volcapy.synthetic package¶
Submodules¶
volcapy.synthetic.build_synth_data module¶
This code generates a synthetic volcano and corresponding dataset. The generated forward operator, inversion grid, data sites coordinates, etc … will be stored as numpy array files (.npy).
See the documentation of the main function below for more information
on the outputs.
-
volcapy.synthetic.build_synth_data.main()[source]¶ - Returns
- F_synth.npy:
The forward operator
- reg_coords_synth.npy:
A regular grid, stored as a n_cells * n_dims array.
- volcano_inds_synth.npy:
A list specifying the indices (in the regular grid) that correspond to volcano cells.
- data_coords_synth.npy:
Coordinates of the data sites.
- data_values_synth.npy:
(Computed) value of the data measurements.
- density_synth.npy:
The matter density inside the synthetic volcano. Note tha this is on the regular grid, with zeros for cells outside the volcano.
volcapy.synthetic.grid module¶
This submodule contains functions for building artificial irregular grids (topographies) when building synthetic volcanoes.
It can also generate data measurement site on the surface of the topography (sites placed at random) and compute the forward operator associated to the topography/data sites.
-
volcapy.synthetic.grid.build_cone(coords)[source]¶ Given a cubic grid, turn it into a cone.
- Parameters
- coords: ndarray
Array of size n_cells * 3. Contains the coordinates of the center of each cell.
- Returns
- ndarray
1 dimensional array containing indices of cells belonging to the cone.
-
volcapy.synthetic.grid.build_cube(nr_x, res_x, nr_y, res_y, nr_z, res_z)[source]¶ Builds a regular gridded cube.
- Parameters
- nr_x: int
Number of cells in x-dimension.
- res_x: float
Size of cell along x_dimension.
- nr_y: int
- res_y: float
- nr_z: int
- res_z: float
- Returns
- ndarray
Array of size n_cells * 3. Contains the coordinates of the center of each cell.
-
volcapy.synthetic.grid.build_random_cone(coords, nx, ny, nz)[source]¶ Given a cubic grid, turn it into a cone.
- Parameters
- coords: ndarray
Array of size n_cells * 3. Contains the coordinates of the center of each cell.
- nx: int
Number of cells along x-dimension.
- ny: int
- nz: int
- Returns
- ndarray
1 dimensional array containing indices of cells belonging to the cone.
-
volcapy.synthetic.grid.compute_forward(coords, res_x, res_y, res_z, data_coords)[source]¶ Compute the forward operator associated to a given topography/irregular grid. In the end, it only need a list of cells.
- Parameters
- coords: ndarray
Cells centroid coordinates, size n_cell * n_dims.
- res_x: float
Length of a cell in x-direction (meters).
- res_y_float
- res_z: float
- data_coords: ndarray
List of data measurements coordinates, size n_data * n_dims.
- Returns
- ndarray
Forward operator, size n_data * n_cells.
-
volcapy.synthetic.grid.generate_regular_surface_datapoints(xl, xh, nx, yl, yh, ny, zl, zh, nz, offset)[source]¶ Put regularly spaced measurement points on the surface of a cube. Note that there will always be measurement sites at the endpoints of the cube. We need an offset because measerements cannot be directly on the endpoints of a cell because of division by zero in the Bannerjee formula.
- Parameters
- xl: float
Lower x-coordinate of the cube.
- xh: float
Higher x-coordinate of the cube.
- nx: int
Number of measurments in x-dimension.
- yl: float
- yh: float
- ny: int
- zl: float
- zh: float
- nz: int
- offset: float
Displace the measurements sites by an offset outside of the cube to avoid division by zero.
- Returns
- ndarray
Coordinates of the measurement sites, size n_data * n_dims.
volcapy.synthetic.reconstruct module¶
This script runs the inversion on the synthetic dataset created using
build_synth_data.
Note that hyperparameters have to be manually specified, so one should run
train before in order to determine the optimal hyperparameters.
volcapy.synthetic.temp module¶
volcapy.synthetic.test_voxel module¶
volcapy.synthetic.train module¶
volcapy.synthetic.try_vtk_str_grid module¶
volcapy.synthetic.vtkutils module¶
Utilities to convert inversion data to VTK format for 3d visualization.