Declutter example notebook

[ashjbarnes]

Here I've removed all descriptions of Gadi / NCI and almost all of Payu from the reanalysis_forced notebook, and deleted the model forced one. This will be migrated to cosima recipes

The old notebook cells for "setup era5" and "set up run directory" are now methods of the experiment class which further streamlines the notebook. The using_payu flag is included here to give users the option of a config.yaml file being generated for their experiment. It defaults to False.

The default input directory has also been decluttered to make things generic and not Gadi-centric

Finally, to address the deprecation of the motu client (issue #79 ) I've removed this entirely and given users step by step instructions for using the GUI on Copernicus's website to download the required data.

Related to #59 in that I've updated the model-forced example but I won't delete it until we find a good home for it elsewhere (see COSIMA/cosima-recipes#308 )

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

Attention: 92 lines in your changes are missing coverage. Please review.

Comparison is base (9388117) 72.97% compared to head (5fea3a0) 61.02%.

Files	Patch %	Lines
regional_mom6/regional_mom6.py	9.80%	92 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             main      #83       +/-   ##
===========================================
- Coverage   72.97%   61.02%   -11.95%     
===========================================
  Files           3        3               
  Lines         444      544      +100     
===========================================
+ Hits          324      332        +8     
- Misses        120      212       +92     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ashjbarnes]

I'm going to update / fix the 'model forced' one then re-open this PR. Until we find a landing place for it I'll at least iron out some of the issues pointed out by Luwei

[ashjbarnes]

I've re-added the ACCESS-OM2 example but made it clear in the markdown that it's intended for NCI users. This should be migrated elsewhere in COSIMA when we agree on the space

[ashjbarnes]

Codecov is down because I've moved the process era5 functionality to the package. A future PR that expands testing should address this rather than me making this PR longer. I've manually tested both notebooks and run the corresponding MOM6 experiments they produce

[ashjbarnes]

@angus-g latest commit that re-introduces yaml for string manipulations fails on example notebooks. Could we please not make any commits that fail for the example notebooks? Ideally this would have been caught by unit testing that's not implemented yet, but the execution of the example notebooks start to finish are a good way to ensure commits aren't code breaking

I'm currently using a branch from declutter_notebook for my own research, which will eventually be the next PR incorporating tides into the workflow, hence this branch needs to remain functional.

I think it would just be easiest if the string manipulation stuff is dealt with in a separate PR, as it's beyond the scope of what this PR intended to do (i.e, migrate existing code from the example notebooks into the codebase)

The error was ultimately: RepresenterError: cannot represent an object: [/scratch/v45/ab8992/regional_mom6_configs/tasmania-example](https://file+.vscode-resource.vscode-cdn.net/scratch/v45/ab8992/regional_mom6_configs/tasmania-example)

[angus-g]

Apologies, fixed. I had no clue you were working live off a PR branch, I would strongly advise against doing that.

[ashjbarnes]

If not, then from a github workflow perspective what's the best way for me to keep working for the moment? I'm losing a lot of time trying to juggle the codebase when this PR is unresolved and I'm already working on the next one (and need it for my day-to-day research). It was becoming too much admin to continually merge changes from this branch into the withtides branch, especially when merge conflicts cropped up.

I then manually made a new branch forked from declutter_notebook to build off but then had to troubleshoot why things weren't working. Perhaps there's a better way of doing things that I'm not aware of?

[ashjbarnes]

Apologies, fixed. I had no clue you were working live off a PR branch, I would strongly advise against doing that.

Yes that makes sense! I wanted to be able to use the improvements from this branch in my day-to-day research, but at some point there were a bunch of annoying merge conflicts when I kept trying to pull changes into withtides. Is it generally best practice to just work on branches from main and wait for PRs to be merged before merging them into your other dev branches, even if it means you're working without some of the new features?

[navidcy]

I haven't followed any discussion here. @ashjbarnes, @angus-g is this PR under control? If you need help ping me.

[ashjbarnes]

Latest commit implements what we talked about today @angus-g . Notebook has been tested but I'll go back and modify notebooks to add some more documentation - that'll be the last planned commit from me unless something else comes up / breaks!

Thanks @navidcy ! Almost finished :)

[navidcy]

tip:
before pushing you should format with black otherwise tests don't run... and tests that fail might hint you about something wrong that is happening. If you postpone formatting black (and thus running the tests) you may not see a bug you introduced until way after you've put substantial effort in refactoring something that you may need to undo.

Fixed

[ashjbarnes]

Everything tested and working with latest executable. Ready to merge