dctools.data.datasets.dataloader

Dataloder.

Functions

add_coords_as_dims(ds[, coords])	Add given coordinates as dimensions to all data variables in the dataset.
add_time_dim(ds, input_df, n_points_dim, ...)	Ensure that dataset has a 'time' dimension compatible with swath/n-points structure.
concat_with_dim(datasets, concat_dim[, sort])	Concatenate datasets along a dimension eagerly.
concat_with_dim_delayed(datasets, concat_dim)	Concatenate datasets along a dimension using dask.delayed.
filter_by_time(df, t0, t1)	Filter the DataFrame to keep only entries where the interval overlaps with [t0, t1].
preprocess_argo_profiles(profile_sources, ...)	Load ARGO data through ArgoManager for a single time window.
preprocess_batch_obs_files(local_paths, ...)	Preprocess all unique observation files on the driver into a single zarr.
preprocess_one_npoints(source, is_swath, ...)	Preprocess a single N-point dataset (e.g., swath or track).
swath_to_points(ds[, drop_coords, ...])	Convert a swath-style Dataset into a 1D point collection along 'n_points'.

Classes

EvaluationDataloader(params)	Class to manage loading and batching of evaluation data.
ObservationDataViewer(source, load_fn, ...)	Class to view and preprocess observation data.