更好地使用Atom支持基于Jupyter的Python开发

有关于使用Atom进行Python开发的网上资料比较少，最近发现使用Atom结合Hydrogen插件进行Python开发，尤其是数据挖掘相关的工作，整体体验要好于Vscode，Vscode虽然说也有连接Jupyter的工具，但是交互式的开发Hydrogen体验更好。

这里放了个动图来展示一下Hydrogen的强大

插件安装

Python

• Hydrogen
• atom-ide-ui
• ide-python

这里要注意，本地的pip需要 安装 python-language-server[all]，在ide-python的readme中有详细说明

远程连接

• Remote FTP

美化

• simplified-chinese-menu（汉化补丁）

• file-icons（文件夹图标）

• bracket-colorizer（彩虹括号，找了好久，确定就是必须配合暗色主题）

• atom-bracket-highlight（括号高亮）

• atom-clock（加个时钟在右下角）

• highlight-selected（高亮选择）

• minimap（类似sublime的右侧map栏）

• minimap-highlight-selected（选择代码后，map上也高亮，方便定位代码）

插件配置

Remote FTP

这里先讲一下我的需求，我是需要利用其连接公司服务器上的内容，但是公司服务器是需要跳板机的，所以我需要通过跳板机才能访问，因此配置上会有些复杂

{
    "protocol": "sftp",
    "host": "跳板机域名", // string - Hostname or IP address of the server. Default: 'localhost'
    "port": 跳板机端口, // integer - Port number of the server. Default: 22
    "user": "用户名", // string - Username for authentication. Default: (none)
    "pass": "如果用密钥这里就不用填", // string - Password for password-based user authentication. Default: (none)
    "promptForPass": false, // boolean - Set to true for enable password/passphrase dialog. This will prevent from using cleartext password/passphrase in this config. Default: false
    "remote": "实际的服务器目录，例如：/服务器域名/用户名/目录", // try to use absolute paths starting with /
    "agent": "", // string - Path to ssh-agent's UNIX socket for ssh-agent-based user authentication. Linux/Mac users can set "env" as a value to use env SSH_AUTH_SOCK variable. Windows users: set to 'pageant' for authenticating with Pageant or (actual) path to a cygwin "UNIX socket." Default: (none)
    "privatekey": "本地私钥path", // string - Absolute path to the private key file (in OpenSSH format). Default: (none)
    "passphrase": "", // string - For an encrypted private key, this is the passphrase used to decrypt it. Default: (none)
    "hosthash": "", // string - 'md5' or 'sha1'. The host's key is hashed using this method and passed to the hostVerifier function. Default: (none)
    "ignorehost": true,
    "connTimeout": 10000, // integer - How long (in milliseconds) to wait for the SSH handshake to complete. Default: 10000
    "keepalive": 10000, // integer - How often (in milliseconds) to send SSH-level keepalive packets to the server (in a similar way as OpenSSH's ServerAliveInterval config option). Set to 0 to disable. Default: 10000
    "keyboardInteractive": 如果要用动态令牌，这里就要填true, // boolean - Set to true for enable verifyCode dialog. Keyboard interaction authentication mechanism. For example using Google Authentication (Multi factor)
    "keyboardInteractiveForPass": false, // boolean - Set to true for enable keyboard interaction and use pass options for password. No open dialog.
    "remoteCommand": "",
    "remoteShell": "",
    "watch":[],
    "watchTimeout":500, // integer - The duration ( in milliseconds ) from when the file was last changed for the upload to begin.
}

ide-python

需要配置一下Python Executable

填写你的python路径，比如使用的是conda虚拟环境，就这样写

/xxx/anaconda3/envs/xxx/bin/python

Hydrogen

连接本地Kernel

首先需要在上面填写的路径下的python环境中安装ipykernel

python -m ipykernel install --user --name py37

然后用atom打开一个py文件，输入

# %%
print('hello atom')

# %%
print('每一个# %%代表一个新的cell')

然后再第2行末尾按ctrl+enter就会自动弹出来让你选择环境的弹窗，选择刚刚新建的环境即可

然后在Hydrogen里面通过使用# %%来分割每一个cell，在Mac中使用option+shift+enter组合键来实现运行当前整个ceil，使用command+enter实现运行当前行，使用shift+enter实现运行当前行并跳转下一行，具体可参考官方文档

连接远程Kernel

连接远程的Jupyter只需要配置一下Hydrogen设置里面的Kernel Gateways，填上如下内容即可

[{"name": "Remote server",
  "options": {
    "baseUrl": "jupyter url",
    "token": "jupyter token"
  }
}]

然后点击Connect to Remote Kernel即可