def numeric_sparse()

in utils_data.py [0:0]


def numeric_sparse(queries, skip_idxs, k, T, epsilon, sens, delta=0):

    count = 0
    idxs = onp.array([], onp.int)
    pos = 0

    sigma = compute_sigma(k, epsilon, sens, delta)

    T_hat = T + onp.random.laplace(loc=0, scale=sigma)

    while pos < len(queries) and len(idxs) < k:

        if pos in skip_idxs:
            pos += 1
            continue

        v_i = onp.random.laplace(loc=0, scale=2 * sigma)

        if queries[pos] + v_i >= T_hat:
            idxs = onp.append(idxs, pos)
            count += 1
            T_hat = T + onp.random.laplace(loc=0, scale=sigma)

        pos += 1

    return idxs