1.1.0

We are now starting the changes suggested in the pyOpenSci review.

A lot of the changes since the last release are bugfixes and documentation updates. The non-documentation, non-bugfix changes are:

• Images can now be accessed using the plenoptic.data module (e.g., po.data.einstein()), which means that the example notebooks can be downloaded and run directly!
• The Portilla-Simoncelli model has been refactored and made more efficient!
• We're now on conda and can be installed with conda install plenoptic -c conda-forge (no Windows support for conda, unfortunately).
• All FrontEnd models (e.g., po.simul.LinearNonlinear) now have a pretrained argument, which loads in the parameters used in Alex Berardino's thesis.
• Deprecates a variety of functions and objects. Most importantly, Geodesic is deprecated and will be removed soon, because it needs more work to be as stable and robust as the other methods. Work will continue in the new geodesics repo.
• The included models now use register_buffers, so that you can call e.g.,

  ps = po.simul.PortillaSimoncelli((256, 256))
  ps.cuda()

  (in addition to ps.to('cuda')

What's Changed

• Add PortillaSimoncelliMinimal to tutorial get original paper's statistics by @dherrera1911 in #216
• Adds info about pip minimum version by @billbrod in #228
• Citation.cff by @billbrod in #231
• Updates installation instructions and CONTRIBUTING by @billbrod in #230
• Switch models to using register_buffers by @billbrod in #232
• Citation update by @billbrod in #249
• Fix failing urlchecker by using nbstripout by @billbrod in #253
• Adds description of torch/pyrtools dependencies by @billbrod in #248
• fix for test_stop_criterion by @billbrod in #251
• Adds repo status badge by @billbrod in #255
• Clarify what model means by @billbrod in #250
• Small citation update by @billbrod in #256
• Updates action versions, change how notebooks are tested by @billbrod in #257
• Updates python versions and trove classifiers by @billbrod in #254
• 235 include example data by @BalzaniEdoardo in #252
• Refactor Portilla-Simoncelli model by @billbrod in #225
• Update deploy by @billbrod in #258
• Torch tensor replacements by @hmd101 in #260
• Update Deprecated PyTorch Functions to torch.linalg Namespace by @hmd101 in #261
• Fix Minor Typos in Documentation by @hmd101 in #267
• small update to quickstart text by @billbrod in #272
• Updates matplotlib config for animations by @billbrod in #274
• Fix links in documentation index, updates CCN logo by @billbrod in #275
• Dynamically Adjust Number of Columns in plot_representation Based on Number of Loss Plots by @hmd101 in #276
• Adds conda install instructions by @billbrod in #278
• adds Hanna to citation.cff by @billbrod in #280
• Avoid unnecessary computation of magnitude variance in Portilla-Simoncelli model by @billbrod in #281
• adds Edoardo to citation by @billbrod in #282
• Updates github org name by @billbrod in #285
• Fix jenkins by @billbrod in #286
• Small restructure of how we use importlib.resources by @billbrod in #289
• Make penalize_range more efficient by @billbrod in #287
• Deprecations by @billbrod in #288
• Specify docutils version by @billbrod in #290
• Improvements to FrontEnd models by @billbrod in #291
• Fix broken to method, test non-standard models/metrics, don't assume dtype by @billbrod in #292
• Updates binder link to 1.1.0 by @billbrod in #293

New Contributors

• @dherrera1911 made their first contribution in #216
• @BalzaniEdoardo made their first contribution in #252
• @hmd101 made their first contribution in #260

Full Changelog: 1.0.2...1.1.0