Installation

1. git clone https://github.com/stair-lab/bo4ag
2. cd your_repository
3. Create a python 3.10 conda environment
4. pip install -r requirements.txt

Downloading the data:

1. The coheritability data can be found on [huggingface](https://huggingface.co/datasets/stair-lab/coh2 or google drive.
2. Download the files to ./Benchmark/data.

Running Bayesian Optimization (BO) Benchmarks

1. cd Benchmarks
2. Run python run_BO.py --env <environment>

Options:

Argument	Description	Default Value	Example
--env	Environment to run the search (e.g. narea, sla, pn, ps)	None	--env narea
--kernel	Kernel function for the Gaussian process	rbf	--kernel matern52
--acq	Acquisition function	EI	--acq EI
--transform	Transforming on the search space	None	-- transform log
--n	Number of iterations	300	--n 300
--gpu	GPU id to run the job	0	--gpu 0
--run_name	Name of the folder to move outputs	None	--run_name test

You can copy and paste this markdown table into your README.md file.

Example: python main.py --env narea --n 30 --kernel matern12 --acq EI --run_name test

Fitting A Surrogate Model

1. cd Benchmarks
2. Run python gp_fit.py --env <trait> --n <number of iterations> --kernel <kernel> --acq <acquisition function>