A Python 2.7 and 3.4+ implementation of the Language Server Protocol.

The base language server requires Jedi to provide Completions, Definitions, Hover, References, Signature Help, and Symbols:

pip install python-language-server

If the respective dependencies are found, the following optional providers will be enabled:

• Rope for Completions and renaming
• Pyflakes linter to detect various errors
• McCabe linter for complexity checking
• pycodestyle linter for style checking
• pydocstyle linter for docstring style checking (disabled by default)
• autopep8 for code formatting
• YAPF for code formatting (preferred over autopep8)

Optional providers can be installed using the extras syntax. To install YAPF formatting for example:

pip install 'python-language-server[yapf]'

All optional providers can be installed using:

pip install 'python-language-server[all]'

If you get an error similar to 'install_requires' must be a string or list of strings then please upgrade setuptools before trying again.

pip install -U setuptools

Installing these plugins will add extra functionality to the language server:

• pyls-mypy Mypy type checking for Python 3
• pyls-isort Isort import sort code formatting
• pyls-black for code formatting using Black

Please see the above repositories for examples on how to write plugins for the Python Language Server. Please file an issue if you require assistance writing a plugin.

Configuration is loaded from zero or more configuration sources. Currently implemented are:

• pycodestyle: discovered in ~/.config/pycodestyle, setup.cfg, tox.ini and pycodestyle.cfg.
• flake8: discovered in ~/.config/flake8, setup.cfg, tox.ini and flake8.cfg

The default configuration source is pycodestyle. Change the pyls.configurationSources setting to ['flake8'] in order to respect flake8 configuration instead.

Overall configuration is computed first from user configuration (in home directory), overridden by configuration passed in by the language client, and then overriden by configuration discovered in the workspace.

To enable pydocstyle for linting docstrings add the following setting in your LSP configuration: ` "pyls.plugins.pydocstyle.enabled": true `

Auto Completion:

Code Linting with pycodestyle and pyflakes:

Signature Help:

Go to definition:

Hover:

Find References:

Document Symbols:

Document Formatting:

To run the test suite:

pip install .[test] && tox

If your code and virtual environment are on a remote host, but you want to edit it on a local machine that has SSHFS-mounted a directory on the remote host, here is how to make your local VS Code use the Python Language Server running on the remote host.

In these instructions, let's pretend that we're mounting /home/kedo/workspace (on a remote Linux host) as /Users/kedo/workspace (on a local macOS computer).

We refer to the Linux host as the server, and the macOS computer running VS Code as the client.

# Check out this repository to a known location.
# It can be anywhere. For this example, we'll put it in the home dir.
$ git clone https://github.com/kennydo/python-language-server.git ~/python-language-server

# Create a new virtual environment for the language server.
# The Python interpreter you choose for the venv here does not have to match
# the Python version in the development virtualenvs.
$ cd ~/python-language-server
$ virtualenv -p python3.6 venv

# Install the `pyls` command
$ pip install .

# Go to your project repository
$ cd ~/path/to/my/python/project

# Make sure there is a virtualenv, and that it has the requirements of your project installed.
$ source virtualenv/bin/activate
$ pip install ...

# Run the `pyls` server.
# The port can be whatever you want.
# Pass in the virtualenvs that you want the language server to look in
# via the `VIRTUALENV_PATHS` environment variable.
# You can pass in multiple venvs by separating them by `:`.
# This command has been split onto multiple lines for ease of viewing.
$ VIRTUALENV_PATHS=~/path/to/my/python/project/virtualenv:~/path/to/another/virtualenv \
    SERVER_MOUNT_PATH='file:///home/kedo/workspace/' \
    CLIENT_MOUNT_PATH='file:///Users/kedo/workspace/' \
    ~/python-language-server/venv/bin/pyls \
    --host 0.0.0.0 \
    --port 2525 \
    --tcp

1. Install VSCode for Mac
2. From within VSCode View -> Command Palette, then type shell and run install 'code' command in PATH
3. Install yarn (ex: brew install yarn)

# Check out this repository
$ git clone https://github.com/kennydo/python-language-server.git

# Enter the repository
$ cd python-language-server

# Install the vscode-client extension
$ cd vscode-client
$ yarn install

# Run VSCode, which is configured to use pyls.
# Pass in the full hostname (or IP address) of
# the remote host via environment variable.
# Pass in the same port number as the port argument to `pyls`.
$ PYLS_HOST=remote.hostname.here PYLS_PORT=2525 yarn run vscode ...

Then to debug, click View -> Output and in the dropdown will be pyls. To refresh VSCode, press Cmd + r. If VS Code loses connectivity to the pyls server stops, you will have to refresh VSCode.

The VSCode started by that yarn command has a separate user data directory than normal. If you want to copy your settings over, run this:

# From the vscode-client directory
cp ~/Library/Application\ Support/Code/User/settings.json .vscode-dev/user-data/User/settings.json

This project is made available under the MIT License.