fix(python): Include pl. qualifier for inner dtypes in to_init_repr

[datenzauberai]

Closes #15802

This is a minimal fix for #15802 by adding a method that prefixes representations, which will be called "recursively" for nested types.

[datenzauberai]

I somehow managed to close the preceding PR #16223 where @stinodego suggested to create a utility function instead of putting this into the DataType classes.

[datenzauberai]

It would be an option to include an optional prefix parameter for DataFrame.to_init_repr and Series.to_init_repr, but I don't think this flexibility is really needed.

[codecov]

Codecov Report

Attention: Patch coverage is 87.50000% with 4 lines in your changes missing coverage. Please review.

Project coverage is 80.80%. Comparing base (a7f9c8d) to head (266ef82).
Report is 1 commits behind head on main.

❗ Current head 266ef82 differs from pull request most recent head 316f772

Please upload reports for the commit 316f772 to get more accurate results.

Files	Patch %	Lines
py-polars/polars/datatypes/_utils.py	87.09%	2 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16235      +/-   ##
==========================================
- Coverage   81.32%   80.80%   -0.52%     
==========================================
  Files        1423     1394      -29     
  Lines      187177   179407    -7770     
  Branches     2721     2927     +206     
==========================================
- Hits       152214   144975    -7239     
+ Misses      34468    33927     -541     
- Partials      495      505      +10     

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

[datenzauberai]

Noted one extra thing: Is it really possible with the current version to have a pl.List/pl.Array without a properly defined inner dtype (I think that was possible in the past)? In the constructor both call py_type_to_dtype which would raise an Exception. Would simplify things a little bit...

[stinodego]

Thanks, looks good! A few comments.

Noted one extra thing: Is it really possible with the current version to have a pl.List/pl.Array without a properly defined inner dtype (I think that was possible in the past)? In the constructor both call py_type_to_dtype which would raise an Exception. Would simplify things a little bit...

This is possible for user defined types, but not for data types returned by Series.dtype. So you don't have to handle it in this function.

[ritchie46]

Any news on this one @stinodego

[datenzauberai]

Just a friendly ping regarding the open pull request @stinodego

[stinodego]

It looks good, but the Array repr has been updated. I'll send an update and this can be merged.

[stinodego]

I updated the Array repr and added a parametric test to make sure we will not diverge with the official repr.

Thanks for the PR!