feat(scantask-2): Implement new module for splitting Parquet ScanTask

[jaychia]

Implements a new module daft-scan/src/scan_task_iters/split_parquet which contains all the functionality for splitting an iterator of ScanTasks into smaller ScanTasks if they are Parquet files.

The public interface is SplitParquetScanTasks, which under the hood uses private functionality in the form of DecideSplitIterator -> RetrieveParquetMetadataIterator -> flatten to produce a new (split) iterator of ScanTasks.

[codspeed-hq]

CodSpeed Performance Report

Merging #3628 will degrade performances by 41.5%

_{Comparing jay/split-all-files-iter-2 (18dc97b) with main (836297a)}

Summary

❌ 2 regressions
✅ 25 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	main	jay/split-all-files-iter-2	Change
❌	test_iter_rows_first_row[100 Small Files]	110.4 ms	188.7 ms	-41.5%
❌	test_show[100 Small Files]	15.7 ms	18.9 ms	-16.95%

[codecov]

Codecov Report

Attention: Patch coverage is 11.32075% with 47 lines in your changes missing coverage. Please review.

Project coverage is 78.04%. Comparing base (ff2619b) to head (18dc97b).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...task_iters/split_parquet/fetch_parquet_metadata.rs	0.00%	17 Missing ⚠️
...daft-scan/src/scan_task_iters/split_parquet/mod.rs	0.00%	10 Missing ⚠️
src/daft-scan/src/scan_task_iters/mod.rs	40.00%	9 Missing ⚠️
...task_iters/split_parquet/split_parquet_decision.rs	0.00%	9 Missing ⚠️
...can_task_iters/split_parquet/split_parquet_file.rs	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3628      +/-   ##
==========================================
+ Coverage   77.28%   78.04%   +0.76%     
==========================================
  Files         721      725       +4     
  Lines       90549    89217    -1332     
==========================================
- Hits        69978    69633     -345     
+ Misses      20571    19584     -987     

Files with missing lines	Coverage Δ
...can_task_iters/split_parquet/split_parquet_file.rs	0.00% <0.00%> (ø)
src/daft-scan/src/scan_task_iters/mod.rs	91.28% <40.00%> (-2.68%)	⬇️
...task_iters/split_parquet/split_parquet_decision.rs	0.00% <0.00%> (ø)
...daft-scan/src/scan_task_iters/split_parquet/mod.rs	0.00% <0.00%> (ø)
...task_iters/split_parquet/fetch_parquet_metadata.rs	0.00% <0.00%> (ø)

... and 19 files with indirect coverage changes