Repository navigation
Modernize OpenLSR-X for publication: 3D orthogonal fusion, petrophysical metrics, benchmarks, CI/CD & JOSS docs - #1
Merged
Conversation
…cal metrics, benchmarks, CI, and JOSS docs
…abot, issue/PR templates, Mypy checks, and 89% test coverage
… I13-2 Zn-13X citations
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Brings
BASE-Laboratory/OpenLSR-Xfrom a partial PhD handoff fork to a reproducible, publication-ready Python package and porous-media XCT super-resolution benchmark suite.Key Fixes & Architectural Upgrades
src/openlsrx/data/: Fixed the unanchoreddata/rule in.gitignorethat had dropped the entire data package during thesrc/layout refactor; implemented vectorized, leakage-free 3D spatial slab splitting (split_volume_spatiallywith configurable guard margins), PoreSpy 2.x/3.x synthetic porous-media generation, unified TIFF/slice-folder/NeXus-HDF5 (PairedVolumeDataModule) loading, and an open micro-CT benchmark fetcher (openlsrx download-benchmarks).openlsrx.inference): Reconstructs isotropic 3D super-resolved XCT volumes from 2D-trained generators acrossyz,xz, andxyplanes using mean, median, or directional 3D Fourier weighting—suppressing single-axis slice-stacking ("venetian blind") discontinuities.openlsrx.models,openlsrx.losses): Converted building blocks andVGGLossto puretorch.nn.Module, isolated frozen VGG-19 weights (~550 MB) from checkpoints while adding ImageNet normalization, addedSRResNetablation + generator warmup support, eliminated allos.chdir/os.systemside effects, and fixed contiguous 3D multi-page BigTIFF export (contiguous=True).openlsrx.metrics,openlsrx.baselines): Added 2D/3D classical baselines (Nearest,Trilinear,Tricubic,Lanczos4), volumetric PSNR & isotropic tri-planar 3D SSIM, plus porous-media petrophysical metrics: segmented porosity error (scripts/benchmarks/porespy_study.py), Open Rock Micro-CT (scripts/benchmarks/public_rock_xct.py), and automatedbooktabsLaTeX table + 300-DPI figure builder (scripts/benchmarks/make_paper_outputs.py).openlsrxCLI (fit,validate,test,predict,evaluate,download-benchmarks) with schema-validated YAML configs (configs/).ci.yml(Ruff, Mypy, MkDocs--strict, wheel build, Python 3.10–3.12 test matrix withdocs.yml(GitHub Pages CD),release.yml(wheel/sdist + JOSSopenjournals/inaraPDF compilation + GitHub Releases onv*tags), Dependabot, issue/PR templates,CITATION.cff,CONTRIBUTING.md,CODE_OF_CONDUCT.md, and JOSS manuscript draft (paper/paper.md+paper/paper.bib).Verification
ruff check src tests scripts&ruff format --check src tests scriptspass (0 errors)mypy src tests scriptspasses (0 errors across 30 files)pytest --cov=openlsrx --cov-fail-under=85passes (52/52 passed, 89.03% coverage)mkdocs build --strictbuilds cleanly--fast) +make_paper_outputs.pyverified end-to-end