Skip to content

Releases

v0.1.0

Choose a tag to compare

@github-actions github-actions released this 11 Nov 21:10

The first release of pcsf, the reference implementation of the
Portable Common Subsurface Format (format version 0.1.0): one
exchange format for electromagnetic and magnetotelluric inversion
models, whatever code produced them.

Every inversion code writes its own format, with its own logarithm
base, axis order, coordinate frame, depth datum and discretization.
PCSF states these conventions in the file, so a model can be read,
checked and compared without knowing the code that made it.

Install

pip install pcsf              # core: NumPy + h5py
pip install "pcsf[all]"       # + algorithms, EMC, MARE2DEM, CRS, schemas, Studio

PCSF Studio — download

The desktop application, with installers that need no Python:

Platform File
Windows 10/11 (64-bit) pcsf-studio-0.1.0-setup.exe (also pcsf-studio-setup.exe)
Linux (x86-64), portable pcsf-studio-0.1.0-x86_64.AppImage (also pcsf-studio-x86_64.AppImage)
Linux (x86-64), installs in ~/.local pcsf-studio-0.1.0-linux.sh (also pcsf-studio-linux.sh)

The files without a version number always point to the latest release.
Windows SmartScreen warns about unsigned installers: choose More
info → Run anyway
. Check a download against SHA256SUMS.

Highlights

  • Two encodings, one model: PCSF (HDF5) and PCSM (plain text,
    optionally .pcsm.gz) convert into each other bit for bit.
  • What the file holds: canonical resistivity in linear Ω·m with
    the solver's own values and encoding kept; the native geometry
    (2D grid, 3D grid, triangular mesh or set of profiles), never
    resampled; stations, topography and coordinate reference system;
    and an inversion record (data, errors, regularization, misfit,
    convergence, depth of investigation) with a completeness check.
  • Solver output without the solver: adapters for Occam2D, ModEM,
    MARE2DEM and the EarthScope Earth Model Collaboration (EMC) netCDF
    archives, on NumPy-only parsers.
  • Checks: conformance rules C1–C10, JSON Schemas of the metadata,
    and a conformance suite of 27 reference files for independent
    implementations.
  • Algorithms: sampling, point location on meshes, comparison of
    models, uncertainty propagation and frame registration, NumPy only.
  • Tools: the pcsf command line (info, validate, convert,
    compare, studio) and PCSF Studio, which views and inspects
    models and converts many solver runs to .pcsf, .pcsm or
    .pcsm.gz in one go.
  • Reproducibility: paper/ reproduces every number, table and
    figure of the paper that describes the format.

Documentation

pcsf is released under the BSD 3-Clause licence. For users of
pycsamt.format: pyCSAMT now uses pcsf as its format core; see
Changed relative to pycsamt.format in the changelog.