[CHORE] Remove user-facing arguments for casting to Ray's tensor type

[jaychia]

Summary

Cleanup PR.

1. Removes cast_tensors_to_ray_tensor_dtype as a user-facing argument in our export methods (e.g. to_arrow, to_pandas etc) -- this is really only intended to be used when a user is converting a Daft dataframe to a Ray dataset anyways and there isn't a need to expose this functionality to a user
2. Instead, the logic for casting daft.DataType.tensor data to a Ray Data tensor type is done inside of the conversion code for Ray Data (_make_ray_block_from_micropartition). This lets us contain the ickiness of that code without having it touch all of our to_arrow logic
3. Also removes _trim_pyarrow_large_arrays which was a legacy codepath that doesn't get hit anymore

[codspeed-hq]

CodSpeed Performance Report

Merging #2802 will degrade performances by 13.33%

_{Comparing jay/arrow-encode-decode (b2a1e6b) with main (e3fbf88)}

Summary

⚡ 1 improvements
❌ 1 regressions
✅ 14 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	main	jay/arrow-encode-decode	Change
❌	test_count[1 Small File]	20.5 ms	23.6 ms	-13.33%
⚡	test_show[100 Small Files]	298.8 ms	50.9 ms	×5.9

[jaychia]

I omitted this logic in this refactor because I have no idea what this is doing. Also there aren't any tests to help me understand so 🤷

[jaychia]

Actually, added this back in to pass tests

[codecov]

Codecov Report

Attention: Patch coverage is 96.07843% with 2 lines in your changes missing coverage. Please review.

Project coverage is 63.11%. Comparing base (6fe408c) to head (b2a1e6b).
Report is 5 commits behind head on main.

Files with missing lines	Patch %	Lines
daft/runners/ray_runner.py	91.30%	2 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #2802    +/-   ##
========================================
  Coverage   63.11%   63.11%            
========================================
  Files        1008     1007     -1     
  Lines      114269   114135   -134     
========================================
- Hits        72117    72038    -79     
+ Misses      42152    42097    -55     

Files with missing lines	Coverage Δ
daft/dataframe/dataframe.py	86.05% <100.00%> (+0.04%)	⬆️
daft/datatype.py	91.10% <100.00%> (ø)
daft/runners/partitioning.py	81.33% <100.00%> (ø)
daft/series.py	89.50% <100.00%> (-0.03%)	⬇️
daft/table/micropartition.py	91.07% <100.00%> (ø)
daft/table/table.py	60.56% <100.00%> (+1.36%)	⬆️
src/daft-core/src/python/datatype.rs	81.29% <100.00%> (-0.62%)	⬇️
daft/runners/ray_runner.py	88.03% <91.30%> (+0.12%)	⬆️

... and 22 files with indirect coverage changes