355 lines
9.2 KiB
Markdown
355 lines
9.2 KiB
Markdown
|
# 原创
|
|||
|
|
： 云服务器搭建神器JupyterLab（多图）
|
|||
|
|
|
|||
|
|
# 云服务器搭建神器JupyterLab（多图）
|
|||
|
|
|
|||
|
|
# 云服务器搭建神器JupyterLab（多图）
|
|||
|
|
|
|||
|
|
>
|
|||
|
|
[`JupyterLab`](https://github.com/jupyterlab/jupyterlab)是一个交互式的开发环境，其用于应对包含着`notebook`、代码以及数据的工作场景。
|
|||
|
|
|
|||
|
|
|
|||
|
|
## 1、前言
|
|||
|
|
|
|||
|
|
**如果说`vim`是编辑器之神，那么`JupyterLab`就是笔记本之神。**
|
|||
|
|
|
|||
|
|
从2017年开始我注意到这一神奇的`IDE`笔记本，第一眼见到它，就觉得它真的太强大了，作为一个交互式的`Python`开发工具，其实也不算开发工具，准确的来说，它是一个演示代码的科学数据工具，支持`markdown`预览，支持[Draw](https://www.draw.io/)扩展，支持丰富的文件格式和多种开发语言，拥有众多[插件](https://github.com/jupyterlab?utf8=%E2%9C%93&q=extension&type=&language=)诸于`GitHub`， `Google-Dirve`， `Git`和`TOC`，更重要的是，它是把浏览器当作开发工具，十分有创意。[`JupyterLab`](https://github.com/jupyterlab/jupyterlab)的开发者众多，`GitHub`上面的`isuue`也十分活跃，众多的大牛开发者也纷纷加入到`JupyterLab`的阵营，为`JupyterLab`的发展作贡献，`JupyterLab`得到迅速发展。
|
|||
|
|
|
|||
|
|
好的工具当然是要好好利用了，回归正题，怎样通过云服务器搭建一个可远程使用的`JupyterLab`?
|
|||
|
|
|
|||
|
|
## 2、购买云服务器`ECS`
|
|||
|
|
|
|||
|
|
这里我就直接买了阿里的学生优惠的`ECS`，9.9元/月，价格算是良心了，这里可以领一下优惠券：[阿里云限时礼包](https://promotion.aliyun.com/ntms/yunparter/invite.html?userCode=vya2etaw)。亚马逊还有每个账户免费使用一年云服务器的机会，也可以用一用。购买之后创建实例，启动云服务器，我这里用的是`Ubuntu 16.04`镜像，记住你设置的`root`密码。如果没有特殊要求，可以直接使用`root`用户进行下面操作，不必新建用户。
|
|||
|
|
|
|||
|
|
## 3、登录ECS并安装必要软件
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install ssh
|
|||
|
|
ssh root@公网ip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
这个ip是公网ip，在你购买的服务器运营商的实例列表里，如果不出意外就可以登录ECS了。
|
|||
|
|
|
|||
|
|
### 3.1 先做好准备工作
|
|||
|
|
|
|||
|
|
比喻你添加ppa源的时候出现这种情况
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
root@iZwz9huxtbd86xp91s3j16Z:~# sudo add-apt-repository ppa:chronitis/jupyter
|
|||
|
|
sudo: add-apt-repository: command not found
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
解决办法
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install software-properties-common
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
### 3.2 关于`pip`的问题
|
|||
|
|
|
|||
|
|
#### 安装pip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install python-pip python-dev build-essential
|
|||
|
|
sudo pip install --upgrade pip
|
|||
|
|
sudo pip install --upgrade virtualenv
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install python-setuptools python-dev build-essential
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo easy_install pip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo pip install --upgrade virtualenv
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install python3-pip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get install python-pip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
#### 升级`pip`
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo pip3 install --upgrade pip
|
|||
|
|
sudo pip2 install --upgrade pip
|
|||
|
|
sudo pip install --upgrade pip
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
#### 更换`pip`源加速下载
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
cat > ~/.pip/pip.conf
|
|||
|
|
[global]
|
|||
|
|
index-url = https://pypi.tuna.tsinghua.edu.cn/simple
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
`npm`和`yarn`也可以换源，更简单，自行`google`。
|
|||
|
|
|
|||
|
|
`pip`国内的一些镜像包,换源之后出现`python2`版本过低的情况导致以前的包下载不了，那就直接将文件夹`~/.pip/pip.conf`删除就可以恢复原来的源。
|
|||
|
|
|
|||
|
|
#### `pip`指向问题
|
|||
|
|
|
|||
|
|
有时候会出现`pip,pip2,pip3`都TM指向`python2`，这个之后就需要改一下这这三个文件。
|
|||
|
|
|
|||
|
|
编辑这三个文件，将第一行注释分别改为python\python2\python3
