Separate Alm class from the atomic scattering kernel

Currently, when creating an Alm array, it also (re)compiles the CL kernel to add scattering from atoms with a given x,y,z and form factor, and initializes the array with zeros. The Alm class should be separate from this operation, to allow e.g. returning an Alm from the ROTALM function (#6 (closed)) or loading an Alm from a file (#1).

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information