xugrid.UgridDataset.from_structured2d#

static UgridDataset.from_structured2d(dataset: Dataset, topology: dict | None = None) → UgridDataset[source]#

Create a UgridDataset from a (structured) xarray Dataset.

The spatial dimensions are flattened into a single UGRID face dimension. By default, this method looks for:

  1. “x” and “y” dimensions

  2. “longitude” and “latitude” dimensions

  3. “axis” attributes of “X” or “Y” on coordinates

  4. “standard_name” attributes of “longitude”, “latitude”, “projection_x_coordinate”, or “projection_y_coordinate” on coordinate variables

Parameters:

  • dataset (xr.Dataset) – The structured dataset to convert.

  • topology (dict, optional) – Either: * A mapping of topology name to (x, y) coordinate names * A mapping of topology name to a dict containing: - “x”: x-dimension name - “y”: y-dimension name - “bounds_x”: x-bounds variable name - “bounds_y”: y-bounds variable name Defaults to {“mesh2d”: (None, None)}.

Returns:

The unstructured grid dataset.

Return type:

UgridDataset

Notes

When using bounds, they should have one of these shapes: * x bounds: (M, 2) or (N, M, 4) * y bounds: (N, 2) or (N, M, 4) where N is the number of rows (along y) and M is columns (along x). Cells with NaN bounds coordinates are omitted.

Examples

Basic usage with default coordinate names:

>>> uds = xugrid.UgridDataset.from_structured2d(dataset)

Specifying custom coordinate names:

>>> uds = xugrid.UgridDataset.from_structured2d(
...     dataset,
...     topology={"my_mesh2d": {"x": "xc", "y": "yc"}}
... )

Multiple grid topologies in a single dataset:

>>> uds = xugrid.UgridDataset.from_structured2d(
...     dataset,
...     topology={
...         "mesh2d_xy": {"x": "x", "y": "y"},
...         "mesh2d_lonlat": {"x": "lon", "y": "lat"}
...     }
... )

Using bounds for non-monotonic coordinates (e.g., curvilinear grids):

>>> uds = xugrid.UgridDataset.from_structured2d(
...     dataset,
...     topology={
...         "my_mesh2d": {
...             "x": "M",
...             "y": "N",
...             "bounds_x": "grid_x",
...             "bounds_y": "grid_y"
...         }
...     }
... )