def equi_freq()

in rules/predicate.py [0:0]

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

def equi_freq(values: Set, num_bins: int):
    cutoffs = set()
    sorted_values = sorted(values)

    num_values = len(sorted_values)

    # number of bins > number of unique values
    if num_bins > num_values:
        # Use square-root choice
        num_bins = math.ceil(math.sqrt(num_values))

    values_ratio = int(num_values / num_bins)

    for i in range(0, num_bins):
        arr = []
        for j in range(i * values_ratio, (i + 1) * values_ratio):
            if j >= num_values:
                break
            arr = arr + [sorted_values[j]]

        cutoff = arr[len(arr) - 1]
        cutoffs.add(cutoff)

    return cutoffs