Installing

There are different ways to install Harmonica:

conda/mamba

Using the conda package manager (or mamba) that comes with the Anaconda, Miniconda, or Miniforge distributions:

conda install harmonica --channel conda-forge

pip

Using the pip package manager:

python -m pip install harmonica

Development version

You can use pip to install the latest unreleased version from GitHub (not recommended in most situations):

python -m pip install --upgrade git+https://github.com/fatiando/harmonica

Tip

The commands above should be executed in a terminal. On Windows, use the cmd.exe or the “Anaconda Prompt” / “Miniforge Prompt” app if you’re using Anaconda / Miniforge.

Which Python?

See Supported Python versions for a list of supported Python versions.

Note

We recommend using the Miniforge distribution to ensure that you have the conda package manager available. Installing Miniforge does not require administrative rights to your computer and doesn’t interfere with any other Python installations in your system. It’s also much smaller than the Anaconda distribution and is less likely to break when installing new software.

Dependencies

The required dependencies should be installed automatically when you install Harmonica using conda or pip. Optional dependencies have to be installed manually.

Required:

• numpy

• pandas

• numba

• scipy

• xarray

• scikit-learn

• pooch

• verde

• bordado

• boule

• xrft

Optional:

• pyvista and vtk (>= 9): for 3D visualizations. See harmonica.prism_to_pyvista.

• numba_progress for printing a progress bar on some forward modelling computations. See harmonica.prism_gravity.

Note

See Supported dependency versions for our policy of oldest supported versions of each dependency.

The examples in the Gallery also use:

• ensaio for downloading sample datasets

• pygmt for plotting maps

• pyproj for cartographic projections