emsarray

The emsarray package provides a common interface for working with the many model geometry conventions used at CSIRO. It enhances xarray Datasets and provides a set of common operations for manipulating datasets.

To use, open the dataset using the emsarray.open_dataset() function and use the dataset.ems attribute:

import emsarray
from shapely.geometry import Point

dataset = emsarray.tutorial.open_dataset('gbr4')
capricorn_group = Point(151.869, -23.386)
point_data = dataset.ems.select_point(capricorn_group)

Some methods take a DataArray as a parameter:

# Plot the sea surface temperature for time = 0
temp = dataset['temp'].isel(time=0, k=-1)
dataset.ems.plot(temp)

A number of operations provide further functionality to manipulate datasets, export geometry, and select subsets of data:

from emsarray.operations import geometry
geometry.write_geojson(dataset, './gbr4.geojson')
geometry.write_shapefile(dataset, './gbr4.shp')

Links

• Source code on Github
• Documentation on Read The Docs
• emsarray on PyPI
• emsarray on conda-forge

Examples

Examples of using emsarray are available in the emsarray-notebooks repository. You can explore these notebooks online with Binder.

Developing

To get set up for development, make a virtual environment and install the dependencies:

$ python3 -m venv
$ source venv/bin/activate
$ pip install --upgrade pip>=21.3
$ pip install -e . -r continuous-integration/requirements.txt

Tests

To run the tests, install and run tox:

$ python3 -m venv
$ source venv/bin/activate
$ pip install --upgrade pip>=21.3 tox
$ tox

Documentation

The documentation for the current stable version of emsarray is available on Read The Docs.

To build the documentation, install the development requirements as above and invoke Sphinx:

$ make -C docs/ html

While updating or adding to the documentation, run the live target to automatically rebuild the docs whenever anything changes. This will serve the documentation via a livereload server.

$ make -C docs/ live

You can the view the docs at http://localhost:5500