Debugging Jupyter

Debugging Jupyter#

Install Jupyter on Linux#

On Linux install nodejs and Jupyter:

sudo apt-get install nodejs
sudo apt-get install jupyter

Issues Starting Jupyter Lab#

For some reasons, the jupyter lab or jupyter-lab commands do not work in another conda environment than base, or on Linux environments. In these cases, open Anaconda Prompt (and activate the concerned environment), or Linux Terminal and tap:

python -m pip install jupyter-lab

Alternatively, pip install jupyterlab (or pip3 install jupyterlab) works just as well.

Try to run jupyter lab or jupyterlab. If it still does not work, try (on Linux):

/usr/bin/env python /home/USER-NAME/.local/python3.X/site-packages/jupyterlab

Make sure to replace USER-NAME with your local user name and python3.X with the installed version of Python (e.g., python3.7).

Jupyter Does not Recognize Installed Packages in Conda Env#

If you launch Jupyter from within a conda environment and the jupyter notebooks are not able to import the packages installed in the environment, you will need to install the ipykernel for the particular environment, too. For instance, this error applies when the import from osgeo import gdal in a jupyter notebook Python code cell started from the active flussenv environment results in an ImportError. To troubleshoot this error:

  • Shutdown Jupyter(Lab)

  • In Anaconda prompt, make sure the target environment is activated (e.g., conda activate flussenv)

  • Verify if ipykernel is installed by typing conda list

  • If ipykernel is not installed, type conda install -c anaconda ipykernel to install it

  • Add a new kernel with:

    • ipython kernel install --user --name=KERNEL_NAME

    • Replace KERNEL_NAME with the name you want to use (e.g., fluss_kernel)

  • Re-run jupyter-lab to select the new kernel from the top menu (Kernel > Change Kernel…)

Contents