[BUG] Fix ScanTask memory estimations when limits are provided

[jaychia]

Fixes memory estimations for ScanTask when limits are pushed down

• Adds a newapprox_num_rows API to ScanTask, which is distinct from the num_rows API which is expected to provide an exact number of rows if available
• Adds some heuristics for determining the approximate number of rows with the available information, including accounting for limits
• Fixes num_rows to account for limit pushdowns as well

[codspeed-hq]

CodSpeed Performance Report

Merging #2735 will degrade performances by 58.47%

_{Comparing jay/fix-memory-estimations-limits-scantask (06c4122) with main (5eb34fc)}

Summary

⚡ 1 improvements
❌ 2 regressions
✅ 13 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	main	jay/fix-memory-estimations-limits-scantask	Change
❌	test_count[1 Small File]	10.1 ms	24.2 ms	-58.47%
⚡	test_explain[100 Small Files]	41.5 ms	37.1 ms	+11.95%
❌	test_show[100 Small Files]	396.6 ms	659 ms	-39.81%

[jaychia]

Interestingly, this is pretty much never used in our codebase... And is also the only place we really use size_bytes_on_disk

[samster25]

Shouldn't this be using estimate_in_memory_size_bytes since this is what we use in the the Scheduling?

[codecov]

Codecov Report

Attention: Patch coverage is 84.00000% with 8 lines in your changes missing coverage. Please review.

Please upload report for BASE (main@20ffed4). Learn more about missing BASE report.
Report is 9 commits behind head on main.

Files	Patch %	Lines
src/daft-scan/src/lib.rs	85.10%	7 Missing ⚠️
src/daft-plan/src/physical_ops/scan.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2735   +/-   ##
=======================================
  Coverage        ?   63.33%           
=======================================
  Files           ?      981           
  Lines           ?   113277           
  Branches        ?        0           
=======================================
  Hits            ?    71748           
  Misses          ?    41529           
  Partials        ?        0           

Files	Coverage Δ
daft/execution/execution_step.py	91.96% <ø> (ø)
src/daft-scan/src/python.rs	69.68% <100.00%> (ø)
src/daft-plan/src/physical_ops/scan.rs	10.11% <0.00%> (ø)
src/daft-scan/src/lib.rs	60.33% <85.10%> (ø)

[samster25]

Shouldn't this be using estimate_in_memory_size_bytes since this is what we use in the the Scheduling?