CI,STYLE: GH38802 narrow down list of ignored words

[chinggg]

• closes CI,STYLE: narrow down list of ignored words #38811
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[MarcoGorelli]

Thanks @chinggg , this looks correct!

Regarding your comment in the closed PR, git is hard for everyone at the beginning - I found the first 3 chapters of this book to be invaluable in learning it https://git-scm.com/book/en/v2

---

Will merge when this is green, if you want to try similarly narrowing down other words, then that might well be possible

--

For the future, I'd advise creating a new branch when opening a PR, see the contributing guide https://pandas.pydata.org/pandas-docs/dev/development/contributing.html

[MarcoGorelli]

Sorry to go back on this, but I think we can do better

The changes you've made for datas are fine, but it might be better to reverse the ones for fox/nox, as fox as at least is a real animal.

The way to do that is:

1. add a new file called .codespellignorelines
2. in that file, add the following:

        1    fo

3. in setup.cfg, add the following to the codespell section so it looks like this:

[codespell]
ignore-words-list=ba,blocs,coo,hist,nd,ser
exclude-file=.codespellignorelines

Like this, it should be possible to get rid of some other words to ignore, such as ba. nd and hist are pretty common, and make sense to keep, but those like ba that just appear sporadically we could add to the .codespellignorelines file

---

Do let me know if any of this is unclear or if you want/need help with git commands

[chinggg]

Thanks a lot for explaining patiently!
I hava realized the general working mechanism of codespell. That is very cool and must be useful in large projects like pandas. We should consider how frequently and commonly the words are used by programmers around the world, and only add those widely-accpeted acronyms or abbreviations to the ignore-words-list. Those temporary symbols should be put in the .codespellignorelines file instead.
Because I do not live in an english-speaking country, I am not sure about the frequency of these shorten words. So I use codespell to check where and how they show up in the code.
I found that

• blocs may be a common abbr. for blocks
• coo seems to be a mathematical term
• nd stands for N-dimensional
• ser is short for series
• hist only has 4 results but it is also used by matplotlib so I consider it as a common practice.
  So I decide to remove only ba and datas(already done by your instruction to rename datas to data).
  I make a new branch named codespell-ignore. Because the branch of this PR is master and seems to be unchangeable, so I have to make a new PR again. I hope it will not be improper.

[MarcoGorelli]

Sure, that's fine, thanks!

And yes, I agree with your assessment of which words to keep in ignore-words-list

[MarcoGorelli]

I have to make a new PR again. I hope it will not be improper.

Closing this one in favour of the new PR then