h5cube v0.2
See CHANGELOG.txt.
_Note:_ The changelog is incorrect about the zero thresholding behavior, per #51. Zero values remain zero after compression/decompression, due to the 0.0
value stored in the SIGNS
dataset. So, isosurface data shouldn't be affected at all; just be aware of the actual behavior if performing volumetric analysis.
In any event, again per #51, a fix is pending.
UPDATE: v0.3, with this problem believed-fixed and the --zero
thresholding option of #21 implemented, will be shipping as soon as I can run it down the release pathway.