FWHM (X)

@staticmethod
def evaluate(image: np.ndarray) -> float:
    """
    Calculate the FWHM of an image across the x axis, at the y position of the image
    centroid. Errors if waveform or scalar provided.

    First, calculates the centroid and extracts the row at that position. Then
    applies smoothing by taking weighted nearest and next-nearest neighbour
    contributions to pixel values whose difference from their neighbours is more
    than 0.2 times the total range in values.

    The global maximum value is then identified along with the first
    positions to either side of the peak where the pixel value is less than
    or equal to half the global maximum. If all pixel values to one side of
    the peak are greater than half the global maximum, the position of the
    global maximum is used. The difference between these two positions in
    number of pixels is returned.
    """
    centroid_position = CentroidY.evaluate(image)
    y = image[centroid_position]
    waveform_variable = WaveformVariable(x=np.arange(len(y)), y=y)

    half_max_left, half_max_right = FWHM.calculate_fwhm(waveform_variable)
    return half_max_right - half_max_left