_fields#

Input fields for benchmarks and testing.

There are some idealized fields, in numpy format. Some functions are provided to add noise. “Real-life” data samples are stored on Zenodo (doi:10.5281/zenodo.15769617) and can be downloaded with pooch.

REGISTRY = <pooch.core.Pooch object>#: File registry of data samples.

add_spikes(field, n_spikes=None)#

Add single pixels spikes to field.

Parameters:

field (ndarray[tuple[Any, ...], dtype[_ScalarT]])
n_spikes (int | None)

Return type:

ndarray[tuple[Any, …], dtype[_ScalarT]]

blobby_gradient(n_grid=512, n_blobs=800, max_blob_size=7)#

Meridional gradient with blobs.

Parameters:

n_grid (int)
n_blobs (int)
max_blob_size (int)

Return type:

ndarray[tuple[Any, …], dtype[_ScalarT]]

ideal_jet(n_grid=512)#

Idealized jet with meanders and eddies.

Parameters:: n_grid (int)
Return type:: ndarray[tuple[Any, …], dtype[_ScalarT]]

sample(name)#

Return sample dataset.

Parameters:: name (str) – Name of the dataset to retrieve. Can be ESA-CCI-C3S or MODIS.
Return type:: Dataset

swap_noise(field, n_swap=None, len_swap=3)#

Add noise by swapping pixels.

Parameters:

field (ndarray[tuple[Any, ...], dtype[_ScalarT]])
n_swap (int | None)
len_swap (int)

Return type:

ndarray[tuple[Any, …], dtype[_ScalarT]]

swap_noise_higher(field, n_swap=None, len_swap=3)#

Add noise by swapping pixel (only towards higher values).

Parameters:

field (ndarray[tuple[Any, ...], dtype[_ScalarT]])
n_swap (int | None)
len_swap (int)

Return type:

ndarray[tuple[Any, …], dtype[_ScalarT]]