def _optimal_kappa()

in spotify_confidence/analysis/frequentist/confidence_computers/sample_size_computer.py [0:0]

• 16 lines of code
• 4 McCabe index (conditional complexity)

def _optimal_kappa(row: Series, is_binary_column) -> float:
    def _binary_variance(p: float) -> float:
        return p * (1 - p)

    if row[is_binary_column]:
        if is_non_inferiority(row[NIM]):
            return 1.0
        else:
            if row[POINT_ESTIMATE] == 0.0:
                # variance will be 0 as well in this case. This if-branch is important to avoid divide by zero problems
                return 1.0
            else:
                hypothetical_effect = row[ALTERNATIVE_HYPOTHESIS] - row[NULL_HYPOTHESIS]
                return np.sqrt(
                    _binary_variance(row[POINT_ESTIMATE]) / _binary_variance(row[POINT_ESTIMATE] + hypothetical_effect)
                )
    else:
        return 1.0