Skip to content

Imports in this page are from the original geo_toolz layout

These design docs were adapted from the geo_toolz design study. Code snippets use the original feature-based import paths (geo_toolz.<module>). In xrtoolz, the domain-agnostic operations live under xrtoolz.geo.<module>; physics-specific operations live under xrtoolz.ocn / xrtoolz.atm / xrtoolz.rs. See xrtoolz/__init__.py and xrtoolz/geo/__init__.py for the current export surface.

geo_toolz — Examples

Usage patterns organized by API layer. Demonstrates the design principles: - P1: Everything is an operator — uniform __call__ interface - P2: Progressive disclosure — L0 functions → L1 operators → L2 graphs - P3: xarray in, xarray out — coordinates and metadata preserved - P4: Bring your own model — inference as a pipeline step

Structure

examples/
├── README.md              # This file
├── primitives.md          # Layer 0 — pure functions, pipe syntax, toolz composition
├── components.md          # Layer 1 — Sequential, operator composition, Hydra, stateful ops
├── models.md              # Layer 2 — Graph API, inference, model comparison
├── integration.md         # Layer 3 — sklearn, xrpatcher, xarray_sklearn, ekalmX
├── validation.md          # Expanded validation graphs and skill diagnostics
├── lagrangian.md          # Particle, trajectory, dispersion, and connectivity examples
├── budgets.md             # Conservation and control-volume budget examples
└── phenomena.md           # Event/object detection and verification examples

Reading Order

  1. primitives.md — L0: pure functions and pipe syntax
  2. components.md — L1: Sequential pipelines and operator patterns
  3. models.md — L2: Graph API, ModelOp, model comparison
  4. integration.md — L3: sklearn, xrpatcher, xarray_sklearn, ecosystem
  5. validation.md — validation graphs and field/spectral/structural/process diagnostics
  6. lagrangian.md — material transport and trajectory diagnostics
  7. budgets.md — heat, salt, volume, and kinetic-energy budgets
  8. phenomena.md — marine heatwave, eddy, and generic event verification

Validation Examples

  • validation.md — end-to-end validation examples with pixel, spectral, lead-time, structural, probabilistic, regional, and process scores.
  • lagrangian.md — particle advection, endpoint error, pair dispersion, residence time, connectivity, and FTLE-style diagnostics.
  • budgets.md — control-volume heat, salt, volume, and kinetic-energy budget residual examples.
  • phenomena.md — marine heatwave, eddy, generic event, and event-graph verification examples.