model.runtime.loaders.grid_data_loader

Grid data-loading helpers for ADELM.

Module Contents

Classes

GridDataLoader

Load and flatten gridded data for ADELM grid simulation.

Data

__all__

API

class model.runtime.loaders.grid_data_loader.GridDataLoader

Bases: model.runtime.loaders.base_data_loader.BaseDataLoader

Load and flatten gridded data for ADELM grid simulation.

classmethod load_static_grid(path, var_names, lat_index=None, lon_index=None)

classmethod load_driver_month(path, var_names)

classmethod load_driver_variable(path, var_name, lat_index=None, lon_index=None)

classmethod load_coords(path)

classmethod load_time_values(path)

static clip_coords_to_bbox(lat, lon, bbox)

static clip_static(arr, lat_index=None, lon_index=None)

static clip_drivers(arr, lat_index=None, lon_index=None)

classmethod all_vars(path)

static resolve_name(nc_name, fcover_map, attris_map, params_map)

static flatten_static(arr, pixel_idx, device)

static flatten_drivers(arr, pixel_idx, device)

static unflatten_static(flat, grid_shape, pixel_idx)

static unflatten_layered(flat, grid_shape, pixel_idx)

static gapfill_interior(arr, passes=4)

Fill small interior NaN gaps by iterative finite-neighbour averaging.

Each pass replaces NaN cells that have at least one finite 8-neighbour with the mean of those neighbours. Bounded passes keeps large gaps (e.g. open ocean) NaN while filling the few-cell holes that some static layers (notably SoilGrids) have over wide rivers / water bodies / cities – holes that otherwise drop those cells from the valid-pixel mask and make the routed river vanish where it flows into them.

classmethod build_valid_pixel_mask(static_arrays)

model.runtime.loaders.grid_data_loader.__all__

[‘GridDataLoader’]