refactor(python): sync the compression argument of sink_ipc and read_ipc

[eitsupi]

compression="uncompressed" is allowed for scan_ipc().

[ritchie46]

This is breaking? And I think we should discuss which direction we want to sync. I like None as uncompressed in python.

[eitsupi]

This is breaking? And I think we should discuss which direction we want to sync. I like None as uncompressed in python.

No, this is not a breaking change.
None will continue to be allowed. The only change is that the new "uncompressed" will be allowed in sink_ipc().

Edit: Sorry, I made a different change than intended. I will correct it.

[codecov]

Codecov Report

Attention: Patch coverage is 50.00000% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 81.30%. Comparing base (7341aee) to head (0718087).
Report is 2 commits behind head on main.

Files	Patch %	Lines
py-polars/polars/lazyframe/frame.py	0.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15644      +/-   ##
==========================================
- Coverage   81.30%   81.30%   -0.01%     
==========================================
  Files        1373     1373              
  Lines      176243   176245       +2     
  Branches     2544     2545       +1     
==========================================
- Hits       143293   143292       -1     
- Misses      32466    32469       +3     
  Partials      484      484              

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

[stinodego]

And I think we should discuss which direction we want to sync. I like None as uncompressed in Python.

I don't think using None is a good idea here. None usually indicates unspecified, which means 'use a default' or 'infer the proper value'. Explicitly passing "uncompressed" is much more clear. I think we should at least support passing "uncompressed".

This is already the case for most write functions on the Python side. None is implicitly supported by converting to "uncompressed" internally, which is what this PR is also trying to do, hence no breaking change here.

On the Rust side, ParquetCompression already includes Uncompressed in the Enum. I think that makes sense for the same reason stated above. If you agree, I think we should update IpcCompression and AvroCompression to match.