centos搭建Jupyter Notebook, 添加多語言支持(R,Ruby,Octave,JS,Java,C++).
https://foreti.me/2018/07/06/centos-jupyter-notebook/
常用命令
主要熟悉常用命令，
wget [url]下載，
tar -xvf解壓.tar.gz，
使用yum包管理工具yum search xx搜索, yum install xx下載。
whereis，which查找文件， find [path] -name xxx查找指定目錄下的文件
ps -ef | grep jupyter 查看jupyter運行的pid
安裝jupyter
- 下載Anaconda最新版本，省事。在官網(wǎng)找到下載鏈接，python3.6版本，然后
wget https://repo.anaconda.com/archive/Anaconda3-5.2.0-Linux-x86_64.sh
然后運行腳本
```sh
bash Anaconda3-3-5.2.0-Linux-x86_64.sh
```
中間需要確定時確定即可，等待安裝完成，會自動寫入環(huán)境變量。
```sh
conda --version
jupyter --version
```
- 然后按照官網(wǎng)步驟，設(shè)置config
jupyter notebook --generate-config
# 會生成config文件為 .jupyter/jupyter_notebook_config.py
設(shè)置密碼
```
from notebook.auth import passwd
passwd()
# 會生成類似 sha1:xxxxxxxxxxxxxxxxxxx， 需要記一下
```
然后編輯 config文件
```
vim .jupyter/jupyter_notebook_config.py
```
修改這些行，去掉注釋，填相應(yīng)值。其他選項也可以自己設(shè)置
```
c.NotebookApp.password = 'sha1:xxxxxxxxxxxxxxxxxxx'
c.NotebookApp.ip = '*'
c.NotebookApp.open_browser = False
c.NotebookApp.port = 8888
```
-
開放和重啟防火墻
centos7：firewall-cmd --zone=public --add-port=8888/tcp --permanent # 永久開放8888端口 firewall-cmd --reload # 重啟firewall firewall-cmd --list-ports # 查看開放端口centos6：
/sbin/iptables -I INPUT -p tcp --dport 8888 -j ACCEPT # 開放8888端口 /etc/rc.d/init.d/iptables save # 保存 service iptables status # 查看防火墻狀態(tài) 啟動server，
jupyter notebook。如果需要后臺啟動，運行
nohup jupyter notebook > jupyter.log 2>&1 &
安裝其他kernel
主要根據(jù) [Jupyter kernels wiki](https://github.com/jupyter/jupyter/wiki/Jupyter-kernels）
py2.7
conda create -n ipykernel_py2 python=2 ipykernel
source activate ipykernel_py2
python -m ipykernel install --user
ruby
iruby, ruby需要版本大于2.1
- 準備工作，安裝環(huán)境。
yum 下載的可能不是最新的ruby，所以用
sudo yum install -y git-core ruby-devel ruby zlib zlib-devel gcc-c++ patch readline readline-devel libyaml-devel libffi-devel openssl-devel make
- ruby kernel需要ZeroMQ
sudo yum install zeromq-devel zeromq czmq
gem install cztop rbczmq ffi_rzmq
- 安裝iruby kernel
gem install cztop iruby
iruby register --force
R
按照irkernel的安裝步驟，
或者直接
conda install -c r r-essentials
octave
pip
pip install octave_kernel
或者conda
conda config --add channels conda-forge
conda install octave_kernel
js
首先安裝nodejs和npm，然后
npm install -g ijavascript
ijsinstall
java
IJava ,需要安裝java 9或10，設(shè)置好環(huán)境變量，
export JAVA_HOME=/usr/java/jdk-10.0.1/
export CLASSPATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar
export PATH=$JAVA_HOME/bin:$PATH
然后
git clone https://github.com/SpencerPark/IJava.git --depth 1
cd IJava/
chmod u+x gradlew
./gradlew installKernel
c++
conda create -n cling
source activate cling
conda install xeus-cling notebook -c QuantStack -c conda-forge
有個問題是需要切換到cling環(huán)境啟動jupyter才會有c++的kernel，找了一下cling 的kernel文件
find anaconda3/envs/cling -name kernels
輸出
anaconda3/envs/cling/lib/python3.6/site-packages/notebook/services/kernels
anaconda3/envs/cling/lib/python3.6/site-packages/notebook/static/services/kernels
anaconda3/envs/cling/share/jupyter/kernels
進入到/share/jupyter/kernels，發(fā)現(xiàn)了
cd anaconda3/envs/cling/share/jupyter/kernels
ll
輸出
total 16
drwxr-xr-x 2 root root 4096 Jul 5 14:48 python3
drwxr-xr-x 2 root root 4096 Jul 5 14:48 xeus-cling-cpp11
drwxr-xr-x 2 root root 4096 Jul 5 14:48 xeus-cling-cpp14
drwxr-xr-x 2 root root 4096 Jul 5 14:48 xeus-cling-cpp17
現(xiàn)在，看一下jupyter kernel的文件目錄在哪
jupyter kernelspec list
輸出
Available kernels:
java /root/.ipython/kernels/java
ruby /root/.ipython/kernels/ruby
ir /root/.local/share/jupyter/kernels/ir
javascript /root/.local/share/jupyter/kernels/javascript
python2 /root/.local/share/jupyter/kernels/python2
python3 /root/.local/share/jupyter/kernels/python3
octave /root/anaconda3/share/jupyter/kernels/octave
bash /usr/local/share/jupyter/kernels/bash
確定是.ipython/kernels/的這一個。
將xeus-cling-cpp11 xeus-cling-cpp14 xeus-cling-cpp17這三個文件夾復(fù)制到.ipython/kernels/下
cp xeus-cling-cpp11 .ipython/kernels/
cp xeus-cling-cpp14 .ipython/kernels/
cp xeus-cling-cpp17 .ipython/kernels/
完成。
DigitalOcean 優(yōu)惠碼，注冊充值 100，鏈接一 鏈接二
Lastly, welcome to follow me on github
