T-appendix.jupyter-installation-and-setup.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": false
   },
   "source": [
    "# Jupyterlab 的安装与配置"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 下载并安装 Anaconda"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Anaconda](https://www.anaconda.com) 是目前最方便的 Python 发行版，搭载了很多我们终将必用的软件包，除了 Python 之外，还有 [R 语言](https://www.r-project.org/)，还包括 [Pandoc](https://pandoc.org/)， [NumPy](http://www.numpy.org/)，[SciPy](https://www.scipy.org/)，[Matplotlib](https://matplotlib.org/)…… 等等。\n",
    "\n",
    "无论是图形化界面安装，还是命令行界面安装，建议都把 Anaconda 安装在本地用户目录内，`~/`。请下载并安装 Python 3.x 的版本。\n",
    "\n",
    "图形化界面安装的教程，官方的很详细，各个操作平台的说明都有：\n",
    "\n",
    "> https://docs.anaconda.com/anaconda/install/\n",
    "\n",
    "在 MacOS 的 Terminal 命令行下，可以直接下载并安装：\n",
    "\n",
    "```bash\n",
    "cd ~/Downloads/\n",
    "wget https://repo.anaconda.com/archive/Anaconda3-2018.12-MacOSX-x86_64.sh\n",
    "./Anaconda3-2018.12-MacOSX-x86_64.sh\n",
    "```\n",
    "\n",
    "安装到最后一步，会问你是否要安装微软出品的 [Visual Studio Code](https://code.visualstudio.com)，选择 `yes` —— 反正以后你的电脑上会不止一个文本编辑器…… 以后你可能还会安装的文本编辑器包括 [SublimeText](https://www.sublimetext.com), [Atom](https://atom.io) 等等。\n",
    "\n",
    "安装完毕之后，打开 Terminal(Windows系统需要打开之前安装的Anaconda Prompt输入)，继续安装几个组件：\n",
    "\n",
    "```bash\n",
    "conda update conda\n",
    "conda update anaconda\n",
    "conda install -c conda-forge nodejs\n",
    "conda install -c conda-forge jupyterlab # 这是用来升级 jupyter lab 到最新版的方法\n",
    "\n",
    "```\n",
    "\n",
    "安装完毕之后，可以看看各个你将要用到的可执行命令都在什么地方，用 `which` 命令：\n",
    "\n",
    "```bash\n",
    "which python\n",
    "python --version\n",
    "which node\n",
    "node -v\n",
    "which jupyter\n",
    "jupyter lab --version\n",
    "jupyter notebook --version\n",
    "which pip\n",
    "pip --version\n",
    "```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 第一次启动 Jupyter lab"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "打开 Terminal，`cd` 到你想打开 Jupyter lab 的目录（就是你保存 `ipynb` 文件的地方，以便在 Jupyter lab 中打开、浏览、编辑 `ipynb` 文件），在这里以用户根目录为例 `~/`：\n",
    "\n",
    "```bash\n",
    "cd ~\n",
    "jupyter lab\n",
    "```\n",
    "此时的 Terminal 窗口不能关闭，否则 Jupyter lab 就停止运行了 —— 就将它放在那里。\n",
    "\n",
    "随后会有个浏览器打开，指向 [http://localhost:8888/lab?](http://localhost:8888/lab?) —— 你就看到 Jupyter lab 的操作界面了。\n",
    "\n",
    "目前，Jupyter lab 和 Jupyter notebook 是并存的，虽然前者是后者的下一步替代者。如果你依然习惯于使用 Jupyter notebook，那么，在浏览器中指向 [http://localhost:8888/tree?](http://localhost:8888/tree?) 看到的就是 Jupyter notebook."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 配置 Jupyter lab"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "打开 Terminal，输入以下命令：\n",
    "\n",
    "```bash\n",
    "jupyter lab --generate-config\n",
    "```\n",
    "\n",
    "这会在 `~/.jupyter/` 目录下生成一个 `jupyter_notebook_config.py` 文件。\n",
    "\n",
    "```bash\n",
    "cd ~/.jupyter\n",
    "code jupyter_notebook_config.py\n",
    "```\n",
    "上面的 code 命令，需要你已经安装 Visual Studio Code，并且在已经在其中设置了 `Install 'code' command in PATH`。参见附录 [Visual Studio Code 的安装与配置](T-appendix.editor.vscode.ipynb)\n",
