setup notes

os dependencies

• conda, eg conda-forge/miniforge
• github cli
• visual studio code
  • install:

    curl -L https://go.microsoft.com/fwlink/?LinkID=760868 --output visual_studio_code.deb
    sudo apt install ./visual_studio_code.deb

  • use tunneling to develop from a remote host (https://code.visualstudio.com/docs/remote/vscode-server)

    code tunnel --accept-server-license-terms --log debug

• git aliases

  curl https://raw.githubusercontent.com/GitAlias/gitalias/master/gitalias.txt -o ~/.gitalias
  git config --global include.path ~/.gitalias

• jupyter-related dependencies to the base conda env

  conda install -y -n base -c conda-forge mamba
  mamba env update --file conda-env-base-extras.yml
  mamba update -n base -c conda-forge --update-all

• other linux stuff

  sudo apt update
  sudo apt install -y \
      pandoc \
      tree \
      texlive-xetex \
      texlive-fonts-recommended \
      texlive-plain-generic

conda env for project code

mamba env create --file conda-env.yml
mamba env update --file conda-env.yml
mamba activate deconv

add deconv kernel to jupyter

# in the base env, in which jupyter lab runs
python -m ipykernel install --user --name=deconv
# or, to mimic nb_conda_kernels
python -m ipykernel install --user --name=conda-env-deconv-py

install helpers

# in the deconv env
pip install --verbose --no-build-isolation --editable .

configuring access to google cloud storage with gcloud auth

gcloud auth application-default login
# and maybe also...
gcloud auth login
gcloud config set project keen-dispatch-316219

restart jupyter lab server

sudo service jupyter status
sudo service jupyter restart

Source: https://cloud.google.com/vertex-ai/docs/general/troubleshooting-workbench#restart_the_jupyter_service