fix: String transform to_titlecase was too narrowly defined

[alexander-beedie]

Closes #18116.

Note, closes the issue by better-aligning to_titlecase to the standard implementation, not by adding new options. (We were defining "word" boundaries by whitespace, but that is a much narrower definition than normal, which considers all non-alphanumeric chars as boundaries; updated the docstrings with a note to explain what constitutes a word boundary).

Example

import polars.selectors as cs
import polars as pl

df = pl.DataFrame({
    "quotes": [
        "'e.t. phone home'",
        "i feel the need--the need for speed",
        "to infinity,and BEYOND!",
        "say 'what' again!i dare you - I\u00a0double-dare you!",
        "What.we.got.here... is#failure#to#communicate",
    ]
})

with pl.Config(fmt_str_lengths=50):
    print(
        df.select(cs.string().str.to_titlecase())
    )

Before

# shape: (5, 1)
# ┌──────────────────────────────────────────────────┐
# │ quotes                                           │
# │ ---                                              │
# │ str                                              │
# ╞══════════════════════════════════════════════════╡
# │ 'e.t. Phone Home'                                │
# │ I Feel The Need--the Need For Speed              │
# │ To Infinity,and Beyond!                          │
# │ Say 'what' Again!i Dare You - I Double-dare You! │
# │ What.we.got.here... Is#failure#to#communicate    │
# └──────────────────────────────────────────────────┘

After

# shape: (5, 1)
# ┌──────────────────────────────────────────────────┐
# │ quotes                                           │
# │ ---                                              │
# │ str                                              │
# ╞══════════════════════════════════════════════════╡
# │ 'E.T. Phone Home'                                │
# │ I Feel The Need--The Need For Speed              │
# │ To Infinity,And Beyond!                          │
# │ Say 'What' Again!I Dare You - I Double-Dare You! │
# │ What.We.Got.Here... Is#Failure#To#Communicate    │
# └──────────────────────────────────────────────────┘

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 80.34%. Comparing base (9dd9569) to head (1136500).

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #18122      +/-   ##
==========================================
+ Coverage   80.33%   80.34%   +0.01%     
==========================================
  Files        1496     1496              
  Lines      197685   197685              
  Branches     2821     2821              
==========================================
+ Hits       158806   158831      +25     
+ Misses      38358    38333      -25     
  Partials      521      521              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ritchie46]

Seems more robust indeed.