    "\n",
    "事实上，你可以用你喜欢的任何编辑器打开 `~/.jupyter/jupyter_notebook_config.py` 文件。\n",
    "\n",
    "文件内容很长，有空可以仔细看。可以直接将以下内容拷贝粘贴到文件底部，根据需求修改：\n",
    "\n",
    "```json\n",
    "#c.NotebookApp.token = ''\n",
    "#c.NotebookApp.open_browser = False\n",
    "#c.NotebookApp.notebook_dir = '~/'\n",
    "#c.NotebookApp.default_url = '/tree'\n",
    "```\n",
    "\n",
    "逐条解释一下：\n",
    "\n",
    "> `c.NotebookApp.token = ''`\n",
    "\n",
    "每次打开 Jupter，它都会给你生成一个新的 Token —— 这是安全策略。但是，如果你只是在自己的电脑上使用，那么，这就给你制造了麻烦，因为若是你想同时用另外一个浏览器打开它，那你就需要从 Terminal 里拷贝那个 Token 出来。所以，你可以在配置文件里直接把它设置为空。\n",
    "\n",
    "> `c.NotebookApp.open_browser = False`\n",
    "\n",
    "每次你执行 `jupyter lab` 或者 `jupyter notebook` 命令的时候，它都会使用系统默认浏览器。\n",
    "\n",
    "每个人的习惯不一样。比如我，会想到用一个平时不怎么用的浏览器专门用在 Jupyter 上，这样会防止自己在关闭其它网页的时候不小心把 Jupyter 关掉…… 那我就会把这项设定为 `False`。\n",
    "\n",
    "> `c.NotebookApp.notebook_dir = '~/'`\n",
    "\n",
    "在 Terminal 中执行 `jupyter` 命令的时候，它默认是在你当前所在的工作目录打开 `jupyter`，这同样是出于安全考虑。但是，如果你只是在自己的电脑上使用，且只有自己在使用，那么莫不如直接把它设置成 `~/`，即，你的用户根目录，这样会很方便地访问各种地方的文件……\n",
    "\n",
    "> `c.NotebookApp.default_url = '/tree'`\n",
    "\n",
    "这一项留给那些依然习惯于使用 jupter notebook 的人，这样设置之后，即便是输入 `jupyter lab` 命令，打开的还是 jupyter notebook。\n",
    "\n",
    "在 Terminal 里常用的与 Jupyter 有关的命令有：\n",
    "\n",
    "```bash\n",
    "jupyter lab\n",
    "jupyter lab --version\n",
    "conda install -c conda-forge jupyterlab # 这是用来升级 jupyter lab 到最新版的方法\n",
    "jupyter notebook list                   # 查看正在运行的 jupyter lab/notebook\n",
    "jupyter notebook stop                   # 停止 jupyter lab/notebook 服务\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 将 Jupyter lab 配置成系统服务"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "如果，你厌烦每次都要跑到 Terminal 里启动 Jupyter lab，可以把它配置成系统服务，每次开机启动它就自动运行。而你需要做的只不过是直接从浏览器中访问 [http://localhost:8888/](http://localhost:8888/)。\n",
    "\n",
    "```bash\n",
    "code ~/Library/LaunchAgents/com.jupyter.lab.plist\n",
    "```\n",
    "\n",
    "这条命令会让 Visual Studio Code 创建 `~/Library/LaunchAgents/com.jupyter.lab.plist`文件并打开。\n",
    "\n",
    "在其中拷贝粘贴以下内容，注意，要把其中的 `your_username` 修改为你的用户名：\n",
    "\n",
    "``` xml\n",
    "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n",
    "<!DOCTYPE plist PUBLIC \"-//Apple//DTD PLIST 1.0//EN\" \"http://www.apple.com/DTDs/PropertyList-1.0.dtd\">\n",
    "<plist version=\"1.0\">\n",
    "<dict>\n",
    "\t<key>Label</key>\n",
    "\t<string>local.job</string>\n",
    "\t<key>ProgramArguments</key>\n",
    "\t<array>\n",
    "\t\t<string>/Users/your_username/anaconda3/bin/jupyter</string>\n",
    "\t\t<string>lab</string>\n",
    "\t\t<string>--no-browser</string>\n",
    "\t\t<string>--notebook-dir=/Users/your_username/</string>\n",
    "\t</array>\n",
    "\t<key>RunAtLoad</key>\n",
    "\t<true/>\n",
    "\t<key>StandardErrorPath</key>\n",
    "\t<string>/tmp/local.job.err</string>\n",
    "\t<key>StandardOutPath</key>\n",
    "\t<string>/tmp/local.job.out</string>\n",
    "</dict>\n",
    "</plist>\n",
    "```\n",
