You can install MoleculeKit either using pip or conda. Various methods of MoleculeKit have additional dependencies which are not installed by default to keep the package lighter and keep it installable by pip. If you encounter errors with missing libraries you can install these manually to resolve the issues.
We generally advise to use a conda installation regardless if installing MoleculeKit over pip or conda as well as creating a separate conda environment for the installation to avoid conflicts with other libraries or python version. If you however prefer to install it to your system python you can ignore the following two steps related to conda installation and environments at your own risk.
Miniconda is a lightweight installer of conda containing just a few basic python libraries. Download miniconda from the following URL https://docs.conda.io/en/latest/miniconda.html and install it following the given instructions.
After installing miniconda you can create a conda environment for MoleculeKit with the following command:
conda create -n moleculekit python=3.6
This will create a new conda environment for python version 3.6. This does not install MoleculeKit. It simply creates a clean empty python environment named moleculekit. To now operate within this new environment use the following command to activate it. Anything installed with pip or conda after this command will be installed into this clean python environment.:
conda activate moleculekit
pip install moleculekit
conda install moleculekit -c acellera