Set no_implicit_reexport = true in pyproject.toml

[matteosantama]

One step closer to enabling strict = true for mypy.

[codecov-commenter]

Codecov Report

Merging #4211 (d695639) into master (95074d9) will decrease coverage by 0.05%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #4211      +/-   ##
==========================================
- Coverage   64.17%   64.11%   -0.06%     
==========================================
  Files         457      457              
  Lines       75611    75613       +2     
==========================================
- Hits        48521    48483      -38     
- Misses      27090    27130      +40     

Impacted Files	Coverage Δ
py-polars/polars/internals/frame.py	91.55% <ø> (ø)
py-polars/polars/__init__.py	100.00% <100.00%> (ø)
py-polars/polars/internals/__init__.py	100.00% <100.00%> (ø)
...s/polars-core/src/chunked_array/ops/unique/rank.rs	40.48% <0.00%> (-8.31%)	⬇️
...s/polars-core/src/chunked_array/upstream_traits.rs	69.34% <0.00%> (-2.58%)	⬇️
...olars-lazy/src/physical_plan/executors/scan/csv.rs	96.15% <0.00%> (-1.93%)	⬇️
...lars/polars-core/src/chunked_array/builder/list.rs	62.56% <0.00%> (-1.12%)	⬇️
polars/polars-core/src/series/series_trait.rs	12.28% <0.00%> (-1.06%)	⬇️
...ars/polars-core/src/chunked_array/ops/any_value.rs	77.39% <0.00%> (-0.87%)	⬇️
...olars/polars-core/src/frame/groupby/into_groups.rs	59.88% <0.00%> (-0.30%)	⬇️
... and 4 more

---

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 95074d9...d695639. Read the comment docs.

[matteosantama]

I should link #4044

[stinodego]

Why not include argsort_by, date_range, and _deser_and_exec in this list? These are now imported but not available through __all__.

Same for the other __init__ file, there are a few instances missing from __all__, like Null.

[matteosantama]

Only because mypy didn't complain about them, meaning they are not imported from anywhere else in the codebase. It might better to remove them as imports from the __init__ than to add them to __all__. Thoughts?

[stinodego]

You're right, let's keep that for a different PR!

[stinodego]

Looks good to me!

I mentioned before that I want to avoid using typing.cast if at all possible, but this seems like a valid use case. Nice work.