getionimage performance issue for multiple mz values

[leorrose]

I want to get ion image for multiple mz value (794.5, 834.5, 888.6) and not one mz value.

Using the existing getionimage method i need to do the following:

img_1= getionimage(p, 794.5, tol=0.5)
img_2 = getionimage(p, 834.5, tol=0.5)
img_3 = getionimage(p, 888.6, tol=0.5)

im_combined = img_1 + img_2 + img_3 

This works but the issue is that it takes a lot of time because we go over the imzML file 3 times.

Is there a better solution to get ion image for multiple mz values?

[leorrose]

I found that this can be achieved by changing the geionimage to accept multiple mz values like this:

def getionimage(p, mz_values, tol=0.1, z=1, reduce_func=sum):
    """
    Get an image representation of the intensity distribution
    of the ion with specified m/z values.

    By default, the intensity values within the tolerance region are summed.

    :param p:
        the ImzMLParser (or anything else with similar attributes) for the desired dataset
    :param mz_values:
        m/z values for which the ion image shall be returned
    :param tol:
        Absolute tolerance for the m/z value, such that all ions with values
        mz_value-|tol| <= x <= mz_value+|tol| are included. Defaults to 0.1
    :param z:
        z Value if spectrogram is 3-dimensional.
    :param reduce_func:
        the bahaviour for reducing the intensities between mz_value-|tol| and mz_value+|tol| to a single value. Must
        be a function that takes a sequence as input and outputs a number. By default, the values are summed.

    :return:
        numpy matrix with each element representing the ion intensity in this
        pixel. Can be easily plotted with matplotlib
    """
    tol = abs(tol)
    im = np.zeros((p.imzmldict["max count of pixels y"], p.imzmldict["max count of pixels x"]))
    for i, (x, y, z_) in enumerate(p.coordinates):
        if z_ == 0:
            UserWarning("z coordinate = 0 present, if you're getting blank images set getionimage(.., .., z=0)")
        if z_ == z:
            mzs, ints = map(lambda x: np.asarray(x), p.getspectrum(i))
            partial_ints = []
            for mz_value in mz_values:
                min_i, max_i = _bisect_spectrum(mzs, mz_value, tol)
                partial_ints.append(ints[min_i:max_i + 1])
            im[y - 1, x - 1] = reduce_func(np.concatenate(partial_ints))
    return im

I added a pull request containing these changes if someone wants to check this out enable multiple MZ values in getionimage