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
~ $which pip
|
|||
|
|
/usr/local/bin/pip
|
|||
|
|
21:36 alien@alien-Inspiron-3443:
|
|||
|
|
~ $which pip2
|
|||
|
|
/usr/local/bin/pip2
|
|||
|
|
21:36 alien@alien-Inspiron-3443:
|
|||
|
|
~ $which pip3
|
|||
|
|
/usr/local/bin/pip3
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
### 3.3 安装`yarn`和`nodejs`
|
|||
|
|
|
|||
|
|
#### 配置仓库
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
curl -sS https://dl.yarnpkg.com/debian/pubkey.gpg | sudo apt-key add -
|
|||
|
|
echo "deb https://dl.yarnpkg.com/debian/ stable main" | sudo tee /etc/apt/sources.list.d/yarn.list
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
#### 安装`yarn`
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo apt-get update
|
|||
|
|
sudo apt-get install yarn
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
注意这里`yarn`自动安装了`nodejs`，不过版本太低，安装`jupyterlab`的时候会出问题，而且使用`n`或者`nvm`安装的`nodejs`也有问题，总之`nodejs`要按照下面这种方式就没事，这个是官方的`bug`，官方`issue`也提到过，但是目前还未解决。
|
|||
|
|
|
|||
|
|
#### 安装`nodejs`
|
|||
|
|
|
|||
|
|
创建一个新文件，输入两行`deb`，结束之后`Ctrl+C`:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
cat > /etc/apt/sources.list.d/nodesource.list
|
|||
|
|
deb https://deb.nodesource.com/node_6.x xenial main
|
|||
|
|
deb-src https://deb.nodesource.com/node_6.x xenial main
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
导入公匙并安装`nodejs`，这个版本的`nodejs`对`jupyterlab`支持比较友好。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
curl -s https://deb.nodesource.com/gpgkey/nodesource.gpg.key | sudo apt-key add -
|
|||
|
|
sudo apt-get update
|
|||
|
|
apt-cache policy nodejs
|
|||
|
|
sudo apt-get install nodejs
|
|||
|
|
nodejs --version
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
#### 安装`ipython,matplotlib,scipy,pandas,numpy`
|
|||
|
|
|
|||
|
|
最好`python2`和`python3`都安装。
|
|||
|
|
|
|||
|
|
## 4、安装`JupyterLab`及其配置
|
|||
|
|
|
|||
|
|
记住使用`pip2`安装，如果你用`pip3`安装的话可能会出现版本兼容问题。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo pip2 install jupyterlab
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
生成密码
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter-notebook password
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
### 4.1 创建哈希密码
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
ipython
|
|||
|
|
from notebook.auth import passwd
|
|||
|
|
passwd()
|
|||
|
|
# 输入你自己设置登录JupyterLab界面的密码，
|
|||
|
|
# 然后就会生产下面这样的密码，将它记下来，待会儿用
|
|||
|
|
'sha1:b92f3fb7d848:a5d40ab2e26aa3b296ae1faa17aa34d3df351704'
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
### 4.2 修改`JupyterLab`配置文件
|
|||
|
|
|
|||
|
|
先生成一个配置文件，记下输出的配置文件地址
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter lab --generate-config
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
修改配置文件，找到下面这几行文件，注释掉并修改成这样。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
c.NotebookApp.allow_root = True
|
|||
|
|
c.NotebookApp.ip = '0.0.0.0'
|
|||
|
|
c.NotebookApp.notebook_dir = u'/root/JupyterLab'
|
|||
|
|
c.NotebookApp.open_browser = False
|
|||
|
|
c.NotebookApp.password = u'sha1:b92f3fb7d848:a5d40ab2e26aa3b296ae1faa17aa34d3df351704'
|
|||
|
|
c.NotebookApp.port = 8080
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
对应每行稍微解释一下
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
允许以root方式运行jupyterlab
|
|||
|
|
允许任意ip段访问
|
|||
|
|
设置jupyterlab页面的根目录
|
|||
|
|
默认运行时不启动浏览器，因为服务器默认只有终端嘛
|
|||
|
|
设置之前生产的哈希密码
|
|||
|
|
设置访问端口
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
