python-language-server


Ⅰ. 插件描述

Microsoft Language Server for Python

Ⅱ. 基本信息

创建日期:  2018-08-23
使用用户:  1
Github星:  585
插件作者:  Microsoft

Ⅲ. 安装方法

使用Vundle管理器安装

在你的.vimrc下添加:
Plugin 'microsoft/python-language-server'
… 然后在Vim中运行以下命令:
:source %
:PluginInstall

对于Vundle版本 < 0.10.2,请用上面的Bundle替换Plugin。

使用NeoBundle管理器安装

在你的.vimrc下添加:
NeoBundle 'microsoft/python-language-server'
… 然后在Vim中运行以下命令:
:source %
:NeoBundleInstall

使用VimPlug管理器安装

在你的.vimrc下添加:
Plug 'microsoft/python-language-server'
… 然后在Vim中运行以下命令:
:source %
:PlugInstall

使用Pathogen管理器安装

在终端中运行以下命令:
cd ~/.vim/bundle
git clone https://github.com/microsoft/python-language-server

Ⅳ. 文档说明

# Microsoft Python Language Server

Microsoft Python Language Server implements the Language Server Protocol.

Its primary clients are the Python extension for VS Code and Python Tools for Visual Studio.

Feel free to file issues or ask questions on our issue tracker, and we welcome code contributions.

Build/contributing instructions

See CONTRIBUTING.md

Troubleshooting and known issues

See TROUBLESHOOTING.md.

Linting options (diagnostics)

The language server implements diagnostics (or linting), which runs on user code.
The following diagnostics are supported:

| Code | Description |
| - | - |
| inherit-non-class | Attempted to inherit something that is not a class. |
| too-many-function-arguments | Too many arguments have been provided to a function call. |
| too-many-positional-arguments-before-star | Too many arguments have been provided before a starred argument. |
| no-cls-argument | First parameter in a class method must be cls |
| no-method-argument | Method has no arguments
| no-self-argument | First parameter in a method must be self
| parameter-already-specified | A argument with this name has already been specified. |
| parameter-missing | A required positional argument is missing. |
| positional-argument-after-keyword | A positional argument has been provided after a keyword argument. |
| positional-only-named | A positional-only argument (3.8+) has been named in a function call. |
| return-in-init | Encountered an explicit return in __init__ function. |
| typing-generic-arguments | An error occurred while constructing Generic. |
| typing-newtype-arguments | An error occurred while constructing NewType. |
| typing-typevar-arguments | An error occurred while constructing TypeVar. |
| unknown-parameter-name | The keyword argument name provided is unknown. |
| unresolved-import | An import cannot be resolved, and may be missing. |
| undefined-variable | A variable has been used that has not yet been defined. |
| variable-not-defined-globally | A variable is not defined in the global scope. |
| variable-not-defined-nonlocal | A variable is not defined in non-local scopes. |

A full list can be seen in the source code.

Linting can be controlled via the user configuration. In VS Code, this is settings.json, but other
clients would send this via workspace/didChangeConfiguration.

If python.linting.enabled is set to false in the user configuration, then no diagnostics
will be collected other than syntax errors and unresolved imports.

To control the visibility and severity of the diagnotics, there are a number of lists
that can be set in the user configuration which make use of each diagnostic's error code.

| Setting | Description |
| - | - |
| python.analysis.errors | Diagnostics which should be shown as errors. |
| python.analysis.warnings | Diagnostics which should be shown as warnings. |
| python.analysis.information | Diagnostics which should be shown as informational. |
| python.analysis.disabled | Diagnotics which should not be shown at all. |

An example of a user configuration which sets these options:

{
    "python.analysis.errors": ["undefined-variable"],
    "python.analysis.warnings": ["unknown-parameter-name"],
    "python.analysis.information": ["unresolved-import"],
    "python.analysis.disabled": ["too-many-function-arguments", "parameter-missing"],
}

Linting can also be controlled on an invidual line basis with a generalized #noqa. Lines with #noqa will have their diagnostic output suppressed.

An example usage:

from python import language_server  # noqa will suppress the linting message for this line

Cache location

During analysis language server produces Python code from compiled modules and builtins which is similar to Python module stubs.
It may also produce database files holding module analysis for faster retrieval later. Cache location is at

Windows

"%LOCALAPPDATA%\Microsoft\Python Language Server" (which is Environment.SpecialFolder.LocalApplicationData). Typically "C:\Users\\%USER_NAME%\AppData\Local\Microsoft\Python Language Server"

Linux

"$XDG_CACHE_HOME/Microsoft/Python Language Server", or if XDG_CACHE_HOME is not set, "$HOME/.cache/Microsoft/Python Language Server"

macOS

"$HOME/Library/Caches/Microsoft/Python Language Server"

添加新评论