Validate inputs in sample_size_necessary_under_cph / power_under_cph

[Qayad-Ali]

Summary

sample_size_necessary_under_cph and power_under_cph in lifelines.statistics raised cryptic errors on out-of-range inputs. This adds input validation that raises a clear ValueError. No change to valid behavior.

Reproductions (current master)

sample_size_necessary_under_cph(0.8, 1, 0.25, 0.35, 1.0)  # ZeroDivisionError: float division by zero
sample_size_necessary_under_cph(1.5, 1, 0.25, 0.35, 0.7)  # ValueError: cannot convert float NaN to integer

A hazard ratio of 1 (no effect) can't be sized for; out-of-range power/probabilities silently produce NaN.

Changes

• Range checks on both functions (power, alpha in (0,1); probabilities in [0,1]; positive sizes/ratio; hazard ratio > 0). postulated_hazard_ratio == 1 is rejected only in sample_size_necessary_under_cph — it's still valid for power_under_cph.
• Tests covering the invalid inputs and confirming valid output is unchanged: (0.8, 1, 0.25, 0.35, 0.7) == (421, 421).