midas-grain-odf

Differentiable per-grain orientation distribution function (ODF) inversion from far-field HEDM data. Closes the gap left open by Nygren et al. 2019 (IOP Conf. Ser. MSE 580 012018) — recovers the intensity-weighted ODF, not just the GOE envelope.

See dev/implementation_plan.md for the design rationale and roadmap.

Status

Alpha. Synthetic plant-and-recover validation only. Real-data path under construction.

Install

cd packages/midas_grain_odf
pip install -e .[dev]

The package depends on the differentiable HEDM forward model in fwd_sim/hedm_forward.py. That module is loaded by relative import; ensure MIDAS/fwd_sim is on PYTHONPATH when using the package outside the MIDAS repo.

Notebooks

Worked-example Jupyter notebooks live in notebooks/. They are not shipped with pip install — get them by cloning the MIDAS repository.

Quickstart (synthetic)

cd packages/midas_grain_odf
python -m pytest tests/test_synth_particle.py -xvs

Public API

from midas_grain_odf import (
    ParticleODF, BinghamMixtureODF, VoxelGridODF,
    fit_grain_odf,
)