In numpy, if I want a random array uniformly distributed between -0.6 and +0.6 with 100 elements, I'd write:

rng = numpy.random.RandomState(seed)
array = rng.uniform(-0.6, 0.6, 100)

Using GalSim deviates, I'd do something like:

ud = galsim.UniformDeviate(seed)
array = np.empty(100)
ud.generate(array)
array = array*1.2 - 0.6

I find the former easier to read. I d

Currently psf.draw(x,y) centers the PSF profile at the sub-pixel values of the provided x,y positions.

This is sometimes a desirable feature, but not always. We should make it easier for the user to specify how they want the PSF image centered. I'm proposing a new center keyword argument to the draw command with the following options:

center=None   # (default) the current behavior u

--runall # if not set, only the main file
# if set, runs the other metacal files (1p,1m,2p,2m), totally 5 times
--head # uses only random subset of source catalog
--nopairs # does not log source-lens pairs
--noclust # skip clusters
--norands # skip randoms
--params PARAMS # forces to use config file at path PARAMS
--metachunk METACHUNK # reruns the i-th

We currently track two coordinate systems: Galactic (gal) and celestial (cel). Celestial just means "sky" and doesn't actually contain information about coordinate frame. The actual system is the equatorial coordinate system, so we should update instances of cel to equ. This is a trivial change, but permeates the codebase.