Skip to content

Types

ArrayCHW = Float[NDArray[float32], 'channels height width'] module-attribute

ArrayHW = Float[NDArray[float32], 'height width'] module-attribute

ArrayHWV = Float[NDArray[float32], 'height width variable'] module-attribute

ArrayIndices2D = Int[NDArray[int64], 'height width'] module-attribute

ArrayTCHW = Float[NDArray[float32], 'time channels height width'] module-attribute

ArrayTHW = Float[NDArray[float32], 'time height width'] module-attribute

DiffMode = Literal['signed', 'absolute', 'smape'] module-attribute

DiffStrategy = Literal['precompute', 'two-pass', 'per-frame'] module-attribute

Hemisphere = Literal['north', 'south'] module-attribute

TensorNCHW = Float[Tensor, 'batch channels height width'] module-attribute

TensorNTCHW = Float[Tensor, 'batch time channels height width'] module-attribute

DataSpace

Description of a CHW data space.

channels = int(channels) instance-attribute

name = name instance-attribute

shape = (int(shape[0]), int(shape[1])) instance-attribute

area property

Return the area of the data space.

chw property

Return a tuple of [channels, height, width].

from_dict(config) classmethod

to_dict()

Return the DataSpace as a DictConfig.

ModelStepOutput dataclass

Bases: Mapping[str, Tensor]

Output of a model step: prediction, target, and loss.

prediction instance-attribute

target instance-attribute

loss instance-attribute

copy()

Return a plain dict copy, required by Lightning's manual optimization loop.

PlotSpec dataclass

Configure how sea-ice plots are rendered.

Attributes:

Name Type Description
variable str

Variable name shown in plots / used for routing.
title_groundtruth str

Title above the ground-truth panel.
title_prediction str

Title above the prediction panel.
title_difference str

Title above the difference panel.
n_contour_levels int

Number of contour levels per panel.
colourmap str

colourmap used for GT/prediction panels.
dpi int

Dots per inch for figure rendering (default 300).
include_difference bool

Whether to draw a difference panel.
diff_mode DiffMode

Difference definition (e.g. "signed", "absolute", "smape").
diff_strategy DiffStrategy

Strategy for animations (precompute, two-pass, per-frame).
selected_timestep int

Slice index when a single timestep is needed.
vmin float | None

Lower bound for GT/prediction colour scale (None = infer).
vmax float | None

Upper bound for GT/prediction colour scale (None = infer).
colourbar_location Literal['vertical', 'horizontal']

"vertical" or "horizontal".
colourbar_strategy Literal['shared', 'separate']

"shared" or "separate" colourbars.
outside_warn float

Threshold for “values outside display range” warnings.
severe_outside float

Severe threshold for clipping warnings.
include_shared_range_mismatch_check bool

If True, add magnitude mismatch nudges.

variable = 'sea_ice_concentration' class-attribute instance-attribute

title_groundtruth = 'Ground Truth' class-attribute instance-attribute

title_prediction = 'Prediction' class-attribute instance-attribute

title_difference = 'Difference' class-attribute instance-attribute

n_contour_levels = 51 class-attribute instance-attribute

colourmap = 'viridis' class-attribute instance-attribute

dpi = 300 class-attribute instance-attribute

include_difference = True class-attribute instance-attribute

diff_mode = 'signed' class-attribute instance-attribute

diff_strategy = 'precompute' class-attribute instance-attribute

selected_timestep = 0 class-attribute instance-attribute

vmin = 0.0 class-attribute instance-attribute

vmax = 1.0 class-attribute instance-attribute

colourbar_location = 'horizontal' class-attribute instance-attribute

colourbar_strategy = 'shared' class-attribute instance-attribute

outside_warn = 0.05 class-attribute instance-attribute

severe_outside = 0.2 class-attribute instance-attribute

include_shared_range_mismatch_check = True class-attribute instance-attribute

hemisphere = None class-attribute instance-attribute

metadata_subtitle = None class-attribute instance-attribute

video_fps = 2 class-attribute instance-attribute

video_format = 'mp4' class-attribute instance-attribute

per_variable_styles = field(default_factory=lambda: {'sic-osisaf:ice_conc': {'cmap': 'Blues_r'}, 'sic-ssmis:ice_conc': {'cmap': 'Blues_r'}}) class-attribute instance-attribute

BetaSchedule

Bases: StrEnum

Enum for diffusion beta schedule types.

