Update indexing material

[negin513]

Adding two new notebooks in support of SciPy 2023:

1. 02.1_indexing_Basic.ipynb (Comprehensive coverage of all basic indexing)
2. 02.2_indexing_Advanced.ipynb

The first notebook includes comprehensive coverage of all basic indexing and can basically replace the 02.1_working_with_labeled_data.ipynb. But I wanted to first ask for feedback before removing the other notebook.

The second notebook includes :

• Vectorized Indexing
• Boolean Indexing and isin, and masking using where

Relevant Issue: #170

We can add materials for Fancy Datetime Indexing here @andersy005 .

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[dcherian]

Very nice! I just took a quick look, will do a more thorough review tomorrow.

The first notebook includes comprehensive coverage of all basic indexing and can basically replace the 02.1_working_with_labeled_data.ipynb. But I wanted to first ask for feedback before removing the other notebook.

Perhaps we can copy the exercises over? That notebook also has some date time indexing material that could be copied?

Can you add entries in _toc.yml please? I think your second notebook should go in the Intermediate section.

[dcherian]

remove the logo.

[dcherian]

I think you'll need one or two more lines to explain "orthogonal" indexing.

We should also link to the xarray docs appropriately.

[dcherian]

Let's add a new subsection for "pointwise indexing"

[dcherian]

For masking, another motivation is selecting regions that are not regular: like lat and lon ranges on a curvilinear grid. A figure might be useful to illustrate the difference. "rasm" is a useful dataset for this.

[dcherian]

Let's use the myst syntax where we can instead of alerts and bold font

[dcherian]

Masking and isin can be grouped under "Boolean indexing". They could be moved to its own notebook.

[dcherian]

We can skip "align and reindex". Reindexing can be its own section left for some future work.

[negin513]

Thanks for your comments.
I significantly updated this notebook based on your comments:
I added a section called orthogonal indexing. I tried to add the pointwise terminology as you recommended.

For the boolean masking :
I created a separate notebook for Boolean masking/indexing (as you recommended).
I updated the notebook to use rasm dataset and created a specific section/exercise telling them why they cannot use sel for such datasets for selecting regions.

[dcherian]

Here's a proposal for organizing:

• Fundamentals
  • Labeled data
    • Selecting Data <- First notebook : positional vs label-based indexing
    • Selection with Times <- Datetime Indexing stuff
• Intermediate
  • Indexing
    • Vectorized and Pointwise indexing
    • Boolean Indexing & Masking

What do you think? We've been evolving towards breaking up the material in to small pieces so it's easy to choose subsets for a syllabus.

We'll need to add the parts you want to cover to the workshops/scipy2023/README.md in the appropriate dropdown.

[negin513]

@dcherian I am not sure why the JupyterBuild of this notebook fails.

It complains about the redirect and a page already existing:

WARNING: (broken) fundamentals/02.1_working_with_labeled_data.html redirects to fundamentals/02.1_indexing_Basic.html but /home/runner/work/xarray-tutorial/xarray-tutorial/_build/html/fundamentals/02.1_working_with_labeled_data.html already exists!

Should I manually deleted the cached htmls ?

[dcherian]

You can change the key to force a new cache:

xarray-tutorial/.github/workflows/main.yaml

Lines 32 to 33 in 2ca47a9

	# NOTE: change key to "jupyterbook-DATE" to force rebuilding cache
	key: jupyterbook-20230626

[andersy005]

You can change the key to force a new cache:

xarray-tutorial/.github/workflows/main.yaml

Lines 32 to 33 in 2ca47a9

	# NOTE: change key to "jupyterbook-DATE" to force rebuilding cache
	key: jupyterbook-20230626

i usually delete stale cache via the actions UI, and this seems to invalidate it and force a new cache (under the same key though)

[andersy005]

@andersy005, please feel free to add a section to the basic or advanced indexing notebook on the fancy datetime indexing. I think at this point I am not going to make any other changes to these notebooks unless I hear back from you or @dcherian .

thank you for putting this together, @negin513! i intend to push changes to this PR tomorrow and sunday

[dcherian]

Thanks all. This is looking great.

I pushed some minor edits.

Some major comments:

• let's delete the numpy advanced indexing piece from "basic indexing" material. That's going to be confusing. It can be moved to the "advanced indexing" notebook if you prefer.
• "basic indexing" uses datetime indexing before it is introduced. Let's edit that to select along lat instead of time. or reorder the material

[dcherian]

We only have 30 minutes for the indexing topic, so perhaps just do only the Advanced and Vectorized Indexing notebook?

Given that I'm just going to merge here.

Thanks @negin513 and @andersy005 !

[negin513]

Sounds great. Thanks @dcherian. Do we have a schedule now and how long each topic is going to take?