suxarray.grid package

Submodules

suxarray.grid.area module

suxarray.grid.grid module

class suxarray.grid.grid.Grid(grid_obj: Dataset | Grid, ds_sgrid_info: Dataset | None = None, source_grid_spec: str = 'UGRID', source_dims_dict: dict | None = {}, **kwargs)[source]

Bases: Grid

uxarray Grid class for SCHISM See uxarray Grid documentation for the details of the methods and properties

Examples

TBD

classmethod from_dataset(ds_out2d: Dataset, ds_zcoords: Dataset, **kwargs)[source]

Create a Grid object from a SCHISM output 2D dataset and a z-coordinate dataset

Parameters

ds_out2dxr.Dataset: SCHISM output 2D dataset
ds_sgrid_infoxr.Dataset, optional: Extra SCHISM grid information

Returns

Grid

get_strtree(elements: str | None = 'nodes')[source]

intersect(geometry: shapely.geometry.BaseGeometry, element: str | None = 'faces', **kwargs)[source]

Find intersecting elements with a Shapely geometry

Parameters

geometry: shapely.geometry.Geometry element: str, optional

Returns

int: Indices of intersecting elements

isel(**dim_kwargs)[source]

Indexes an unstructured grid along a given dimension (n_node, n_edge, or n_face) and returns a new grid.

Currently only supports inclusive selection, meaning that for cases where node or edge indices are provided, any face that contains that element is included in the resulting subset. This means that additional elements beyond those that were initially provided in the indices will be included. Support for more methods, such as exclusive and clipped indexing is in the works.

Parameters

inverse_indiceslist[str] | set[str] | bool, default=False: Indicates whether to store the original grids indices. Passing True stores the original face indices, other reverse indices can be stored by passing any or all of the following: ([“face”, “edge”, “node”], True)
**dims_kwargs: kwargs: Dimension to index, one of [‘n_node’, ‘n_edge’, ‘n_face’]

Example

>> grid = ux.open_grid(grid_path) >> grid.isel(n_face = [1,2,3,4])

property sgrid_info: Dataset

sgrid_isel(**kwargs)[source]

subset: alias of GridSubsetAccessor

suxarray.grid.integrate module

suxarray.grid.integrate.calc_volumes_from_projected(sxda)[source]

Calculate the volume of prism under a selection of nodes for each layer in the vertical direction.

Parameters

sxdaSxDataArray: SxDataArray with z-coordinates and bottom indices. This function enforces single-timestep behavior.

Returns

ndarray, shape (n_layer-1, n_face): Volume of each prism layer under the selection of nodes.

Notes

For external dask/xarray parallelization over time, use calc_volumes_from_projected_kernel with xarray.apply_ufunc.

suxarray.grid.integrate.integrate_prism_variable(sxda, return_per_layer=False)[source]

Integrate a variable under an element over the water column for nodal data.

Parameters

sxdaSxDataArray: SxDataArray with z-coordinates and bottom indices containing the variable to integrate. Variable must be a concentration or density-like quantity (e.g., chlorophyll, salt, energy) that can be integrated over the water column to produce a total quantity per face element. This function enforces single-timestep behavior.
return_per_layerbool, optional: If True, return per-layer integrated values with shape (n_layer-1, n_face). If False (default), return total integrated values per face with shape (n_face).

Returns

ndarray: If return_per_layer=False: shape (n_face), integrated variable per face element summed over all layers. If return_per_layer=True: shape (n_layer-1, n_face), integrated variable per layer and face.

suxarray.grid.integrate.projected_volume_ufunc_kernel(z, node_x, node_y, connectivity, bottom_indices)[source]

Array-kernel companion for calc_volumes_from_projected.

This helper is designed for external orchestration with xarray.apply_ufunc(..., dask='parallelized') where z is passed as an ndarray block/chunk.

Parameters

zndarray, shape (n_node, n_layer): Single-timestep z-coordinates.
node_xndarray, shape (n_node,): x coordinates of nodes.
node_yndarray, shape (n_node,): y coordinates of nodes.
connectivityndarray, shape (n_face, n_node_per_face): Node indices for each face.
bottom_indicesndarray, shape (n_node,): 0-based bottom layer index for each node.

Returns

ndarray, shape (n_layer-1, n_face): Prism volumes per layer and face for one timestep.

suxarray.grid.neighbors module

class suxarray.grid.neighbors.STRTree(grid, elements: str | None = 'nodes')[source]

Bases: object

property faces

nearest(geometry)[source]

property points

query(geometry, predicate=None, distance=None)[source]

Module contents

class suxarray.grid.Grid(grid_obj: Dataset | Grid, ds_sgrid_info: Dataset | None = None, source_grid_spec: str = 'UGRID', source_dims_dict: dict | None = {}, **kwargs)[source]

Bases: Grid

uxarray Grid class for SCHISM See uxarray Grid documentation for the details of the methods and properties

Examples

TBD

classmethod from_dataset(ds_out2d: Dataset, ds_zcoords: Dataset, **kwargs)[source]

Create a Grid object from a SCHISM output 2D dataset and a z-coordinate dataset

Parameters

ds_out2dxr.Dataset: SCHISM output 2D dataset
ds_sgrid_infoxr.Dataset, optional: Extra SCHISM grid information

Returns

Grid

get_strtree(elements: str | None = 'nodes')[source]

intersect(geometry: shapely.geometry.BaseGeometry, element: str | None = 'faces', **kwargs)[source]

Find intersecting elements with a Shapely geometry

Parameters

geometry: shapely.geometry.Geometry element: str, optional

Returns

int: Indices of intersecting elements

isel(**dim_kwargs)[source]

Indexes an unstructured grid along a given dimension (n_node, n_edge, or n_face) and returns a new grid.

Currently only supports inclusive selection, meaning that for cases where node or edge indices are provided, any face that contains that element is included in the resulting subset. This means that additional elements beyond those that were initially provided in the indices will be included. Support for more methods, such as exclusive and clipped indexing is in the works.

Parameters

inverse_indiceslist[str] | set[str] | bool, default=False: Indicates whether to store the original grids indices. Passing True stores the original face indices, other reverse indices can be stored by passing any or all of the following: ([“face”, “edge”, “node”], True)
**dims_kwargs: kwargs: Dimension to index, one of [‘n_node’, ‘n_edge’, ‘n_face’]

Example

>> grid = ux.open_grid(grid_path) >> grid.isel(n_face = [1,2,3,4])

property sgrid_info: Dataset

sgrid_isel(**kwargs)[source]

subset: alias of GridSubsetAccessor