Added troubleshooting for installing cf next to Jupyter and conda

[ThatDesert]

• Added a section in installation do s below conda installation instructions that talks about what to do when installing cf alongside Jupyter (common use case)
• Attempt to give guidance on how to solve Conda, dependency clash when installing cf-python and cf-plot alongside Jupyter. #883

[sadielbartholomew]

Thanks for this, Ollie! Good spot about potential environment issues here.

I have made some suggestions to clarify the edge case and think we should try to check the workaround together tomorrow since I am not convinced it works (I wonder if jupyter would run properly even if the installation commands didn't complain)...

[ThatDesert]

Looks good, thanks for the clarification!

[sadielbartholomew]

Great stuff. Thanks Ollie.

As for the workaround you suggested - it does indeed seem to work (as tested through a new Python 3.12 conda environment)! Logically based on the dependency constraints I don't understand why (although guess that perhaps somehow they relax certain constraints when installation happens in a different order, or something funky), but if it works for both of us independently then we can only assume it should work generally. So, we can ignore my comment here: #884 (comment). Thanks for finding it and apologies for questioning it - as conveyed in person I didn't understand the logic behind it, but really the key is what works in practice 🙂

I am happy to merge this now. Though, since you've made your first contribution to the library, it'd be great f you could add your name to the contributor list at github.com/NCAS-CMS/cf-python/blob/main/docs/source/contributing.rst#contributors, via editing that documentation source RST file to include your name (in your preferred format) in the list according to alphabetical order.

Please let me know if you'd prefer to add your name to the contributor list outside of this PR by comment, else push another commit to do that and once that's up I will merge this PR.

[sadielbartholomew]

To add: in case useful going forward, the versions of relevant libraries I end up with (using Python 3.12) after following your workaround Ollie are:

$ conda list | grep esmpy                                                                ─╯
esmpy                     8.8.1              pyhecae5ae_0    conda-forge
$ conda list | grep jupyter                                                              ─╯
jupyter                   1.1.1           py312h06a4308_0  
jupyter-lsp               2.2.5           py312h06a4308_0  
jupyter_client            8.6.3           py312h06a4308_0  
jupyter_console           6.6.3           py312h06a4308_1  
jupyter_core              5.8.1           py312h06a4308_0  
jupyter_events            0.12.0          py312h06a4308_0  
jupyter_server            2.16.0          py312h06a4308_0  
jupyter_server_terminals  0.5.3           py312h06a4308_0  
jupyterlab                4.4.4           py312h06a4308_0  
jupyterlab_pygments       0.3.0           py312h06a4308_0  
jupyterlab_server         2.27.3          py312h06a4308_0  
jupyterlab_widgets        3.0.15          py312h06a4308_0  
$ conda list | grep zlib                                                                 ─╯
libzlib                   1.2.13               h4ab18f5_6    conda-forge
zlib                      1.2.13               h4ab18f5_6    conda-forge

[sadielbartholomew]

Brilliant, thanks. Merging now.