treegp
is a python gaussian process code that perform 1D and 2D interpolation.
treegp
has some special features compared to other available Gaussian Processes codes:
- Hyperparameters estimation will scale in O(N log(N)) with the the 2-points correlation function estimation compared to O(N^3) with the classical maximum likelihood.
- Gaussian process interpolation can be performed around a mean function
- A tool is provided to compute the mean function (
meanify
)
treegp
was originally developed for Point Spread Function interpolation within Piff. There is a specific article that describes the math used in treegp
in the context of modelling astrometric shifts of the Subaru Telescope due to atmospheric turbulences. This article can be found
here.
The easiest way to install is usually:
pip install treegp
which will install the latest released version.
If you would instead like to install the development version, you can do so via:
git clone https://github.com/PFLeget/treegp.git cd treegp/ python setup.py install
treegp
has for now the following dependencies (see the quick
installs below):
- libraries listed in the requirements file
treegp
is regularly tested on Python 3.8, 3.9, 3.10, and 3.11. It may work in other
versions of Python (e.g. pypy), but these are not currently supported.