The weighted variance stamps were causing a segfault when the number of returned stamps was less than the number of images. This could happen because of use_index which is set by the sigma-G filtering. The fix is to set num_images to the number of returned stamps.
This PR includes a few cleanups found during the debugging, including adding some docstring comments and fixing a numpy conversion error on old (v0.5) data.
The weighted variance stamps were causing a segfault when the number of returned stamps was less than the number of images. This could happen because of
use_indexwhich is set by the sigma-G filtering. The fix is to setnum_imagesto the number of returned stamps.This PR includes a few cleanups found during the debugging, including adding some docstring comments and fixing a numpy conversion error on old (v0.5) data.