Guassian Process Classifiction
Python environment is being managed by conda, if needed, can install via (https://docs.conda.io/en/latest/miniconda.html). Or just execute
cd ~
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh
Configuration is stored in the ENV.yml file.
To create your environment just use conda env create --file ENV.yml from the base of the repo.
Then execute conda activate gproc to activate the environment.
To update using a new config use conda env update --file ENV.yml --prune.
For other conda tasks check out the cheat sheet @ https://docs.conda.io/projects/conda/en/latest/_downloads/843d9e0198f2a193a3484886fa28163c/conda-cheatsheet.pdf .
To install in dev mode run pip install -e . from the root of the project repo.
C++ dependencies are built by pybind11.
This package is installed by pip in the ENV.yml.
To use this we add "pybind11~=2.6.1" to our required build systems in pyproject.toml.
Additionally, we need to add a build step to our setup.py file, which will build .cpp files in the cpp/ directory and map to conigured python module structure.
For more details look at https://pybind11.readthedocs.io/en/latest/basics.html .
Static configuration is managed in setup.cfg.
The data folder for the reported experiments, if required, can be downloaded from https://drive.google.com/drive/folders/17rLJQekKVYesQjRptC9g2GDKa5yv7qV2?usp=sharing . The experiment script may need to be edited to point at the data folder.
Documented in NumPy style.
Documentation is built automatically by sphinx.
To build run make html from the docs directory.
The corresponding docs are then found in docs/_build/index.html.
On merge to main, the docs are rebuilt and deployed automatically to https://tennessee-wallaceh.github.io/gproc/.
For the automatic documentation to be generated the configuration needs to be added to docs/index.rst.