Skip to content

API Reference

The public API is organised by subject, mirroring the three-layer architecture (primitives → operators → graph). Every page pairs Layer-0 pure functions — (xr.Dataset, …) → xr.Dataset — with the Layer-1 Operator wrappers that compose them into Sequential chains and Graph networks.

How to read these pages

Each entry is generated from the source docstring, so the signature, argument types, and shapes you see here are exactly what ships in the package. Layer-0 functions take a dim: (a coordinate name); the private numpy kernels behind them take an axis: (an integer) and are jaxtyped — see the array-typing convention.

Composition core

Page Scope
Composition Operator, Sequential, Graph, Input, Node, Tap, Augment, ApplyToEach, Signature

Geoprocessing

Domain-agnostic xarray geoprocessing. Anything that is not domain physics lives here.

Page Scope
Validation & Coordinates Coordinate validation, CF-time decoding, CF standard-name renaming, grid resolution
Subsetting, Regions & Masks Bounding-box / time / region subsetting, the region registry, land/ocean/country masks
CRS & Reprojection CRS assignment, reprojection, lon/lat ↔ projected coordinates
Climatology & Anomalies Climatologies, seasonal cycles, anomalies, detrending
Wavelet Analysis 1-D / 2-D continuous wavelet transforms, scalograms, significance
Extremes Block extrema and peaks-over-threshold diagnostics

Numerics

Page Scope
Calculus Finite-difference gradient, divergence, curl, laplacian on cartesian / spherical grids
Transforms Fourier / wavelet / DCT spectra, spectral fluxes, matrix decompositions, morphology
Encoders Coordinate-space, cyclical-time, and Fourier-feature encodings
Interpolation Regridding, coarsening, gap-filling, binning, smoothing, point sampling

Domain physics

Page Scope
Oceanography Geostrophy, vorticity, kinetic energy, stratification, SSH diagnostics
Atmosphere Atmospheric and trace-gas physics (planned)

Diagnostics & evaluation

Page Scope
Metrics Pixel, spectral, multiscale, physical, masked, distributional metrics
Budgets Conservation-budget residuals (heat, salt, volume, KE)

Bridges & integration

Page Scope
scikit-learn on xarray XarrayEstimator on DataArrays, Datasets and DataTrees, and the .sklearn accessor
Inference ModelOp wrappers for sklearn / JAX / framework-agnostic models
Visualization Cartopy axes, colormap registry, and the V1–V5 validation panels
Utilities pack_dataset / unpack_dataset

Whole-package dump

For a flat, searchable dump of every public symbol:

xrops — xarray operators for Earth System Data Cubes.

The operator package of the xrtoolz family (pip install xrtoolz-ops). Each family is a subpackage: :mod:xrops.anomaly, :mod:xrops.atm, :mod:xrops.budgets, :mod:xrops.cf, :mod:xrops.extremes, :mod:xrops.geom, :mod:xrops.inference, :mod:xrops.interpolate, :mod:xrops.mask, :mod:xrops.metrics, :mod:xrops.ocn, :mod:xrops.transforms, :mod:xrops.utils, :mod:xrops.viz and :mod:xrops.wavelets. The composition layer (Sequential, Graph, Input, Tap) and Signature are :mod:pipekit; :class:Operator is the xarray-aware base every xrops operator subclasses (DataTree dispatch on top of pipekit.Operator).