repositoryrepositoryThe SNV/SG/PLS baseline expressed as a dag-ml pipeline DSL (nirs4all-compatible).
unvalidatedThe robust SNV/SG/PLS regression baseline, expressed as a dag-ml pipeline DSL.
The repository serves both the nirs4all and dag-ml ecosystems through one storage envelope. This recipe is the dag-ml counterpart of snv_savgol_pls: the same preprocessing and PLS model, but authored as a dag-ml DSL that a dag-ml-aware runtime compiles into a graph + campaign template.
window_length=11, polyorder=2).Provenance, content-addressing, and the cross-language contract are identical to the nirs4all recipes — only the recipe grammar differs.
{
"name": "dag-ml — SNV + Savitzky-Golay + PLS",
"pipeline": [
{"class": "nirs4all.operators.transforms.StandardNormalVariate"},
{"class": "nirs4all.operators.transforms.SavitzkyGolay", "params": {"window_length": 11, "polyorder": 2, "deriv": 1}},
{"class": "sklearn.preprocessing.MinMaxScaler"},
{"split": {"class": "sklearn.model_selection.ShuffleSplit", "params": {"n_splits": 5, "test_size": 0.25, "random_state": 42}}},
{"model": {"class": "sklearn.cross_decomposition.PLSRegression", "params": {"n_components": 12}}}
]
}
# Python
import nirs4all_repository as n4r
pipe = n4r.get("dagml_snv_pls")
config = pipe.to_nirs4all() # ready for nirs4all.run() / predict()# any language: read the index, fetch + verify
curl https://repository.nirs4all.org/data/index.json
curl https://repository.nirs4all.org/data/pipelines/dagml_snv_pls/pipeline.json