LINEAR = 'linear' class-attribute instance-attribute

COSINE = 'cosine' class-attribute instance-attribute

MaskType

Bases: StrEnum

Enum for types of masking.

ACTIVE = 'active' class-attribute instance-attribute

LAND = 'land' class-attribute instance-attribute

NONE = 'none' class-attribute instance-attribute

RangeRestriction

Bases: StrEnum

Enum for bounded output types.

CLAMP = 'clamp' class-attribute instance-attribute

NONE = 'none' class-attribute instance-attribute

SIGMOID = 'sigmoid' class-attribute instance-attribute

TANH = 'tanh' class-attribute instance-attribute

SkipConnectionType

Bases: StrEnum

Enum for decoder skip connection types.

ADDITIVE = 'additive' class-attribute instance-attribute

CONVOLUTIONAL = 'convolutional' class-attribute instance-attribute

GATED = 'gated' class-attribute instance-attribute

NONE = 'none' class-attribute instance-attribute

SupportsMetadata

Bases: Protocol

set_metadata(config, model_name)

AnemoiCleanupArgs dataclass

Arguments for anemoi cleanup.

path instance-attribute

command = 'unused' class-attribute instance-attribute

delta = None class-attribute instance-attribute

AnemoiDatasetStatus

Bases: NamedTuple

Status of an Anemoi dataset.

copy_in_progress instance-attribute

download_complete instance-attribute

is_finalised instance-attribute

AnemoiFinaliseArgs dataclass

Arguments for anemoi finalise.

path instance-attribute

recipe instance-attribute

command = 'unused' class-attribute instance-attribute

AnemoiInitArgs dataclass

Arguments for anemoi init.

path instance-attribute

recipe instance-attribute

command = 'unused' class-attribute instance-attribute

overwrite = False class-attribute instance-attribute

AnemoiInspectArgs dataclass

Arguments for anemoi inspect.

detailed instance-attribute

path instance-attribute

progress instance-attribute

size instance-attribute

statistics instance-attribute

AnemoiLoadArgs dataclass

Arguments for anemoi load.

path instance-attribute

recipe instance-attribute

command = 'unused' class-attribute instance-attribute

DataloaderArgs

Bases: TypedDict

Arguments for the data loader.

batch_sampler instance-attribute

batch_size instance-attribute

drop_last instance-attribute

num_workers instance-attribute

prefetch_factor instance-attribute

persistent_workers instance-attribute

sampler instance-attribute

worker_init_fn instance-attribute

DiffColourmapSpec

Bases: NamedTuple

Specify the colour scale used for a difference panel.

Attributes:

Name Type Description
norm Normalize | None

Normalisation for mapping values to colours (e.g. TwoSlopeNorm for signed diffs).
vmin float | None

Lower bound if no norm is provided.
vmax float | None

Upper bound if no norm is provided.
cmap str

Matplotlib colourmap name.

norm instance-attribute

vmin instance-attribute

vmax instance-attribute

cmap instance-attribute

Metadata dataclass

Structured metadata extracted from training configuration.

Attributes:

Name Type Description
model str | None

Model name (if available).
max_epochs int | None

Maximum number of training epochs (if available).
current_epoch int | None

Current training epoch (if available).
start str | None

Training start date string (if available).
end str | None

Training end date string (if available).
cadence str | None

Training data cadence string (if available).
n_points int | None

Number of training points calculated from date range and cadence.
vars_by_source dict[str, list[str]] | None

Dictionary mapping dataset source names to lists of variable names.
n_history_steps int | None

Number of history steps used as model input window (days).

model = None class-attribute instance-attribute

max_epochs = None class-attribute instance-attribute

current_epoch = None class-attribute instance-attribute

start = None class-attribute instance-attribute

end = None class-attribute instance-attribute

cadence = None class-attribute instance-attribute

n_points = None class-attribute instance-attribute

n_history_steps = None class-attribute instance-attribute

vars_by_source = None class-attribute instance-attribute

ProcessorOutput dataclass

Output of a processor rollout step.

prediction instance-attribute

loss = None class-attribute instance-attribute

UncertaintyArrays

Bases: NamedTuple

The observed, predicted and uncertainty arrays for a standardised difference.

ground_truth instance-attribute

prediction instance-attribute

uncertainty instance-attribute