Broken code in the gallery related to data access

[brian-rose]

Currently we're publishing some broken code on our gallery, see here:
https://projectpythia.org/xbatcher-ML-1-cookbook/notebooks/surface_currents.html#data-and-preprocessing

This is a data access problem that's already been discussed in #4 and #5.

It seems like the workaround was to suppress notebook execution during the build. Currently we have in the _config.yml:

execute:
  # To execute notebooks via a binder instead, replace 'cache' with 'binder'
  execute_notebooks: force
  exclude_patterns:
    - "notebooks/surface_currents_prep.ipynb"
    - "notebooks/surface_currents_model.ipynb"

But those notebooks are executing anyway during the build. I'm not sure why those flags aren't working.

This repo is also missing the flag nb_execution_raise_on_error: true which would normally prevent our build system from publishing to GitHub Pages if there's a notebook error. PR #8 puts that flag in, but is currently failing the build (because the problematic notebook is still executing).

[brian-rose]

Ah, I see the issue. There is nothing wrong with exclude flags. However we are still executing the main notebook notebooks/surface_currents.ipynb, and that notebook itself calls the other two notebooks, e.g. with code like

%run ./surface_currents_prep.ipynb

I don't know if there's a real workaround here.

Maybe this cookbook should be removed from the gallery until the data access problem is resolved.

[brian-rose]

@mgrover1 just making you aware of this since you've been looking at this cookbook recently.

[cmdupuis3]

Just FYI I'm going to ask around again about the data hosting.

[brian-rose]

We are working on implementing a solution for cookbook data storage prior to our next Hackathon this June.

[jukent]

I was trying to work on #12, but it's hard to parse or work on the error messages when i can't get the notebook to run at all locally. The first problem I encountered seems to be related to this issue?

When attempting to call prepare_data(sc5, 200, 1000, 1000, 200) I get the message

ValueError: Bad Request: https://storage.googleapis.com/download/storage/v1/b/pangeo-cesm-pop/o/control%2F.zmetadata?alt=media
User project specified in the request is invalid.

(same as in failing action) which brings me to a blank page that says "Bucket is a requester pays bucket but no user project provided."