Jupyter lab vs. jupyter-lab

JupyterLab
#1

Hi,
since ever I was able to start jupyter lab from the windows 10 CMD console using jupyter lab.
However, since a new installation of python and jupyter lab I am onyl able to open the program using jupyter-lab.
I am asking myelf if this can cause any problems or has any other ramifications going forward. Also is there a change to put it back to the command jupyter lab, wich is the command in the official documentation?
Thanks for any help in advance!
Oscar

#2

From the command line, try:

where jupyter
where jupyter-lab

If they have radically different prefixes, you may want to remove one or the other.

There may also be issues with your %PATH%:

set

See if your PATH (or PYTHONPATH) contain multiple python environments. This is also a recipe for disaster.

As a windows user, it may be worth looking at this screed and the associated contents about keeping Jupyter tools working “the hard way” in Windows with mamba.

1 Like
#3

Hi,
many thanks!
I tried the where commands and got the follwing:

C:\Users\user>where jupyter
C:\Users\user\AppData\Roaming\Python\Python311\Scripts\jupyter.exe

C:\Users\user>where jupyter-lab
C:\Python\Scripts\jupyter-lab.exe

PATH is set as follows:
2022-12-24 14_32_42-Umgebungsvariable bearbeiten

Any hints?
Thanks in advance!
Oscar

#4

I can’t in good conscience recommend using the official python installer on Windows, much less multiple times on your machine or whatever you are seeing. The number of things that won’t work right is just very high.

As recommended in the post linked above:

  • remove all your existing python installations and references to them
    • with the GUI tool if provided
      • rd /s /q if you must
        • then manually clean up the environment variables, desktop shortcuts, registry entries, and wherever else stuff goes on windows
  • start with Mambaforge in a very short path (e.g. C:\mf) on your fastest drive
    • keep that base environment very small (e.g. just mamba)
  • build up from there with per-project environment.yml
    • even pip-only packages, can be managed in the pip: section
  • use mamba env update for everything
    • never pip install anything
      • other than your packages under development with pip install -e .
    • never mamba install anything
    • never use conda for much of anything
  • use per-environment Windows Terminal shortcuts to manage running in the right environment
1 Like
#5

Hi Nicholas,
many thanks and happy holidays!
Best,
Oscar

1 Like