    "如果之前在`jupyter_notebook_config.py` 文件里已经设置过\n",
    "\n",
    "```json\n",
    "c.NotebookApp.open_browser = False\n",
    "c.NotebookApp.notebook_dir = '~/'\n",
    "```\n",
    "那么这两行就可以不要了：\n",
    "\n",
    "```xml\n",
    "\t\t<string>--no-browser</string>\n",
    "\t\t<string>--notebook-dir=/Users/your_username/</string>\n",
    "```\n",
    "\n",
    "\n",
    "而后在 Terminal 里执行：\n",
    "\n",
    "```bash\n",
    "launchctl load ~/Library/LaunchAgents/com.jupyter.lab.plist\n",
    "```\n",
    "\n",
    "如果你想重新启动这个服务，那么执行：\n",
    "\n",
    "```bash\n",
    "launchctl unload ~/Library/LaunchAgents/com.jupyter.lab.plist\n",
    "launchctl load ~/Library/LaunchAgents/com.jupyter.lab.plist\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 关于 Jupyter lab themes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对中文用户来说，Jupyter 的默认字号有点过小，阅读起来不是很舒适。但，最佳的方案不是去寻找合适的 themes，而是直接使用支持 [Stylus](https://github.com/openstyles/stylus) 这类终端 CSS 定制插件的浏览器，Chrome/Firefox/Opera 都支持 Stylus 插件。\n",
    "\n",
    "我用的 Stylus 定制 CSS 是这样的：\n",
    "\n",
    "```css\n",
    "a {color: #2456A4 !important;}\n",
    "strong {color:#6392BF;}\n",
    "em {color: #A9312A; font-style: normal !important;}\n",
    "table {font-size: 90% !important;}\n",
    "\n",
    "#jp-main-dock-panel {background-color: #f9f9f9;}\n",
    ".jp-RenderedHTMLCommon {font-family: \"Yuanti SC\"; font-size: 100%;}\n",
    ".jp-Notebook {background-color: #fbfafa;}\n",
    ".CodeMirror, .jp-RenderedHTMLCommon pre {font-size: 90%;}\n",
    ".jp-RenderedHTMLCommon pre {\n",
    "    padding: 10px 25px;\n",
    "    background-color: #fafafa;\n",
    "    border-left: 4px solid #dadada;\n",
    "    border-radius: 10px;\n",
    "}\n",
    "\n",
    ".jp-RenderedHTMLCommon pre code {\n",
    "    background-color: #fafafa;\n",
    "}\n",
    "\n",
    ".jp-RenderedHTMLCommon h1 code,\n",
    ".jp-RenderedHTMLCommon h2 code,\n",
    ".jp-RenderedHTMLCommon h3 code,\n",
    ".jp-RenderedHTMLCommon h4 code,\n",
    ".jp-RenderedHTMLCommon p code, \n",
    ".jp-RenderedHTMLCommon li code,\n",
    ".jp-RenderedHTMLCommon blockquote p code, \n",
    ".jp-RenderedHTMLCommon blockquote li code,\n",
    ".jp-RenderedHTMLCommon td code {\n",
    "    background-color: #f6f6f6;\n",
    "    font-size: 90%;\n",
    "    color:#2e2e2e;\n",
    "    padding: 4px 4px;\n",
    "    margin: 0 8px;\n",
    "    box-shadow: 0px 1px 2px 0px rgba(0,0,0,0.2);\n",
    "    border-radius: 4px;\n",
    "}\n",
    "```\n",
    "\n",
    "这样就相当于我把 JupyterLab Light 这个 Theme 稍微 Tweak 了一下。\n",
    "\n",
    "另，我写的内容里，为了重点突出，特别定制了 `strong` 和 `em` 两个元素的显示，让它们以不同的颜色展示；又因为中文并不适合斜体展示，所以，把 `em` 的 `font-style` 设定为 `normal`……"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 安装插件"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Jupyter notebook 经过很多年的发展，现在有很多扩展插件，但也有其中一些并不兼容最新的 Jupyter lab。不过，刚开始的时候用不着那么多插件，你只用其中的两个就足够开始了：\n",
    "\n",
    "> * [@jupyterlab/toc](https://github.com/jupyterlab/jupyterlab-toc)\n",
    "> * [ryantam626/jupyterlab_sublime](https://github.com/ryantam626/jupyterlab_sublime)\n",
