fix: to_cnf and to_dnf functions

[kevinzwang]

The pre-order expression tree traversal algorithm I used for to_dnf turned out not to actually fully convert a statement to disjunctive normal form. Since it was also used in to_cnf, that did not fully work either.

This new algorithm recursively converts subtrees to DNF form and propagates that upwards. In each boolean AND, the subtrees are assumed to be fully DNF, and it uses that assumption to merge the two sides into one DNF statement.

Also a quick fix for de Morgan's law on the case of triple negatives.

[codspeed-hq]

CodSpeed Performance Report

Merging #3728 will degrade performances by 36.53%

_{Comparing kevin/to-dnf-fix (5ae59bf) with main (fec399e)}

Summary

❌ 2 regressions
✅ 25 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	test_iter_rows_first_row[100 Small Files]	188.1 ms	296.4 ms	-36.53%
❌	test_show[100 Small Files]	15.9 ms	23.9 ms	-33.64%

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 77.58%. Comparing base (603199f) to head (5ae59bf).
Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3728      +/-   ##
==========================================
+ Coverage   77.55%   77.58%   +0.03%     
==========================================
  Files         729      729              
  Lines       91917    92005      +88     
==========================================
+ Hits        71284    71383      +99     
+ Misses      20633    20622      -11     

Files with missing lines	Coverage Δ
src/daft-algebra/src/boolean.rs	100.00% <100.00%> (ø)
...al-plan/src/optimization/rules/push_down_filter.rs	97.35% <ø> (ø)

... and 5 files with indirect coverage changes

[universalmind303]

approving, but i'd prefer to merge the mod/tests into single file.