Installation

Requirements

pyshqg provides two implementations of the QG model: the first based on NumPy and the second based on TensorFlow.

The NumPy-based variant depends on:

• Python;

• NumPy: for array support;

• pyshtools: for the spherical harmonics coefficients;

• Xarray: for dataset support;

• Zarr: for dataset storage;

• tqdm: for the progress bar.

The TensorFlow-based variant additionally depends on

• TensorFlow: for tensor support.

Instructions

For the NumPy-based variant

The easiest way to install pyshqg is to use pip. This will also install all the dependencies for the NumPy-based variant.

$ pip install pyshqg

For the TensorFlow-based variant

Iy you need support for the TensorFlow-based variant, you just need to have TensorFlow available on top of what pip has installed. For example, you can follow the instructions here.