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.

Layer 0 — Primitive Examples¶

Pure function usage patterns. (P2: progressive disclosure — simplest layer first)

Pipe Syntax with toolz¶

Compose L0 functions into functional pipelines¶

from toolz import compose_left
from functools import partial
from geo_toolz._src.validation.coords import validate_longitude, validate_latitude
from geo_toolz._src.subset.where import subset_bbox
from geo_toolz._src.detrend.climatology import remove_climatology

preprocess = compose_left(
    validate_longitude,
    validate_latitude,
    partial(subset_bbox, lon_bnds=(-30, 45), lat_bnds=(25, 65)),
    partial(remove_climatology, climatology=clim),
)

ds_clean = preprocess(ds_raw)

Direct Function Calls¶

xr.Dataset.pipe for one-off transforms¶

from geo_toolz._src.detrend.climatology import calculate_climatology, remove_climatology

# Learn state from training data
clim = calculate_climatology(ds_train, freq="day", smoothing=60)

# Apply — P3: xarray in, xarray out
ds_anom = remove_climatology(ds_test, clim)