Hypershape recognition (HSR): a general framework for moment-based similarity measures

HSR is a versatile, moment-based similarity measure tailored for three-dimensional (3D) chemical representations annotated with atomic features. It enhances the robustness and versatility of the Ultrafast Shape Recognition (USR) method by incorporating multidimensional features for each atom, such as protons, neutrons, and formal charges.

Getting Started

Installing HSR

You can install HSR using either pip or conda:

pip install hsr

or

conda install hsr -c conda-forge

Build from source

Clone this repository on your machine. Move inside it and create the conda environment:

conda env create -f environment.yml
conda activate HSR_devel

Verify the correct creation of the environment by running:

pytest

To use HSR from CLI you can run:

python -m hsr.hsr_cli 

If HSR is installed with pip or conda, the above command is replace by the simple use of hsr

Basic Usage

Run the folowing command to get help in using HSR from CLI:

hsr -h

For a detailed overview of HSR's methodology check our documentation.

Licensing

HSR is licensed under the GNU Affero General Public License Version 3. For more details, see the LICENSE file.

Citing HSR

If you use HSR in your research, please cite it as follows:

[TODO: Add citation]

Contributing

Contributions to HSR are welcome! Please read our Contributing Guidelines for information on how to get started.