Add pyarrow dependency and use pyarrow backed dtypes

[gsheni]

No description provided.

[codecov]

Codecov Report

Patch coverage: 96.00% and project coverage change: +0.02% 🎉

Comparison is base (a1f3bcf) 90.01% compared to head (2457439) 90.03%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #120      +/-   ##
==========================================
+ Coverage   90.01%   90.03%   +0.02%     
==========================================
  Files          24       24              
  Lines         871      893      +22     
==========================================
+ Hits          784      804      +20     
- Misses         87       89       +2     

Files Changed	Coverage Δ
trane/ops/aggregation_ops.py	98.57% <87.50%> (-1.43%)	⬇️
trane/typing/logical_types.py	93.05% <92.85%> (-1.39%)	⬇️
trane/core/prediction_problem.py	84.00% <100.00%> (+0.43%)	⬆️
trane/core/prediction_problem_generator.py	91.57% <100.00%> (-0.82%)	⬇️
trane/core/utils.py	86.32% <100.00%> (+0.35%)	⬆️
trane/datasets/load_functions.py	100.00% <100.00%> (ø)
trane/ops/threshold_functions.py	100.00% <100.00%> (ø)
trane/typing/inference_functions.py	81.75% <100.00%> (+1.59%)	⬆️

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