BLD: minor break ci/requirements-optional-pip.txt

[minggli]

• closes KeyError on master in TestExcelWrtiter #22595
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

When adding optional dependencies, below error appears from

Invalid requirement: 'openpyxl=2.5.5'
= is not a valid operator. Did you mean == ?

To core dev, should we fix the version or simply specify lower bound (i.e. >=)?

[datapythonista]

Thanks for the PR @minggli. Those files are autogenerated. Can you check the comments in this PR: #22749 and fix in the source file please (and then autogenerate the ones you changed)?

[datapythonista]

I think we want that specific version of openpyxl, so it'd be openpyxl=2.5.5 (not >=). Can you confirm @TomAugspurger

[codecov]

Codecov Report

Merging #22889 into master will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #22889      +/-   ##
==========================================
+ Coverage   92.18%   92.21%   +0.02%     
==========================================
  Files         169      169              
  Lines       50830    50975     +145     
==========================================
+ Hits        46860    47006     +146     
+ Misses       3970     3969       -1

Flag	Coverage Δ
#multiple	90.63% <ø> (+0.02%)	⬆️
#single	42.42% <ø> (+0.05%)	⬆️

Impacted Files	Coverage Δ
pandas/core/frame.py	97.44% <0%> (+0.24%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e45a6c1...57ec327. Read the comment docs.

[TomAugspurger]

Correct. We need to update scripts/convert_deps.py to rewrite exact pins as == for pip.

[minggli]

Thanks for the PR @minggli. Those files are autogenerated. Can you check the comments in this PR: #22749 and fix in the source file please (and then autogenerate the ones you changed)?

Thanks @datapythonista for looking into it. The script/convert_deps.py uses requirements-optional-conda.txt as base to generate, which has been fixed already.

with open("ci/requirements-optional-conda.txt") as f:
    optional = [x.strip() for x in f.readlines()]

so this should conclude the PR.

[datapythonista]

Cool, I didn't remember that. Lgtm. Thanks for the fix @minggli

[datapythonista]

Sorry, I think I misunderstood. If I understand correctly, the conda file required one equal only (so you shouldn't add the extra one), and the script convert_deps.py should be updated to make it two in the pip file. And after that generate the pip file.

[pep8speaks]

Hello @minggli! Thanks for updating the PR.

• There are no PEP8 issues in the file scripts/convert_deps.py !

[datapythonista]

lgtm

It can surely be in a separate PR, but I think at some point it'd be nice to have a single function in the script that we call twice (one for required and one for optional), and that we unit test.

[minggli]

good point. thanks Marc.

[jreback]

thanks. this is a dev requirement only.

[minggli]

thanks for reviewing @jreback @datapythonista @TomAugspurger!