Load the coefficients from Complex spherical harmonics about shtools HOT 8 CLOSED

Unfortunately, you will need to reorder to the coefficients to be in the [2, L_max+1, L_max+1 format. I understand that this is not an efficient storage scheme, but it would be nontrivial to modify all the code to accept a different format.

What we could do is the following:

1. Create some kind of helper function that reorders the coeffiecients to the expected format. We could put it in the utils subpackage.
2. Add some kind of flag to the from_array method allowing the user to specify the format of the array and then convert internally.

I won't have time to work on this for the next couple of months, so if this is something that interests, please put in a PR!

from shtools.

@MarkWieczorek
Thanks for you prompt reply. I am happy to create some helper function to transform the data format.
However, I am not sure how to do such reorder. If we follow the eq(1), there should be only (l_max+1)**2 complex numbers. Can you instruct how to convert them to 2(l_max+1)^2 numbers and then [2, L_max+1, L_max+1] array?

from shtools.

The coeffs array is formatted as [i, l, m] where i=0 is for positive m and i=1 is for negative m. I'll try to add this into the docstrings to make it more obvious.

from shtools.

sorry, I am still confused. If I have 4 coefficients when l_max = 1.
$f_0^0, f_1^{-1}, f_1^{0}, f_1^{1}$,

How shall I arrange them into a [2, 2, 2] array?

from shtools.

$\mbox{coeffs}[0, 0, 0] = f_0^0 $
$\mbox{coeffs}[0, 1, 0] = f_1^0 $
$\mbox{coeffs}[0, 1, 1] = f_1^1 $
$\mbox{coeffs}[1, 1, 1] = f_1^{-1} $

from shtools.

How about the rest elements in the array, e.g. coeffs[1, 0, 0]? Shall we just set them as zero or any number?

from shtools.

Just set them to zero (or any other value you want, it really doesn't matter as these spherical harmonics don't exist...).

from shtools.

I have added additional documentation to the SHCoeffs, SHGravCoeffs and SHMagCoeffs classes as to how the array coeffs is formatted. Hopefully this helps. If you still have questions, feel free to open this issue. (The PR should be merged into develop by the end of the day).

from shtools.