    "\n",
    "首先在用快捷键 `⌘ ,` 打开 Jupter lab 的 Advanced Settings，在 Extension Manager 中，添加 User Overrides：\n",
    "\n",
    "```json\n",
    "{\n",
    "    \"enabled\": true\n",
    "}\n",
    "```\n",
    "\n",
    "而后在 Terminal 执行以下命令安装插件：\n",
    "\n",
    "```bash \n",
    "jupyter labextension install @jupyterlab/toc\n",
    "jupyter labextension install @ryantam626/jupyterlab_sublime\n",
    "jupyter lab build\n",
    "```\n",
    "\n",
    "toc 插件，自动将 ipynb 文件中的标题转换成目录。\n",
    "\n",
    "![](https://github.com/jupyterlab/jupyterlab-toc/raw/master/toc.gif)\n",
    "\n",
    "jupyterlab_sublime 则可以让你在 Jupyter lab 的 cell 中，使用跟 SublimeText 一样的快捷键，比如 `⌘ D` 能够多选其它与当前选中内容一样的内容；比如 `⌘` 加鼠标点击，可以生成多个可编辑点……\n",
    "\n",
    "![](http://blog.rtwilson.com/wp-content/uploads/2016/03/IPyNbSublime.gif)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 常用快捷键"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "以下是 MacOS 下 Jupyter lab 最常用的快捷键。快捷键在两种模式下执行，进入编辑模式用 `⏎`，回到命令模式用 `⎋`（ESC）。\n",
    "\n",
    "另外，代码编辑过程中需要安装 Jupyterlab 插件 [@ryantam626/jupyterlab_sublime](https://github.com/ryantam626/jupyterlab_sublime) 之后才能使用“多行同时编辑功能”。\n",
    "\n",
    "\n",
    "| 快捷键                                  | 说明                                                         | 模式   |\n",
    "| --------------------------------------- | ------------------------------------------------------------ | ------ |\n",
    "| `ESC`                                   | 从编辑模式回到命令模式                                       | 命令   |\n",
    "| `A`                                     | 在当前 Cell 之前插入一个 Cell                                |        |\n",
    "| `B`                                     | 在当前 Cell 之后插入一个 Cell                                |        |\n",
    "| `D`, `D`                                   | 连续按两次 `d` 键，删除当前 Cell                                                |        |\n",
    "| `Y`                                     | 将当前 Cell 设置为 Code Cell                                 |        |\n",
    "| `M`                                     | 将当前 Cell 设置为 Markdown Cell                             |        |\n",
    "| `^ ⇧ - `                                | 将当前 Cell 拆分为两个                                       | 编辑    |\n",
    "| `⇧ M`                                   | 合并选中的 Cells                                             |        |\n",
    "| `⇧ J` or `⇧ ↓`                          | 连续向下选中 Cells                                           |        |\n",
    "| `⇧ K` or `⇧ ↑`                          | 连续向上选中 Cells                                           |        |\n",
    "| `⇧ ⏎` or `^ ⏎`                          | 运行当前 Cell 中的代码                                       |        |\n",
    "| `⇧ L`                                   | 显示/隐藏代码行号                                            |        |\n",
    "| `⏎`                                     | 当前 Cell 进入编辑模式                                       | 编辑   |\n",
    "| `⇥`                                   | 自动补全代码                               |        |\n",
