Significant performance enhancements:
- Use KD-tree for matching observed and reference aperture coordinates
- Implement caching and reuse of zernike polynomial terms when calculating set of slopes
Overall improvement of about a factor of two in performance when doing a full analysis of a typical Shack-Hartmann image.
- Bump
python_requires
to 3.8 - Convert README to
.rst
format - Move dependencies from
all
toinstall_requires
sopip
installs from PyPI pull in everything that's needed.
- The minimum supported Python is now 3.8.
- The minimum supported Astropy is now 5.0.
- The minimum supported Numpy is now 1.18.
- Command-line script for batch re-analysis of archived data.
- Support for MMIRS vignetting model to predict visible spots.
- Support for calculating uncertainties in wavefront model fits.
- Use lenslet optics to calculate spot sizes from diffraction instead of using reference images.
- Fix recentering calculations for Hecto.
- Add support for FLWO engineering wavefront sensors.
- Add configuration for building a container image and automate it in CI.
- Improve reliability of pupil registration.
- Improve reliability in marginal seeing conditions.
- Fix effective wavelengths to be based on more accurate QE and filter response curves.
- Update reference images for MMIRS.
- Various CI fixes, improvements, and eventual migration to GitHub Actions.
- Remove
astropy_helpers
and update/modernize package configuration. - Fix bug in
pup_mask
that was tripped over byscipy >= 1.4
. - Improve background substraction in WFS images.
- Fix DNS resolver to match API changes in upstream package.
- Initial Release