Setup

Installing Python

The iprPy framework is based on Python and therefore requires that Python be installed. Luckily, Python is free and compatible with Linux, MacOS and Windows. You can download Python from python.org, although it is highly recommended to obtain a Python distribution from Anaconda. The Anaconda distributions support multiple Python environments, which can greatly help avoid possible versioning conflicts that may arise between different Python-based programs. Additionally, the Anaconda distributions come with the scipy family of packages already installed, which iprPy heavily relies on.

Installing iprPy

If you wish to use iprPy simply to access and run the existing calculations, then you can easily install it with pip or with conda-forge if you have a conda distribution of Python

pip install iprPy

or

conda install iprPy -c conda-forge

Alternatively, if you have an interest in developing new calculations or modifying the existing ones, you can download the code from github. The github repositories also contain the Jupyter Notebook versions of the calculations for anyone to download and use.

• https://github.com/usnistgov/iprPy always coincides with the latest release of iprPy, i.e. it will be consistent with the pip and conda-forge packaged downloads. This allows users to download the most recent stable version of iprPy.

• https://github.com/lmhale99/iprPy is the development repo of iprPy. It may contain new calculation methods or other updates that are in the works but not yet included in the calculation release. Ideally, any user pull requests should be submitted here.

Once you download a github repository, you can install it to your Python environment in editable/development mode by changing the directory to the repository’s root directory and

pip install -e

Testing the install

Installing iprPy should install an iprPy command line option to the Python environment. To test if iprPy is working, enter the following command in a terminal

python iprPy check_modules

If everything installed properly, this should print informative statements about the modules that make up iprPy. In particular, there should be a list of modules, including calculation methods, that properly loaded or failed to load. Those that failed to load should display messages that either state that the modules are not supported by the current version of iprPy or that those modules require additional Python packages to be installed in order to work. Should you wish to use those modules, be sure to add the necessary Python packages.

If it displays an error that the iprPy command cannot be found, then the installer failed to add the executable script to the proper location. The executable script can be found in the iprPy repository at iprPy/bin/iprPy. Ideally, you should then copy/download this file to the associated Python environment’s Scripts directory. Alternatively, you can place the script file in any other directory and call it directly or “install” it by adding it to any directory listed in your system’s PATH environment variable.

For any other issues, feel free to email potentials@nist.gov for support.

Updating iprPy

For packaged installs, ipyPy can be updated using pip or conda update options. This should update not only iprPy’s version but also the versions of any required packages.

For editable/development installs, the code can be updated by doing a pull request from the github repository. Note, however, that this will not update required packages. Check the requirements.txt file for any new version updates, most notably for the related packages “potentials” and “atomman”. If needed, updating atomman should fix any requirement issues.