Outline
Virtual Environment
Everything starts w/ an independent virtual environment. Install conda will make it easier.
After installation, it might be necessary to add this line to .bashrc or .zshrc:
export PATH="${PATH_TO_ANACONDA}/bin:$PATH"
To create a virtual environment:
conda create -n lb python=3 numpy pandas
To activate the virtual environment:
conda activate lb
To install more packages to virtual environment:
conda install -n lb matplotlib
To export an environment configuration:
conda activate myenv
conda env export > environment.yml
To recreate a venv w/ given environment.yml:
conda env create -f environment.yml
To delete a virtual env:
conda remove -n ${myenv} --all
or:
conda env remove --name ${myenv}
Reference: conda wiki page
Jupyter Notebook
To install jupyter notebook:
conda install jupyter notebook
To easily manage conda virtual environment in jupyter notebook, install nb_conda:
conda install nb_conda
To install virtual environment into ipykernel jupyter notebook, run:
python -m ipykernel install --user --name lb --display-name "Python3 (lb)"
To remove virtual environment from ipykernel jupyter notebook:
# notice ${unwanted-kernel} should be name instead of display name
jupyter kernelspec uninstall ${unwanted-kernel}
To open jupyter notebook w/o browser and do port forwarding stuffs:
jupyter notebook --no-browser --port=9000
To get access to this jupyter notebook from laptop:
ssh -N -L localhost:9000:localhost:9000 [email protected] & sleep 0.5; open -a Safari http://localhost:9000
Then the ssh job will be running in the background. To list all jobs, just run jobs. In case you want to kill the job, get its id and run kill %{ssh_job_id}. For more info on CLI job, go here.
To convert a notebook to an HTML file:
jupyter nbconvert --to html notebook.ipynb
Jupyter nbextensions:
conda install -c conda-forge jupyter_contrib_nbextensions