search

yt-project / website

The source for http://yt-project.org/
7 stars 33 forks source link

Update examples to use load_sample #94

Closed matthewturk closed 9 months ago

matthewturk commented 3 years ago

This updates the examples to use load_sample. Once #91 is included it'll be just those changes.

neutrinoceros commented 3 years ago

wait the images are still shown as changed.

Also, for posterity: as discussed in #91, this shouldn't be merged before yt.load_sample is shipped (4.0)

matthewturk commented 3 years ago

a merge fixed it :)

neutrinoceros commented 3 years ago

Btw you may prefer to keep the full paths in every example. It's not required, but it produces identical results and may be considered best practice (depending on the example we want to give). Normally, the whole patch could just be about substituting yt.load( with yt.load_sample(, and if we don't remove the full path data, the reversed substitution will also work as long as the data is already installed (as was assumed in existing versions of the docs).

matthewturk commented 2 years ago

Hey, what do we think about this nowadays?

neutrinoceros commented 2 years ago

it's in better shape now that yt.load_sample has been shipped for 4 releases over a year. However the function still requires two optional deps (pooch and pandas), and personally I don't like it that we force users to install pandas just for this (for the record it was my idea from 2 years ago), so I would like to refactor it so pandas isn't needed, but I digress.

We wrap ImportError for both these dependencies so it limits the risk of confusion, so I wouldn't oppose this change now. Tell me if you want to add to it, otherwise I can just merge as is

chrishavlin commented 9 months ago

@neutrinoceros looks like as of the last comment you're OK with this PR, that still true? @matthewturk how you feeling about it?

chrishavlin commented 9 months ago

for what it's worth, +1 from me for being merged as-is at this point.

neutrinoceros commented 9 months ago

yes, go ahead. I never found the motivation to refactor pandas out of the function and at this point I doubt I ever will.