    "| `⇧ ⇥`                                   | 呼出当前光标下词汇的 Docstring                               |        |\n",
    "| `⌘ D`                                   | Sublime Keymap: 选中下一个相同字符串                         |        |\n",
    "| `⇧ ⌘ L`                                 | Sublime Keymap: 在选中的行内启动多行同时编辑                 |        |\n",
    "| `⌘ + Mouse Click`                       | 生成下一个可同时编辑的光标点                                 |        |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 增加一些必要的快捷键"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在 Settings > Keyboard Shortcuts 中，可以设定一些常用但系统并未给出的快捷键：\n",
    "\n",
    "```json\n",
    "{\n",
    "\t\"notebook:move-cells-down-down\": {\n",
    "\t\t\"command\": \"notebook:move-cell-down\",\n",
    "\t\t\"keys\": [\n",
    "\t\t\"Alt J\"\n",
    "\t\t],\n",
    "\t\t\"selector\": \".jp-Notebook:focus\",\n",
    "\t\t\"title\": \"Move Cells Down\",\n",
    "\t\t\"category\": \"Notebook Cell Operations\"\n",
    "\t},\n",
    "\t\"notebook:move-cells-down-up\": {\n",
    "\t\t\"command\": \"notebook:move-cell-up\",\n",
    "\t\t\"keys\": [\n",
    "\t\t\"Alt K\"\n",
    "\t\t],\n",
    "\t\t\"selector\": \".jp-Notebook:focus\",\n",
    "\t\t\"title\": \"Move Cells Down\",\n",
    "\t\t\"category\": \"Notebook Cell Operations\"\n",
    "\t},\n",
    "\t\"notebook:enable-output-scrolling\": {\n",
    "\t\t\"command\": \"notebook:enable-output-scrolling\",\n",
    "\t\t\"keys\": [\n",
    "\t\t\"S\"\n",
    "\t\t],\n",
    "\t\t\"selector\": \".jp-Notebook:focus\",\n",
    "\t\t\"title\": \"Enable output scrolling\",\n",
    "\t\t\"category\": \"Notebook Cell Operations\"\n",
    "\t},\n",
    "\t\"notebook:disable-output-scrolling\": {\n",
    "\t\t\"command\": \"notebook:disable-output-scrolling\",\n",
    "\t\t\"keys\": [\n",
    "\t\t\"Alt S\"\n",
    "\t\t],\n",
    "\t\t\"selector\": \".jp-Notebook:focus\",\n",
    "\t\t\"title\": \"Enable output scrolling\",\n",
    "\t\t\"category\": \"Notebook Cell Operations\"\n",
    "\t}\n",
    "}\n",
    "```\n",
    "这样就添加了 4 个快捷键：\n",
    "> * `⌥ J`: Move selected cells up\n",
    "> * `⌥ K`: Move selected cells down\n",
    "> * `S`: Enable output scrolling\n",
    "> * `⌥ S`: Disable output scrolling\n",
    "\n",
    "比如 Move Selected cells up：\n",
    "\n",
    "![](https://user-images.githubusercontent.com/86304/37438938-b5bc4994-27b2-11e8-8f58-184a58b33ba4.gif)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 输出所有变量内容"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "默认情况下，Code Cell 只输出最后一个可以被 evaluate 的值，用 `_` 代表之前刚刚被 evaluate 的值。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 2, 3]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[1, 2, 3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 2, 3]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "_ # 执行完上面的 Cell，试试这个 Cell; 而后执行完下面的 Cell 之后再重新执行一次当前这个 Cell"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{1, 2, 3}"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(1, 2, 3)\n",
