Skip to content

v0.4.0
Compare
Choose a tag to compare
@ehpor ehpor released this 23 Feb 01:14
· 595 commits to master since this release
v0.4.0
b542f29
This commit was created on github.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

HCIPy 0.4.0 Release notes

This HCIPy release provides significant speed improvements for Fourier transforms, propagations and the vortex coronagraphs, among some other minor improvements and bugfixes. See the list of all changes below for a summary. We dropped support for Python 3.5. This version supports Python 3.6+.

List of major changes

  • Major bug fix in caching algorithm for optical elements (@ehpor)
  • Major improvement in computational efficiency of FFTs and MFTs (@ehpor)
  • FFTs now use MKL when installed (@ehpor)
  • Fourier transforms now retain bit depth and do not automatically cast to double precision (@ehpor)
  • A new class FourierFilter for efficient correlations and convolutions (@ehpor)
  • The vortex coronagraphs now use smooth windowing for their multiscale Fourier transforms (@ehpor)
  • Added VLT aperture (@dsdoelman, @ehpor, @syhaffert & @jmilou)
  • The perfect coronagraph can now handle polarized wavefronts (@ehpor)
  • Optical elements can now be pickled for multiprocessing (@ehpor)
  • Detectors can now do subsampling of the incoming light (@spbos & @ehpor)
  • Fixed the OD wavefront sensors (@syhaffert)
  • Fixed bug for the ZernikeWavefrontSensor when physical units were used (@yinzi-xin & @ehpor)
  • Two new tutorials (@jmilou & @ehpor)
  • Tutorial notebooks are now allowed to use up to 10mins per cell (@ehpor)
  • Added support for Python 3.8 and 3.9 (@ehpor)
  • Migration to Azure Pipelines for CI tests on all operating systems (@ehpor)
  • Miscellaneous minor bug fixes (@syhaffert, @jmilou & @ehpor)
Assets 4
Loading