feat: add DaskExpr.quantile

[anopsy]

Edit: Thanks for help @FBruzzesi

Hello, that's my attempt at solving Dask Expr.quantile- the method works as expected (according to my own tests).However I've been trying to find a way to include dask-case in our tests, since dask can work only with interpolation="nearest" and not the other ones.
What I tried:

• adding is_dask param to the parameters in tests and running the tests using nullcontext, but compare dict doesn't want to cooperate with error messages,
• writing separate test for dask cases and using if "dask" in str(contructor) to call that test but also failed here
  I was thinking about including every single constructor type in the parameters, but that seems a bit too verbose and the number of possible combinations is humongous.
  At the moment I'm exploring other options, but feel free to 🛟 send help 🛟 if you know how to approach it

What type of PR is this? (check all applicable)

• 💾 Refactor
• ✨ Feature
• 🐛 Bug Fix
• 🔧 Optimization
• 📝 Documentation
• ✅ Test
• 🐳 Other

Related issues

• Related issue feat: add methods to Dask backend #637
• Closes #

Checklist

• Code follows style guide (ruff)
• Tests added
• Documented the changes

If you have comments or can explain your changes, please do so below.

[FBruzzesi]

Hey @anopsy thanks for the PR 🙌🏼

My 2 cents: I would edit test_quantile_expr test as:

- if "dask" in str(constructor):
+ if "dask" in str(constructor) and interpolation != "nearest":
    request.applymarker(pytest.mark.xfail)

[anopsy]

I hope you enjoy your holiday 😃 Thank you so much I knew I'm missing some really simple solution 🫶

[FBruzzesi]

I think it is only missing an disclaimer/note in the exposed method and then it should be good to go?!

[anopsy]

like that?
Note: dask has its own method to approximate quantile and it doesn't implement 'nearest', 'higher', 'lower', 'midpoint' as interpolation method - use 'linear' instead

[FBruzzesi]

like that? Note: dask has its own method to approximate quantile and it doesn't implement 'nearest', 'higher', 'lower', 'midpoint' as interpolation method - use 'linear' instead

Yes, you could add it in the already existing note and create bullet points. If possible please consider adding an embedded link to the dask method as well.

[MarcoGorelli]

thanks @anopsy (and @FBruzzesi for reviewing!)

content looks good, merge when ready

[anopsy]

@FBruzzesi cool, thanks for suggestion, was wondering how to arrange those two notes, also what do you mean by an embedded link to the dask method? A link to the dask docs?
Edit : this one? https://docs.dask.org/en/stable/generated/dask.dataframe.DataFrame.quantile.html

[FBruzzesi]

We end up using Series.quantile