Adding a Custom IPython Kernel in a Notebook Instance

Scenarios

If the default engine environment in notebook does not meet your needs, you can create a conda environment to set up your own environment as required. This section uses python3.6.5 and tensorflow1.2.0 as an example to describe how to set up an IPython kernel.

Procedure

Create a conda environment.

Run the command below on the notebook terminal. In the command, my-env indicates a customizable virtual environment name. For details about conda parameters, see the conda official website.

conda create --quiet --yes -n my-env python=3.6.5

After the creation, run the conda info --envs command to view existing virtual environments, where you can see the my-env virtual environment.

sh-4.4$conda info --envs
# conda environments:
#
base                  *  /home/ma-user/anaconda3
TensorFlow-2.1           /home/ma-user/anaconda3/envs/TensorFlow-2.1
my-env                   /home/ma-user/anaconda3/envs/my-env
python-3.7.10            /home/ma-user/anaconda3/envs/python-3.7.10                  /opt/conda/envs/my-env

Access the conda environment.

source /home/ma-user/anaconda3/bin/activate /home/ma-user/anaconda3/envs/my-env

Install the following dependency packages in my env:

pip install jupyter
pip install jupyter_core==5.3.0 
pip install jupyter_client==8.2.0 
pip install ipython==8.10.0
pip install ipykernel==6.23.1

Add the virtual environment as an IPython kernel.

The value of --name can be customized.

python3 -m ipykernel install --user --name "my-py3-tensorflow-env"

After the command is executed, the following information is displayed:

(my-env) sh-4.4$python3 -m ipykernel install --user --name "my-py3-tensorflow-env"
Installed kernelspec my-py3-tensorflow-env in /home/ma-user/.local/share/jupyter/kernels/my-py3-tensorflow-env

Customize the environment variables of the virtual environment kernel.

Run the cat /home/ma-user/.local/share/jupyter/kernels/my-py3-tensorflow-env/kernel.json command to view the default configuration.

{
 "argv": [
  "/home/ma-user/anaconda3/envs/my-env/bin/python3",
  "-m",
  "ipykernel_launcher",
  "-f",
  "{connection_file}"
 ],
 "display_name": "my-py3-tensorflow-env",
 "language": "python"
}

Add content in the env field as needed. For reference, see the following configuration. PATH indicates the path to the Python package of the virtual environment.

{
 "argv": [
  "/home/ma-user/anaconda3/envs/my-env/bin/python3",
  "-m",
  "ipykernel_launcher",
  "-f",
  "{connection_file}"
 ],
 "display_name": "my-py3-tensorflow-env",
 "language": "python",
 "env": {
      "PATH": "/home/ma-user/anaconda3/envs/my-env/bin:/opt/conda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/home/ma-user/modelarts/ma-cli/bin",
      "http_proxy": "http://proxy-notebook.modelarts-dev-proxy.com:8083",
      "https_proxy": "http://proxy-notebook.modelarts-dev-proxy.com:8083",
      "ftp_proxy": "http://proxy-notebook.modelarts-dev-proxy.com:8083",
      "HTTP_PROXY": "http://proxy-notebook.modelarts-dev-proxy.com:8083",
      "HTTPS_PROXY": "http://proxy-notebook.modelarts-dev-proxy.com:8083",
      "FTP_PROXY": "http://proxy-notebook.modelarts-dev-proxy.com:8083"
 }
}

Access the IPython kernel of the virtual environment.
Refresh the JupyterLab page. The custom virtual environment kernel is displayed.

Click my-py3-tensorflow-env to verify that the current environment is used.
Clear the environment.
Delete the IPython kernel of the virtual environment.
```
jupyter kernelspec uninstall my-py3-tensorflow-env
```
Delete the virtual environment.
```
conda env remove -n my-env
```