Configure server-side jupyter under linux, local access & add conda virtual environment in jupyter

Configure server-side jupyter under linux, local access

0. Background

The laboratory server does not have a graphical interface and can only be accessed through sshlogin, but there are many scripts involved in the alchemy process, and I want to experience the interactive process. After checking, you can configure the server-side jupyter, and then use the local browser to remotely access.

1. Install jupyter notebook

pip install jupyter

2. Generate configuration files

jupyter notebook --generate-config

3. Set login password and generate secret key

• sha1The secret key needs to be copied and used in the next step

4. Modify the configuration file

• Enter the folder where the configuration file is located:cd ~/.jupyter

• Edit the configuration file jupyter_notebook_config.py:vim jupyter_notebook_config.py

• After entering vim, press i to enter edit mode, add configuration information at the end of the file

c.NotebookApp.ip='*'               # *表示所有ip可以访问
c.NotebookApp.password = 'sha1:秘钥'  #第三步生成的秘钥
c.NotebookApp.open_browser = False
c.NotebookApp.port =8888           #端口号
c.NotebookApp.notebook_dir = '/home/user/jupyter_file' #ipython文件保存的路径，按自己的路径修改，必须要配置，不然访问时会出现404错误。

5. Generate jupyter access address on the server

• The following ipinto your serverip
• Replace the port with the port number configured in the above file

jupyter notebook --ip=10.141.222.152 --port=8888 --allow-root

6. Set up the server to run jupyter in the background

nohup jupyter notebook --allow-root > jupyter.log 2>&1 &

• or

nohup jupyter notebook --ip=10.141.222.152 --port=8888 --allow-root > jupyter.log 2>&1 &

• View
  ps -a
• Delete
  kill -9 pidterminate the process

Reference

https://www.cnblogs.com/nxf-rabbit75/p/11906199.html

Add conda virtual environment in jupyter

1. First install ipykernel, execute the command line under terminal

conda install ipykernel

2. Create a kernel file in the virtual environment and execute the command line in the terminal

conda install -n 环境名称 ipykernel

3. Activate the conda environment and execute the command line in the terminal:

source activate 环境名称

4. Write the environment to the notebook's kernel

python -m ipykernel install --user --name 环境名称 --display-name "在jupyter中显示的环境名称" 

The name in the quotation marks here can be freely used for identification in jupyter. Here, my jupyterenvironment name is called py36, so my command is:python -m ipykernel install --user --name py36 --display-name "py36"

**5. Open the notebook server (see above for remote configuration)

• In the terminalimplementation of the command linejupyter notebook
• The above related steps can complete jupyterthe related configuration, but if you need to use it often jupyter notebook, it is best to install it when you create the virtual environmentipykernel
  • The command is as follows

conda create -n 环境名称 python=3.5 ipykernel

6. In addition, list and delete the kernel environment

• List the kernelenvironment

jupyter kernelspec list

• Delete kernelenvironment

jupyter kernelspec remove 环境名称

Refer

reference