feat: Add DataType inference from Python types

[jaychia]

Closes #3549

[codspeed-hq]

CodSpeed Performance Report

Merging #3555 will degrade performances by 12.53%

_{Comparing jay/type-convenience (0f8be2c) with main (6ae4e77)}

Summary

❌ 1 regressions
✅ 26 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	jay/type-convenience	Change
❌	test_iter_rows_first_row[100 Small Files]	185.1 ms	211.7 ms	-12.53%

[andrewgazelka]

perhaps could test

• dict (our DataType::Map)
• tuple (our DataType::Struct I think with just arbitrary names)
• bytearray

[andrewgazelka]

trust your judgement here though :)

[jaychia]

Oh yes the dict map one is a good idea. Let me add that.

[andrewgazelka]

I wonder if we could create a type alias in pythong for DataType | type as it seems to be used in a lot of places

[jaychia]

I remember documentation being a bit messy when I used type aliases for something else. Let me give it a shot again and see how it shows up on docs.

[andrewgazelka]

ah I forgot how it shows up in documentation... yea I guess I am thinking like a Rust user. it might be complicated for something that isn't statically typed having to go deeper in docs 🤔. Trust your judgement here.

[codecov]

Codecov Report

Attention: Patch coverage is 92.50000% with 3 lines in your changes missing coverage. Please review.

Project coverage is 77.79%. Comparing base (f23ee37) to head (0f8be2c).
Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
daft/datatype.py	90.32%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3555      +/-   ##
==========================================
+ Coverage   77.69%   77.79%   +0.09%     
==========================================
  Files         710      716       +6     
  Lines       86941    87801     +860     
==========================================
+ Hits        67552    68305     +753     
- Misses      19389    19496     +107     

Files with missing lines	Coverage Δ
daft/expressions/expressions.py	93.60% <100.00%> (+0.01%)	⬆️
daft/udf.py	95.58% <100.00%> (+0.02%)	⬆️
daft/datatype.py	91.73% <90.32%> (-0.16%)	⬇️

... and 41 files with indirect coverage changes