feat(rust, python): add dt.quarter_start, dt.quarter_end, dt.year_start and dt.year_end

[SozinM]

closes #14511

• [x] Implement functionality
• [x] Implement tests
• [x] Correct docs

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 96.48241% with 14 lines in your changes are missing coverage. Please review.

Project coverage is 81.28%. Comparing base (802d6c8) to head (e9d5f52).

:exclamation: Current head e9d5f52 differs from pull request most recent head 86a0755. Consider uploading reports for the commit 86a0755 to get more accurate results

Files	Patch %	Lines
crates/polars-time/src/period_start.rs	94.41%	10 Missing :warning:
...ates/polars-plan/src/dsl/function_expr/datetime.rs	92.45%	4 Missing :warning:

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14513      +/-   ##
==========================================
+ Coverage   81.14%   81.28%   +0.13%     
==========================================
  Files        1362     1356       -6     
  Lines      174846   175938    +1092     
  Branches     2531     2522       -9     
==========================================
+ Hits       141881   143010    +1129     
+ Misses      32481    32444      -37     
  Partials      484      484              

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[MarcoGorelli]

thanks - looks like it's on the right track, will do a full review next week

[SozinM]

All comments were addressed @MarcoGorelli

[MarcoGorelli]

thanks @SozinM for updating - could you update the merge conflict please? It should just be a matter of passing NonExistent::Raise to replace_time_zone

[SozinM]

@MarcoGorelli done :)

[SozinM]

@alexander-beedie Hi! Would you like to review this?

[SozinM]

@MarcoGorelli could you suggest a person to ask for an additional review?

[alexander-beedie]

@alexander-beedie Hi! Would you like to review this?

~This is one for @MarcoGorelli 😎👍 But the functionality looks useful!~

Ahh, just spotted that it's already approved by @MarcoGorelli and you're just waiting for a merge. Looks like you need to rebase once more (there are currently some conflicts), and then we can take a final look and get it in👌

[SozinM]

@alexander-beedie Hi! Rebased, take a look please :)

[SozinM]

@alexander-beedie soft bump

[SozinM]

Hi @c-peters and @stinodego, I randomly chose you to ask to take a look if you have free time :)

[SozinM]

Hi @reswqa , @orlp ! Maybe you will save the day? (or save this PR from being buried in history)

[SozinM]

Hi @ritchie46 maybe you will be able to help?

[MarcoGorelli]

hi @SozinM - seeing as this is still open, it may be good to get your thoughts on https://github.com/pola-rs/polars/issues/16094

if month_end were indeed disallowed for Datetimes, it may make sense to remove it from this PR too. do you have thoughts on that?

[stinodego]

I've felt a bit apprehensive pressing the merge button, because I'm not sure these methods are really that useful. Don't we already offer the right tools to do this yourself? E.g.

from datetime import date

import polars as pl

df = pl.DataFrame({"date": [date(2024, 2, 29)]})
result = df.with_columns(
    year_end=pl.date(year=pl.col("date").dt.year(), month=12, day=31)
)
print(result)

shape: (1, 2)
┌────────────┬────────────┐
│ date       ┆ year_end   │
│ ---        ┆ ---        │
│ date       ┆ date       │
╞════════════╪════════════╡
│ 2024-02-29 ┆ 2024-12-31 │
└────────────┴────────────┘

The API bloat just doesn't seem worth it.

[SozinM]

Okay, this PR got a little out of hand so I will close it. If somebody wants to pick it up - feel free to repurpose the code :)