EVMS v0.1.1

EVMS v0.1.1 — Volumetric Inversion of Geological Radioactivity from Surface Gamma Measurements

EVMS is a scientific Python framework for volumetric inversion of geological radioactivity from surface gamma measurements, with sparse forward modeling, regularized inversion, and a multipage Streamlit interface.

Highlights

• Sparse voxel-based forward operator with attenuation and distance kernel.
• Tikhonov inversion with graph regularization.
• Layer-aware smoothing and finite fracture barrier support.
• Optional calibration workflow (relative units -> physical units).
• Inversion diagnostics: residual map, ||AS - M||, ||LS||, optional holdout error, trust report.
• Automatic parameter search for mu and R_max.
• Publication-ready Streamlit UI with:
  • Inversion Workspace
  • How EVMS Works
  • Credits
• Mesh export with baked texture (OBJ + MTL + PNG) and volumetric export (.npy).

Validation

• Test suite included (pytest) covering geometry, forward model, inversion, calibration, and diagnostics.

Installation

conda env create -f environment.yml
conda activate evms-env
pip install -e .

Run

streamlit run streamlit_app.py

Project links

Website maxsc4.github.io
Contact : soarescorreia@ipgp.fr

Release notes

v0.1.1 :

• Release-ready README with cleaner scientific summary.
• Added CITATION.cff.
• DOI integrated in README and Credits page.
• Package and citation metadata updated to v0.1.1.

EVMS v0.1.0 — First Public Release

EVMS v0.1.0 — First Public Release

EVMS is a scientific Python framework for volumetric inversion of geological radioactivity from surface gamma measurements, with sparse forward modeling, regularized inversion, and a multipage Streamlit interface.

Highlights

• Sparse voxel-based forward operator with attenuation and distance kernel.
• Tikhonov inversion with graph regularization.
• Layer-aware smoothing and finite fracture barrier support.
• Optional calibration workflow (relative units -> physical units).
• Inversion diagnostics: residual map, ||AS - M||, ||LS||, optional holdout error, trust report.
• Automatic parameter search for mu and R_max.
• Publication-ready Streamlit UI with:
  • Inversion Workspace
  • How EVMS Works
  • Credits
• Mesh export with baked texture (OBJ + MTL + PNG) and volumetric export (.npy).

Validation

• Test suite included (pytest) covering geometry, forward model, inversion, calibration, and diagnostics.

Installation

conda env create -f environment.yml
conda activate evms-env
pip install -e .

Run

streamlit run streamlit_app.py

Project links

• Website maxsc4.github.io
• Contact : soarescorreia@ipgp.fr