    "{1, 2, 3}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "于是，为了显示最近 evaluate 的多个值，我们总是不得不使用很多的 `print()`……\n",
    "\n",
    "如果觉得这事儿比较烦的话，可以在 Cell 最上面写上：\n",
    "\n",
    "```python\n",
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "```\n",
    "\n",
    "如果还想更省事儿一点，就把这个设置写入配置文件：\n",
    "\n",
    "```python\n",
    "c.InteractiveShell.ast_node_interactivity = \"all\"\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 2, 3)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "{1, 2, 3}"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "\n",
    "(1, 2, 3)\n",
    "{1, 2, 3}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 魔法函数"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在 Code Cell 里，可以运行一些“魔法函数”（Magic Functions），这是秉承了 IPython 的特性。绝大多数在 IPython 里能够使用的魔法函数在 Jupyterlab 里都可以直接使用。完整的 IPython 魔法函数请参照：\n",
    "\n",
    "> https://ipython.readthedocs.io/en/stable/interactive/magics.html\n",
    "\n",
    "Jupyterlab 里较为常用的魔法函数整理如下：\n",
    "\n",
    "| 魔法函数             | 说明                                                         |\n",
    "| -------------------- | ------------------------------------------------------------ |\n",
    "| `%lsmagic`           | 列出所有可被使用的 Jupyter lab 魔法函数                      |\n",
    "| `%run`               | 在 Cell 中运行 `.py` 文件：`%run file_name`                  |\n",
    "| `%who`               | 列出所有当前 Global Scope 中的变量；类似的还有：`%who df`，`%whos` |\n",
    "| `%env`               | 列出当前的环境变量                                           |\n",
    "| `%load`              | 将其他文件内容导入 Cell，`%load source`，`source` 可以是文件名，也可以是 URL。 |\n",
    "| `%time`              | 返回 Cell 内代码执行的时间，相关的还有 `%timeit`             |\n",
    "| `%writefile`         | 把 Cell 的内容写入文件，`%write file_name`；%write -a file_name，`-a` 是追加 |\n",
    "| `%matplotlib inline` | 行内展示 matplotlib 的结果                                   |\n",
    "| `%%bash`             | 运行随后的 shell 命令，比如 %%bash ls；与之类似的还有 `%%HTML`， `%%python2`， `%%python3`， `%%ruby`， `%%perl`……                      |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": true
   },
   "source": [
    "## 桌面版 Jupyter App"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Nteract"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "支持各个操作系统，很好看、很好用。有一个小缺点是，不支持 `input()` 函数的调用。\n",
    "\n",
    "> https://nteract.io/desktop\n",
    "\n",
    "![https://cloud.githubusercontent.com/assets/836375/18421299/d95ad398-783b-11e6-8b23-d54cf7caad1e.png](https://cloud.githubusercontent.com/assets/836375/18421299/d95ad398-783b-11e6-8b23-d54cf7caad1e.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": true
   },
   "source": [
    "### Pineapple"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "只支持 MacOS，也很好用 —— 缺点就是很难看……\n",
    "\n",
    "> https://nwhitehead.github.io/pineapple/\n",
    "\n",
    "![https://nwhitehead.github.io/pineapple/images/sshots.png](https://nwhitehead.github.io/pineapple/images/sshots.png)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  },
  "toc-autonumbering": true
 },
 "nbformat": 4,
 "nbformat_minor": 2
}