyright), you may want to defer to that LSP for certain capabilities, like textDocument/hover:

local on_attach = function(client, bufnr)
  if client.name == 'ruff_lsp' then
    -- Disable hover in favor of Pyright
    client.server_capabilities.hoverProvider = false
  end
end

require('lspconfig').ruff_lsp.setup {
  on_attach = on_attach,
}

And, if you'd like to use Ruff exclusively for linting, formatting, and organizing imports, you can disable those capabilities in Pyright:

require('lspconfig').pyright.setup {
  settings = {
    pyright = {
      -- Using Ruff's import organizer
      disableOrganizeImports = true,
    },
    python = {
      analysis = {
        -- Ignore all files for analysis to exclusively use Ruff for linting
        ignore = { '*' },
      },
    },
  },
}

Ruff also integrates with coc.nvim:

"languageserver": {
  "ruff-lsp": {
    "command": "ruff-lsp",
    "filetypes": ["python"]
  }
}

Example: Sublime Text

To use ruff-lsp with Sublime Text, install Sublime Text's LSP and LSP-ruff package.

Upon successful installation, you should see errors surfaced directly in your editor:

Example: Helix

To use ruff-lsp with Helix, add something like the following to ~/.config/helix/languages.toml (in this case, with auto-format enabled):

[language-server.ruff]
command = "ruff-lsp"

[[language]]
name = "python"
language-servers = [ "ruff" ]
auto-format = true

Upon successful installation, you should see errors surfaced directly in your editor:

As of v23.10, Helix supports the use of multiple language servers for a given language. This enables, for example, the use of ruff-lsp alongside a language server like pyright:

[[language]]
name = "python"
language-servers = [ "pyright", "ruff" ]

[language-server.ruff]
command = "ruff-lsp"

[language-server.ruff.config.settings]
args = ["--ignore", "E501"]

Example: Lapce

To use ruff-lsp with Lapce, install the lapce-ruff-lsp plugin (which wraps ruff-lsp) from the Lapce plugins panel.

Upon successful installation, you should see errors surfaced directly in your editor:

Example: Kate

To use ruff-lsp with Kate, add something like the following to the LSP client's settings.json:

{
  "servers": {
    "python": {
      "command": ["ruff-lsp"],
      "url": "https://github.com/astral-sh/ruff-lsp",
      "highlightingModeRegex": "^Python$"
    }
  }
}

Fix safety

Ruff's automatic fixes are labeled as "safe" and "unsafe". By default, the "Fix all" action will not apply unsafe fixes. However, unsafe fixes can be applied manually with the "Quick fix" action. Application of unsafe fixes when using "Fix all" can be enabled by setting unsafe-fixes = true in your Ruff configuration file or adding --unsafe-fixes flag to the "Lint args" setting.

See the Ruff fix docs for more details on how fix safety works.

Jupyter Notebook Support

ruff-lsp has support for Jupyter Notebooks via the Notebook Document Synchronization capabilities of the Language Server Protocol which were added in 3.17. This allows ruff-lsp to provide full support for all of the existing capabilities available for Python files in Jupyter Notebooks, including diagnostics, code actions, and formatting.

This requires clients, such as Visual Studio Code, to support the notebook-related capabilities. In addition to the editor support, it also requires Ruff version v0.1.3 or later.

Settings

The exact mechanism by which settings will be passed to ruff-lsp will vary by editor. However, the following settings are supported:

Settings	Default	Description
codeAction.disableRuleComment.enable	true	Whether to display Quick Fix actions to disable rules via noqa suppression comments.
codeAction.fixViolation.enable	true	Whether to display Quick Fix actions to autofix violations.
fixAll	true	Whether to register Ruff as capable of handling source.fixAll actions.
format.args	[]	Additional command-line arguments to pass to ruff format, e.g., "args": ["--config=/path/to/pyproject.toml"]. Supports a subset of Ruff's command-line arguments, ignoring those that are required to operate the LSP, like --force-exclude and --verbose.
ignoreStandardLibrary	true	Whether to ignore files that are inferred to be part of the Python standard library.
interpreter	[]	Path to a Python interpreter to use to run the linter server.
lint.args	[]	Additional command-line arguments to pass to ruff check, e.g., "args": ["--config=/path/to/pyproject.toml"]. Supports a subset of Ruff's command-line arguments, ignoring those that are required to operate the LSP, like --force-exclude and --verbose.
lint.enable	true	Whether to enable linting. Set to false to use Ruff exclusively as a formatter.
lint.run	onType	Run Ruff on every keystroke (onType) or on save (onSave).
logLevel	error	Sets the tracing level for the extension: error, warn, info, or debug.
organizeImports	true	Whether to register Ruff as capable of handling source.organizeImports actions.
path	[]	Path to a custom ruff executable, e.g., ["/path/to/ruff"].

Development

• Install just, or see the justfile for corresponding commands.
• Create and activate a virtual environment (e.g., python -m venv .venv && source .venv/bin/activate).
• Install development dependencies (just install). To run the test_format.py test, you need to install a custom ruff build with --features format, e.g. maturin develop --features format -m ../ruff/crates/ruff_cli/Cargo.toml.
• To automatically format the codebase, run: just fmt.
• To run lint and type checks, run: just check.
• To run tests, run: just test. This is just a wrapper around pytest, which you can use as usual.

Release

• Bump the version in ruff_lsp/__init__.py.
• Make sure you use Python 3.7 installed and as your default Python.
• Run python -m venv .venv to create a venv and activate it.
• Run python -m pip install pip-tools to install pip-tools.
• Run rm requirements.txt requirements-dev.txt and then just lock to update ruff.
• Create a new PR and merge it.
• Create a new Release, enter v0.0.x (where x is the new version) into the Choose a tag selector. Click Generate release notes, curate the release notes and publish the release.
• The Release workflow publishes the LSP to PyPI.

License

MIT