Sam points out on IRC that overlap_proj works as expected - the issue is with QuadTreeProj. -Nathan On 11/29/12 3:23 PM, Nathan Goldbaum wrote:
Hi all,
This discussion is based on the work we're doing for the enzo testing suite. As part of this, we're doing a number of 1D and 2D enzo simulations. Unfortunately, we're getting some strange results which makes me curious whether it actually makes any sense to project 1D and 2D AMR data in yt. This is motivated by inspection of my data:
In [58]: pf.dimensionality Out[58]: 1
In [59]: prj = pf.h.proj(0,'Density',weight_field='Density') yt : [INFO ] 2012-11-29 15:17:55,860 Projection completed
In [60]: prj['Density'] Out[60]: array([ 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 1. , 2.24475454, 2.54903927, 2.54903927, 2.54903927])
Formally the simulation has no extent along y or z, yet the projection I'm getting back isn't a scalar.
For reference, this is the Toro-3-ShockTubeAMR test that comes with enzo.
I guess the real question I have is what is the best tool inside yt to quantitatively inspect and compare 1D and 2D datasets? Am I making a mistake in the way the test simulation is set up, causing yt to misinterpret it somehow?
Thanks very much for your help with this.
Cheers,
Nathan