到此，`JupyterLab`已经安装成功了。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter-lab --version
|
|||
|
|
0.33.12
|
|||
|
|
jupyter lab build
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
`jupyter lab build`时间有点久，如果没报错就成功了。但此时你还不能访问`JupyterLab`，还需要添加端口规则，也就是所谓的添加安全组。
|
|||
|
|
|
|||
|
|
## 5、添加安全组
|
|||
|
|
|
|||
|
|
<img alt="" src="https://raw.githubusercontent.com/ds19991999/githubimg/master/picgo20181102230911.png"/><br/> 去ECS控制台添加安全组，不然你无法通过本地浏览器访问`JupyterLab`，设置`8080`端口入方向。<br/> <img alt="" src="https://raw.githubusercontent.com/ds19991999/githubimg/master/picgo20181102231126.png"/>
|
|||
|
|
|
|||
|
|
## 远程访问`JupyterLab`
|
|||
|
|
|
|||
|
|
运行下面命令。
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
# nohup表示ssh终端断开后仍然运行
|
|||
|
|
# &表示允许后台运行
|
|||
|
|
nohup jupyter lab &
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
浏览器输入`公网ip:8080`，就可以访问你的`JupyterLab`了，第一次访问比较慢，耐心一点，如果最终还是无法访问，那么就是你的安全组配置错啦。<br/> <img alt="" src="https://raw.githubusercontent.com/ds19991999/githubimg/master/picgo20181102231458.png"/><br/> 输入密码就可以用了。
|
|||
|
|
|
|||
|
|
## 6、`JupyterLab`扩展安装
|
|||
|
|
|
|||
|
|
查看已经安装的扩展及其状态：
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter labextension list
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
比喻安装一个扩展`jupyterlab_html`，支持html预览:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter labextension install @mflevine/jupyterlab_html
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
卸载扩展:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter labextension uninstall @mflevine/jupyterlab_html
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
更新所有扩展:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter labextension update --all
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
下面以安装[GitHub](https://github.com/jupyterlab/jupyterlab-github)扩展为例。
|
|||
|
|
|
|||
|
|
### 安装`GitHub`扩展
|
|||
|
|
|
|||
|
|
先去`GitHub`生成一个`token`，记下`token`，待会儿配置要用。
|
|||
|
|
|
|||
|
|
下载安装扩展:
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter labextension install @jupyterlab/github
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
配置`token`
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
# 在之前的生成的config文件中添加
|
|||
|
|
c.GitHubConfig.access_token = '< YOUR_ACCESS_TOKEN >'
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
需要其他扩展的在[GitHub](https://github.com/)可以自行下载。<br/> <img alt="" src="https://raw.githubusercontent.com/ds19991999/githubimg/master/picgo20181102234041.png"/>
|
|||
|
|
|
|||
|
|
## 7、内核安装与卸载
|
|||
|
|
|
|||
|
|
安装Python内核
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
sudo pip2 install ipykernel
|
|||
|
|
sudo pip3 install ipykernel
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
如果`pip`指向正常的话就可以看到两个`Python`内核了。
|
|||
|
|
|
|||
|
|
查看已经安装的内核
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter kernelspec list
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
删除你不需要的内核
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
jupyter kernelspec remove <kernel_name>
|
|||
|
|
/root/JupyterLab
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
## 8、域名和`https`配置
|
|||
|
|
|
|||
|
|
域名和SSL配置可以参考这两篇文章：
|
|||
|
|
|
|||
|
|
其实为觉得没必要，毕竟是个人用的工具，没必要搞个域名，不过强迫症就另说了。具体效果是这样的:
|
|||
|
|
|
|||
|
|
## 9、结语
|
|||
|
|
|
|||
|
|
`JupyterLab`的搭建就是这么简单，好的工具就应该好好利用，支持做图，`markdown`，多标签，内部打开网页，`latex`，网页预览，这么好的工具我应该早点发现呀。最后，以秀图结束本文，多多指教！
|
|||
|
|
|
|||
|
|
>
|
|||
|
|
原创博文，转载注明请注明出处： [https:www.ds-vip.top](https:www.ds-vip.top)
|
|||
|
|
|