add running mean power line (yellow)

opt = dict()

def running_mean(x: np.ndarray, N: int) -> np.ndarray:

    """

    Compute `N` elements wide running average over `x`.

    :param x: 1-Dimensional NumPy array

    :param N: how many items to average. Should be even for optimal results.

    """

    # to ensure that output.shape == input.shape, we need to insert data

    # at the boundaries

    boundary_array = np.insert(x, 0, np.full((N // 2), x[0]))

    boundary_array = np.append(boundary_array, np.full((N // 2 + N % 2), x[-1]))

    print(boundary_array)

    cumsum = np.cumsum(boundary_array)

    return (cumsum[N:] - cumsum[:-N]) / N

def measure_data(filename, duration):

    # libmsp430.so must be available

    if not 'LD_LIBRARY_PATH' in os.environ:

        print('Average per-peak energy (delta over baseline): {:.0f} µJ         : {:.9f}'.format(

            delta_energy * 1e6 / len(peaks), delta_energy / len(peaks)))

    power_from_current = data[:, 1] * data[:, 2] * 1e-12

    power_from_energy = ((data[1:, 3] - data[:-1, 3]) * 1e-9) / ((data[1:, 0] - data[:-1, 0]) * 1e-6)

    mean_power = running_mean(power_from_energy, 10)

    if 'stat' in opt:

        mean_voltage = np.mean(data[:, 2])

    if 'plot' in opt:

        # nA * mV = pW

        pwrhandle, = plt.plot(data[:, 0] * 1e-6, data[:, 1] * data[:, 2] * 1e-12, 'b-', label='P=U*I', markersize=1)

        energyhandle, = plt.plot(data[1:, 0] * 1e-6, ((data[1:, 3] - data[:-1, 3]) * 1e-9) / ((data[1:, 0] - data[:-1, 0]) * 1e-6), 'r-', label='P=ΔE/Δt', markersize=1)

        pwrhandle, = plt.plot(data[:, 0] * 1e-6, power_from_current, 'b-', label='P=U*I', markersize=1)

        energyhandle, = plt.plot(data[1:, 0] * 1e-6, power_from_energy, 'r-', label='P=ΔE/Δt', markersize=1)

        energyhandle2, = plt.plot(data[1:, 0] * 1e-6, mean_power, 'y-', label='E/Δt', markersize=1)

        plt.legend(handles=[pwrhandle, energyhandle])

        plt.xlabel('Time [s]')

        plt.ylabel('Power [W]')