mdtanker/invert4geom

3D geometric gravity inversions

Invert4geom

3D geometric gravity inversions

Documentation Link

Invert4geom is a Python library for performing 3D geometric gravity inversions, where the aim is to recover the geometry of a density contrast.

Typical use cases include modeling the topography of the Moho, the sediment-basement contact, or bathymetry. These density contrasts are represented by a layer of vertical right-rectangular prisms. Since we use vertical prisms, they don't take the curvature of the Earth into account. For large-scale applications, such as continental studies, it would be better to use tesseroids instead of prisms.

See the overview for further description of this package and what it can be used for.

What Invert4Geom is for:

• invert gravity data to recover the topography of a density contrast
• use with gridded gravity data
• discretization using vertical right-rectangular prisms
• performing regional-residual separation
• stochastic uncertainty analysis of inversion results
• basic visualization tools

What Invert4Geom is NOT for:

• performing parameter-estimation inversions (i.e. recovering density values) -> see SimPEG or pyGIMLi
• a point-and-click GUI -> see Geosoft/Seequent Oasis Montaj or MiraGeoscience VPmg
• a ready-to-use inversion with minimal user input
• use with discrete (un-gridded) gravity data -> see Harmonica for gridding your data
• for processing gravity data -> see Harmonica for gravity processing tools
• for use with tesseroids, or non-regular grids (raise an issue request if you want this implemented!)
• publication-quality visualization -> see PyGMT for plotting tools

invert4geom's Issues

Add support for Python 3.12

Once numba supports 3.12, tracked here, we should extend support to 3.12.

• add 3.12 to classifiers in pyproject.toml
• replace 3.11 with 3.12 in ci.yml tests

Show links to code in documentation

Since switching from autodoc to autoapi, on RTD there are no links to show the code for each function of the API. It works locally. I've tried switching from sphinx.ext.viewcod to sphinx.ext.linkcode with no success.

add Zenodo DOI

Interactive widgets in documentation

It would be great to implement interactive figures with pyvista in the documentation. The new release of PyGimli has examples of this.

Switch from `sphinx.autodoc` to `sphinx.autoapi`

Hopefuly this switch would remove the need to install runtime dependencies with the Readthedocs environement, as defined in the file RTD_env.yml.

See autoapi instrustions.

Changelog template issues

Same as mdtanker/polartoolkit#177, the changelog template needs to be update to remove commit categories for excluded commits.

`zref` and `density_contrast` cross validation issue when using constraint point minimization for regional estimation.

The reference level and density contrast cross validation currently uses constraint points to calculate a score. This works well, except for when constraint points are also used in the regional field calculation. We should both raise a warning to users for this situations that the cross validation will not work properly, and also create a separate cross-validation function for this specific scenario. It would split the constraints into testing and training sets, so that the cross validation should work properly.