Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2026
|
| Online Access: | https://doi.org/10.5281/zenodo.20173237 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902077000646656 |
|---|---|
| author | Mohsen Ghanbari |
| author_facet | Mohsen Ghanbari |
| contents | <p>What's New:</p> <p>-Modular codebase refactor: The entire library has been restructured from a single monolithic lib.py into a clean, maintainable package layout: skema.preprocessing: band extraction, spectral indices, normalization, slope calculation, and static layer warping/merging skema.model: model architecture (SegModel) and HuggingFace weight loader skema.inference: tile-based inference engine with halo-weighted stitching (DatasetInference) skema.masking: exclusion zones, depth mask, and eelgrass mask skema.postprocessing: BC Albers mosaic creation</p> <p>-All existing CLI behavior and public API (segment(), create_mosaic()) are fully backward-compatible.</p> <p>-Test suite: 63 unit tests added across all submodules, covering spectral index correctness, normalization, Horn slope calculation, NoData filling, tile generation, weight map symmetry, and model instantiation. Tests require no GPU, no network access, and no real Sentinel-2 data.</p> <p>-CI/CD via GitHub Actions: Automated test runs on every push and pull request across Ubuntu, macOS, and Windows on Python 3.10, 3.11, and 3.12.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_20173237 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | m5ghanba/skema: SKeMa v0.3.5 Mohsen Ghanbari <p>What's New:</p> <p>-Modular codebase refactor: The entire library has been restructured from a single monolithic lib.py into a clean, maintainable package layout: skema.preprocessing: band extraction, spectral indices, normalization, slope calculation, and static layer warping/merging skema.model: model architecture (SegModel) and HuggingFace weight loader skema.inference: tile-based inference engine with halo-weighted stitching (DatasetInference) skema.masking: exclusion zones, depth mask, and eelgrass mask skema.postprocessing: BC Albers mosaic creation</p> <p>-All existing CLI behavior and public API (segment(), create_mosaic()) are fully backward-compatible.</p> <p>-Test suite: 63 unit tests added across all submodules, covering spectral index correctness, normalization, Horn slope calculation, NoData filling, tile generation, weight map symmetry, and model instantiation. Tests require no GPU, no network access, and no real Sentinel-2 data.</p> <p>-CI/CD via GitHub Actions: Automated test runs on every push and pull request across Ubuntu, macOS, and Windows on Python 3.10, 3.11, and 3.12.</p> |
| title | m5ghanba/skema: SKeMa v0.3.5 |
| url | https://doi.org/10.5281/zenodo.20173237 |