fix: make skip_partial_aggregation_probe_ratio_threshold match the docs

[haohuaijin]

Which issue does this PR close?

• Closes #.

Rationale for this change

The config skip_partial_aggregation_probe_ratio_threshold was documented as triggering skip when the ratio is greater than the threshold, but the code used >=. This meant setting the threshold to 1.0 (to disable the feature) still skipped rows when cardinality was exactly 100%.

What changes are included in this PR?

• Changed >= to > in the ratio comparison to match the docs.
• Return None for SkipAggregationProbe when probe_ratio_threshold >= 1.0, effectively disabling the feature since the ratio can never exceed 1.0.

Are these changes tested?

Yes. Added test_skip_aggregation_disabled_at_threshold_one which sets threshold to 1.0 with 100% cardinality input and asserts that no rows are skipped.

Are there any user-facing changes?

Yes. Setting skip_partial_aggregation_probe_ratio_threshold = 1.0 now reliably disables skip aggregation, matching the documented behavior.

[kosiew]

@haohuaijin
Thanks for the fix. I do not see any blocking issues, just a couple of small suggestions that could make the regression coverage and config setup a bit clearer.

[haohuaijin]

Thanks for you reviews @kosiew , apply suggestion in 0853632

[haohuaijin]

the failed should not related to my change, let me retrigger the ci to see.