fix: correct pvalue logic and add discretization step

[drbh]

This PR supersedes #53 and ensures that the experimental output is identical to the current output with a relative accuracy of <1e-6 and absolute accuracy of <1e-12.

Additionally this PR renames and improves the tests/bench_expr.py file. The new tests/test_experimental_correctness.py file will run with the other test (prefixed with test_) and compares the output between the reference and experimental outputs to ensure they are within a extremely small tolerance

note these changes slow down the overall computation by a small amount (mainly when the discretization step is needed) which can be avoid but providing a the raw integer data

uv run tests/test_experimental_correctness.py

======================================================== test session starts ========================================================
platform darwin -- Python 3.10.15, pytest-8.4.1, pluggy-1.6.0
benchmark: 5.1.0 (defaults: timer=time.perf_counter disable_gc=False min_rounds=5 min_time=0.000005 max_time=1.0 calibration_precision=10 warmup=False warmup_iterations=100000)
rootdir: /Users/drbh/Projects/pdex
configfile: pyproject.toml
plugins: benchmark-5.1.0
collected 9 items

tests/test_experimental_correctness.py .........                                                                              [100%]

========================================================= warnings summary ==========================================================
tests/test_experimental_correctness.py::test_correctness_comparison
tests/test_experimental_correctness.py::test_detailed_correctness_metrics
tests/test_experimental_correctness.py::test_edge_cases_numerical_stability
tests/test_experimental_correctness.py::test_edge_cases_numerical_stability
  /Users/drbh/Projects/pdex/.venv/lib/python3.10/site-packages/anndata/_core/aligned_df.py:68: ImplicitModificationWarning: Transforming to str index.
    warnings.warn("Transforming to str index.", ImplicitModificationWarning)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html

------------------------------------------------------------------------------------------------------------ benchmark: 6 tests ------------------------------------------------------------------------------------------------------------
Name (time in ms)                                                    Min                   Max                  Mean             StdDev                Median                IQR            Outliers       OPS            Rounds  Iterations
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_benchmark_parameterized_datasets[True-500-100-10]            7.3009 (1.0)          8.3150 (1.0)          7.9079 (1.0)       0.4193 (1.0)          7.9405 (1.0)       0.6697 (1.0)           1;0  126.4562 (1.0)           5           1
test_benchmark_parameterized_datasets[True-1000-300-100]         81.2446 (11.13)       83.1028 (9.99)        82.1008 (10.38)     0.7036 (1.68)        82.1472 (10.35)     0.9652 (1.44)          2;0   12.1802 (0.10)          5           1
test_benchmark_parameterized_datasets[True-2000-500-50]         113.6737 (15.57)      124.0537 (14.92)      118.8101 (15.02)     4.5535 (10.86)      119.7459 (15.08)     8.1712 (12.20)         2;0    8.4168 (0.07)          5           1
test_benchmark_parameterized_datasets[False-500-100-10]       1,121.4055 (153.60)   1,163.2690 (139.90)   1,146.4066 (144.97)   15.3490 (36.61)    1,150.1841 (144.85)   13.6819 (20.43)         2;0    0.8723 (0.01)          5           1
test_benchmark_parameterized_datasets[False-2000-500-50]      3,598.1983 (492.84)   3,636.3652 (437.33)   3,620.5203 (457.84)   14.3509 (34.23)    3,620.4431 (455.95)   17.2623 (25.78)         2;0    0.2762 (0.00)          5           1
test_benchmark_parameterized_datasets[False-1000-300-100]     3,952.2674 (541.34)   4,037.7101 (485.59)   3,976.9548 (502.91)   34.6742 (82.70)    3,966.9077 (499.58)   30.1742 (45.06)         1;1    0.2514 (0.00)          5           1
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Legend:
  Outliers: 1 Standard Deviation from Mean; 1.5 IQR (InterQuartile Range) from 1st Quartile and 3rd Quartile.
  OPS: Operations Per Second, computed as 1 / Mean
============================================= 9 passed, 4 warnings in 71.21s (0:01:11) ==============================================