Full Changelog#

1.0.0 (2026-04-14)#

Updated the minimum supported dependencies for demregpy 1.0. demregpy now targets Python 3.12 and newer, uses NumPy 2 or newer, and adds matplotlib as a core dependency. (#21)

Extended demregpy.dn2dem() to accept channel-last inputs with up to three leading axes, including stacks such as (ntime, nx, ny, nf). The weighting inputs are also more flexible, including support for per-filter gloci masks and broadcasting a single dem_norm0 curve across larger inputs, and invalid shapes or response matrices now raise explicit errors earlier in the solve. (#21)
Added small public helper APIs for common workflows, including demregpy.load_aia_response() for bundled AIA response files, demregpy.plotting.plot_dem() for one-dimensional DEM plots, and demregpy.synthetic.synthesize_counts() for generating synthetic observations from response matrices. (#21)
Improved inversion performance substantially, with typical speedups of around 4 to 7 times depending on the input shape and solver settings. (#21)
Added demregpy.plotting.plot_loci_curves() for plotting filter loci curves on their own or overplotting them on a DEM axis. The helper supports bundled AIA responses and synthetic response matrices, and uses the same DEM-space scaling as demregpy.dn2dem() by default so loci curves can be compared directly with recovered DEMs. (#22)

Reworked the documentation and example gallery around smaller, focused examples instead of a few monolithic scripts. The updated docs now cover installation, a dn2dem tutorial, topic guides, AIA response tooling, synthetic-count generation, and a wider range of runnable gallery examples. (#21)
Added synthetic and AIA gallery examples showing how to plot loci curves and how to overlay them on recovered DEMs. (#22)