DOC: Upload cheatsheets to site

[trallard]

• closes DOC: Fix up auto build procedure to upload Cheat Sheet to web site #40949
• tests added / passed
• Ensure all linting tests pass, see here for how to run them
• whatsnew entry

I went ahead and modified the .github/workflows/ci.yml so that the cheatsheets are copied to the docs build directory once the Sphinx site is built

[jreback]

cc @Dr-Irv @datapythonista over to you

[Dr-Irv]

This isn't going to solve the problem. Right now, we have that https://pandas.pydata.org/docs/getting_started/index.html#tutorials has a link
to https://pandas.pydata.org/Pandas_Cheat_Sheet.pdf .

I think we just want to copy Pandas_Cheat_Sheet.pdf to web/build to fix this.

Having said that, maybe @datapythonista knows more about how the web site is built/structured to confirm.

We could consider changing the link in the tutorial, but there are a lot of links to the current location on the web, so we would then need to put a redirect on the web site, which is probably harder.

[datapythonista]

Agree with @Dr-Irv, we shouldn't change the location of the cheatsheets, we try ro be quite conservative with breaking links. And if we want to consider changing the location, let's keep it separated from this change. Besides that detail, the PR looks great. I do think we need to remove the exclude above @Dr-Irv, otherwise the chetsheets wouldn't upload. So, it's just changing the path were we upload the files and we're good to go IMO.

[Dr-Irv]

@datapythonista Do you agree with this statement that I made above? (I'm not 100% sure it's correct)

I think we just want to copy Pandas_Cheat_Sheet.pdf to web/build to fix this.

[datapythonista]

There are more cheatsheets than that, there is at least a Japanese one iirc, but other than that I agree. The file needs to be available at the current location https://pandas.pydata.org/Pandas_Cheat_Sheet.pdf

[trallard]

Thanks, folks I have made the requested changes so now the cheatsheets should be under web/build

Apologies for messing the path in the first place - there were some talks of having a dedicated folder for cheatsheets in #4090 so I think I jumped the gun

[Dr-Irv]

I'm fine with this now, but @datapythonista how can we test this? How often does this script get run so that the web site is updated?

[datapythonista]

Thanks @trallard, really nice.

The CI for this commit (PR merge) will update the website, and the cheatsheet in https://pandas.pydata.org/Pandas_Cheat_Sheet.pdf should be updated with the new version. The new cheatsheet contains 3 columns instead of two for the "Subset" sections, in the bottom right part of the first page. In 30 minutes or so we should see the change. If the job fails, the cheatseet is not replaced, or whatever, we'll follow up with another PR. There is no easy way to test this further without merging the PR.

Note that the copying of the cheatsheet to the directory to upload was already being run in this PR. So the job shouldn't be failing. It's just the upload that hasn't been tested already.

[datapythonista]

Ups, just realized now that we're copying the cheatsheets to the directory after the directory is uploaded, so the cheatsheets won't be pushed, and will be delete from the server, my bad. I'll open a PR to fix it.