Previous topic

hyperion.grid.AMRGrid

Next topic

Contributing to Hyperion

This Page

hyperion.grid.OctreeGrid

class hyperion.grid.OctreeGrid(*args)

An octree grid.

To initialize an Octree object, use:

>>> grid = OctreeGrid(x, y, z, dx, dy, dz, refined)

where x, y, and z are the cartesian coordinates of the center of the grid, dx, dy, and dz are the half-widths of the grid, and refined is a sequence of boolean values that indicate whether a given cell is refined.

The first value of the refined sequence indicates whether the parent cell is sub-divided. If it is, then the the second element indicates whether the first cell of the parent cell is sub-divided. If it isn’t, then the next value indicates whether the second cell of the parent cell is sub-divided. If it is, then we need to specify the booleans for all the children of that cell before we move to the third cell of the parent cell.

For example, the simplest grid is a single cell that is not sub-divided:

refined = [False]

The next simplest grid is a single grid cell that is only sub-divided once:

refined = [True, False, False, False, False, False, False, False, False]

It is easier to picture this as a hierarchy:

refined = [True,
             False,
             False,
             False,
             False,
             False,
             False,
             False,
             False,
             ]

If we sub-divide the third sub-cell in the parent cell into cells that are themselves not sub-divided, we get:

refined = [True,
             False,
             False,
             True,
               False,
               False,
               False,
               False,
               False,
               False,
               False,
               False,
             False,
             False,
             False,
             False,
             False,
             ]

and so on. The order of the sub-cells is first along x, then along y, then along z.

OctreeGrid objects may contain multiple quantities (e.g. density, specific energy). To access these, you can specify the name of the quantity as an item:

>>> grid['density']

which is no longer an OctreeGrid object, but a OctreeGridView object. When setting this for the first time, this can be set either to another OctreeGridView object, an external h5py link, or an empty list. For example, the following should work:

>>> grid['density_new'] = grid['density']

OctreeGridView objects allow the specific dust population to be selected as an index:

>>> grid['density'][0]

Which is also an OctreeGridView object.

Methods

set_walls(x, y, z, dx, dy, dz, refined)
read(group[, quantities]) Read the geometry and physical quantities from an octree grid
read_geometry(group) Read in geometry information from a cartesian grid
read_quantities(group[, quantities]) Read in physical quantities from a cartesian grid
write(group[, quantities, copy, ...]) Write out the octree grid
write_single_array(group, name, array[, ...]) Write out a single quantity, checking for consistency with geometry
add_derived_quantity(name, function)

Methods (detail)

set_walls(x, y, z, dx, dy, dz, refined)
read(group, quantities='all')

Read the geometry and physical quantities from an octree grid

Parameters :

group : h5py.Group

The HDF5 group to read the grid from. This group should contain groups named ‘Geometry’ and ‘Quantities’.

quantities : ‘all’ or list

Which physical quantities to read in. Use ‘all’ to read in all quantities or a list of strings to read only specific quantities.

read_geometry(group)

Read in geometry information from a cartesian grid

Parameters :

group : h5py.Group

The HDF5 group to read the grid geometry from.

read_quantities(group, quantities='all')

Read in physical quantities from a cartesian grid

Parameters :

group : h5py.Group

The HDF5 group to read the grid quantities from

quantities : ‘all’ or list

Which physical quantities to read in. Use ‘all’ to read in all quantities or a list of strings to read only specific quantities.

write(group, quantities='all', copy=True, absolute_paths=False, compression=True, wall_dtype=<type 'float'>, physics_dtype=<type 'float'>)

Write out the octree grid

Parameters :

group : h5py.Group

The HDF5 group to write the grid to

quantities : ‘all’ or list

Which physical quantities to write out. Use ‘all’ to write out all quantities or a list of strings to write only specific quantities.

copy : bool

Whether to copy external links, or leave them as links.

absolute_paths : bool

If copy is False, then this indicates whether to use absolute or relative paths for links.

compression : bool

Whether to compress the arrays in the HDF5 file

wall_dtype : type

The datatype to use to write the wall positions (ignored for this kind of grid)

physics_dtype : type

The datatype to use to write the physical quantities

write_single_array(group, name, array, copy=True, absolute_paths=False, compression=True, physics_dtype=<type 'float'>)

Write out a single quantity, checking for consistency with geometry

Parameters :

group : h5py.Group

The HDF5 group to write the grid to

name : str

The name of the array in the group

array : np.ndarray

The array to write out

copy : bool

Whether to copy external links, or leave them as links.

absolute_paths : bool

If copy is False, then this indicates whether to use absolute or relative paths for links.

compression : bool

Whether to compress the arrays in the HDF5 file

wall_dtype : type

The datatype to use to write the wall positions

physics_dtype : type

The datatype to use to write the physical quantities

add_derived_quantity(name, function)
class hyperion.grid.OctreeGridView(grid, quantity)

Methods

append(grid) Used to append quantities from another grid
add(grid) Used to add quantities from another grid

Methods (detail)

append(grid)

Used to append quantities from another grid

Parameters :

grid : 1D Numpy array or OctreeGridView instance

The grid to copy the quantity from

add(grid)

Used to add quantities from another grid

Parameters :

grid : 1D Numpy array or OctreeGridView instance

The grid to copy the quantity from