cedarkit is the meta package for the CEMC meteorological data-processing and
visualization toolkit. Installing it provides the core data-access,
meteorological-computation, and plotting packages as a compatible set.
cedarkit itself intentionally ships no Python modules. It exists to provide a
single installation target and a tested dependency set, following the meta-
package model used by earthkit. The
cedarkit namespace is supplied by component packages as a PEP 420 namespace
package.
This project is at the Sandbox maturity level. Component APIs and the dependency set may still evolve.
| Package | Namespace | Responsibility |
|---|---|---|
| reki | reki |
Locate, read, query, and process meteorological data, including GRIB2 data in CMADaaS-mounted directories. |
| cedarkit-comp | cedarkit.comp |
Meteorological computation utilities, such as smoothing and grid calculations. |
| cedarkit-plots | cedarkit.plots |
Plotting primitives, map domains, styles, colormaps, and declarative plot recipes. |
Together, the components form a simple workflow:
CMADaaS-mounted data → reki → cedarkit.comp → cedarkit.plots → figures
Use the components independently when only one layer is needed, or install this meta package when an application needs the complete toolkit.
Python 3.11 or later is required.
pip install cedarkitThe runtime dependencies install reki, cedarkit-comp, and
cedarkit-plots. Reading GRIB2 data requires an ecCodes installation provided
by the target environment.
After installation, the component namespaces should import directly:
import reki
from cedarkit.comp.smooth import smth9
from cedarkit.plots.chart import Panel
print("cedarkit components are available")For CMADaaS-mounted GRIB data, use reki with the local cmadaas data class
and the mount root:
import reki
reader = reki.from_source(
"local",
"cma_gfs_gmf/grib2/orig",
start_time="2025081900",
forecast_time="24h",
data_class="cmadaas",
storage_base="/CMADAAS",
)
t2m = reader.sel(
parameter="2t",
level_type="heightAboveGround",
level=2,
).first().to_xarray()The mounted-directory workflow reads local files and does not require CMADaaS remote-service credentials.
In this workspace, the three components are configured as local editable dependencies. Install the meta package and its test extra from this directory:
cd repo/cedarkit
uv pip install -e ".[test]"
pytestThe version is derived from Git tags by setuptools_scm. Since this is a meta
package, it has no module-level version file and ships no code of its own.
Copyright © 2021-2026, developers at cemc-oper.
cedarkit is licensed under the Apache License 2.0.