v0.2.0

This release makes GPfit object-oriented, with fit-type-specific sub-classes (e.g. SoftmaxAffine) derived from a _Fit parent class. These classes have print and plot methods (plot_surface and plot_slices for 2D fits). They also include the ability to save fits, both to a pickle file and to a text file. A convenience function fit remains for one-line creation of fits.

Finally, the documentation also gets a facelift, a restructure, installation instructions and a tutorial page.

v0.1.0

• Python 3 only
• General code tightening (more uniform and concise code)
• Deleted legacy Matlab code
• Better variable naming
• Better function naming
• Better class naming
• Code restructuring
  • Grouping similar functions into modules
  • Separate xfoil sub-directory
  • Remove examples/tests redundancy
• Better test coverage
• Uploaded to PyPI (https://pypi.org/project/gpfit/0.1.0/)

v0.0.0 - OG GPfit

This project originally started as a library of MATLAB code written by @whoburg during his PhD for fitting GP-compatible functions to data. This repository was created when the code was ported to python in late 2014/early 2015. Though some additional features have been added in the intervening 6.5 years, the code in this release has much of the same structure and naming as the original MATLAB code. It has had some refactors/cleanups over the years but it still very much resembles a project developed mostly by graduate students, some of whom were very new to python when they created it :)