MaxQSARing

An autoML toolkit for QSAR Modelling for molecular properties.

Installation

conda env create -n maxqsaring python=3.8.8
conda activate maxqsaring
pip install -r requirements.txt

# Set local data dir:
export DATA_ROOT_DIR={your path of data}

Usage

• Training:

python main-v1.py train -tn {task_name} -s {scaffold, random-cv} -d tempdata

• --task_name, -tn: set task name, eg. herg, bbb_wang

• --split, -s: set the split method, eg. random-cv, scaffold, default is scaffold

• --tmp_dir, -d: set temp directory where the cache data is saved at.

• Evaluating:

python main-v1.py eval -tn {task_name} -d tempdata

• --task_name, -tn: set task name, eg. herg, bbb_wang

• --tmp_dir, -d: set temp directory where the cache data is saved at.

• Predicting:

python main-v1.py predict -tn {task_name} -tf {test_path}

• --task_name, -tn: set task name, eg. herg, bbb_wang
• --test_file, -tf: the test