surface_grid

Functions to define 2D grids based on coordinates

coordinates_to_grid(x, y=None, z=None)

Convert coordinates to grid coordinates

X, Y, Z = coordinates_to_grid(x, y, z)

Parameters:

Name	Type	Description	Default
x	Array | tuple[list, list, list]	array of coordinates, either x=(n,) or x=(n,3).	required
y	Array | None	array of y coordinates, or None if x=(n,3)	None
z	Array | None	array of z coordinates, or None if x=(n,3)	None

Returns:

Type	Description
	x,y,z grid coordinates, each with shape ~(a,b) where a+b ~ n

Source code in mmg_toolbox/utils/surface_grid.py

def coordinates_to_grid(x: Array | tuple[list, list, list], y: Array | None = None, z: Array | None = None):
    """
    Convert coordinates to grid coordinates

        X, Y, Z = coordinates_to_grid(x, y, z)

    :param x: array of coordinates, either x=(n,) or x=(n,3).
    :param y: array of y coordinates, or None if x=(n,3)
    :param z: array of z coordinates, or None if x=(n,3)
    :return: x,y,z grid coordinates, each with shape ~(a,b) where a+b ~ n
    """
    x, y, z = _prepare_coordinates(x, y, z)

    x_diff = abs(np.diff(x))
    x_space = min(x_diff[x_diff > 0]) / 2  # reduce spacing to ensure rounding is unique
    y_diff = abs(np.diff(y))
    y_space = min(y_diff[y_diff > 0]) / 2

    x_unique, x_index, x_inv = np.unique(np.round(x / x_space), return_index=True, return_inverse=True)
    y_unique, y_index, y_inv = np.unique(np.round(y / y_space), return_index=True, return_inverse=True)

    grid_z = np.full((len(y_unique), len(x_unique)), np.nan)
    grid_z[y_inv, x_inv] = z
    grid_x, grid_y = np.meshgrid(x[x_index], y[y_index])
    return grid_x, grid_y, grid_z

interpolate_to_grid(x, y=None, z=None, grid_shape=None, method='linear')

Convert coordinates to grid coordinates

X, Y, Z = interpolate_to_grid(x, y, z, (101, 101))

Parameters:

Name	Type	Description	Default
x	Array | tuple[list, list, list]	array of coordinates, either x=(n,) or x=(n,3).	required
y	Array | None	array of y coordinates, or None if x=(n,3)	None
z	Array | None	array of z coordinates, or None if x=(n,3)	None
grid_shape	tuple[int, int] | None	grid shape, (x, y)	None
method	Literal['nearest', 'linear', 'cubic']	interpolation method for scipy.interpolate.griddata. One of: "nearest", "linear", "cubic"	'linear'

Returns:

Type	Description
tuple[ndarray, ndarray, ndarray]	x,y,z grid coordinates, each with shape ~(a,b) where a+b ~ n

Source code in mmg_toolbox/utils/surface_grid.py

def interpolate_to_grid(x: Array | tuple[list, list, list], y: Array | None = None, z: Array | None = None,
                        grid_shape: tuple[int, int] | None = None,
                        method: Literal["nearest", "linear", "cubic"] = "linear") -> tuple[np.ndarray, np.ndarray, np.ndarray]:
    """
    Convert coordinates to grid coordinates

        X, Y, Z = interpolate_to_grid(x, y, z, (101, 101))

    :param x: array of coordinates, either x=(n,) or x=(n,3).
    :param y: array of y coordinates, or None if x=(n,3)
    :param z: array of z coordinates, or None if x=(n,3)
    :param grid_shape: grid shape, (x, y)
    :param method: interpolation method for scipy.interpolate.griddata. One of: "nearest", "linear", "cubic"
    :return: x,y,z grid coordinates, each with shape ~(a,b) where a+b ~ n
    """
    x, y, z = _prepare_coordinates(x, y, z)

    # Create grid
    grid_shape = grid_shape or (len(x), len(x))
    grid_shape = (grid_shape, grid_shape) if isinstance(grid_shape, int) else grid_shape
    x_range = np.linspace(np.min(x), np.max(x), grid_shape[0])
    y_range = np.linspace(np.min(y), np.max(y), grid_shape[1])
    grid_x, grid_y = np.meshgrid(x_range, y_range)
    grid_z = griddata((x, y), z, (grid_x, grid_y), method=method)
    return grid_x, grid_y, grid_z