avni.models.model3d module#
This script/module contains routines that are used to analyze/visualize the data sets in the standard AVNI format.
- class avni.models.model3d.Model3D(file=None, **kwargs)[source]#
Bases:
objectA class for 3D reference Earth models used in tomography
- property name#
- property num_resolutions#
- property num_realizations#
- add_realization(coef=None, name=None, resolution=None)[source]#
Added a set of realizations to the object. resolution is the tesselation level at which the instance is update with name and coeff (default: None, last resolution).
- add_resolution(metadata=None)[source]#
Added a resolution level to the object. num_realizations is the number of realizations of model coefficients within that object.
- readhdf5(hf, query=None)[source]#
Reads a standard 3D model file from a hdf5 file
- queryif None, use the model available if only one is included.
Choose from query hf.keys() if multiple ones are available
hf: hdf5 handle from h5py
- buildtree3D(resolution=0, dbs_path='/home/pm5113/Github/avni-private/avni/files')[source]#
Build a KDtree interpolant based on the metadata
- average_in_polygon(depth_in_km, parameter, polygon_latitude, polygon_longitude, num_cores=1, orientation='anti-clockwise', threshold=1e-06, grid=1.0, outside=False, mask=None, **kwargs)[source]#
Get the average values within a polygon
parameter: variable whose value will be returned
depth_in_km: depth where values are being queried at (km)
- polygon_latitude,polygon_longitude: closed points that define the polygon.
First and last points need to be the same.
grid: fineness of the grid to use for averaging (in degrees)
num_cores: Number of cores to use for the calculations
orientation: clockwise or anti-clockwise orientation of points specified above
- threshold: limit to which the sum of azimuth check to (-)360 degrees is permitted
to be defined as within the polygon.
outside: give average values outside polygon
average: average value within the polygon of interest
- tree: if kwarg argument to evaluate_at_location has interpolated=True and
tree==None, returns the tree data structure.
- evaluate_unit(parameter, values, units, depth_in_km=None, add_reference=True, resolution=0)[source]#
- get_reference(parameter, depth_in_km, resolution=0, interpolation='linear', dbs_path='/home/pm5113/Github/avni-private/avni/files')[source]#
- evaluate_slice(parameter, data=None, grid=10.0, depth_in_km=None, **kwargs)[source]#
checks whether the data input is a DataArray and the coordinates are compatible
- Parameters
- parameter: 2D or 3D parameter
- dataxarray Dataset or a dictionary with values
- gridsize of pixel in degrees, ignore if data is Dataset
- depth_in_km: list of depths in km for a 3D parameter
- Returns
- data: new or updated xarray Dataset
- evaluate_at_location(parameter, latitude, longitude, depth_in_km=None, resolution=0, realization=0, grid=False, interpolated=False, tree=None, nearest=None, units=None, add_reference=True, dbs_path='/home/pm5113/Github/avni-private/avni/files')[source]#
Evaluate the mode at a location (latitude, longitude,depth)
parameter: variable whose value will be returned
depth_in_km: depth where values are being queried at (km)
- interpolated: If True, use KDTree from a predefined grid. If False, evaluated
exactly using kernel_set instance.
- nearest: number of points to interpolate. If None, defaults to values based on
self[resolution][‘interpolant’]
grid: make a grid by unraveling (depth_in_km,latitude,longitude)
- units: if ‘absolute’ try converting to absolute units of the parameter.
If ‘default’, give the default units from model instance. If None, simply return the sparse array
- add_reference: add the reference paramter value to perturbation, only valid
if units queried is ‘absolute’
- coeff2modelarr(resolution=0, realization=0, parameter=None)[source]#
Convert the coeffiecient matrix from the file to a sparse model array. Add up all the resolutions if a list is provided.
realization : index of the set of coefficients in an ensemble. Default is 0 as there is only one set of coefficients when read from a model file.
resolution : list of resolutions to include the the modelarray
parameter: parameters to select. Default is to include all available.
- readprojbinary(lateral_basis)[source]#
Reads Projection matrix created by plot_3dmod_pm. lateral_basis can be M362 or pixel1
- get_projection(parameter, latitude, longitude, depth_in_km=None, resolution=0, grid=False)[source]#
Get the projection matrix from a lateral basis to another and for particular depths
- depth_in_kmdepth in km where the projection matrix is needed.
If None, returns the projection matrix for the lat/lon and radial basis as a dirac delta.
grid: make a grid by repeating (latitude,longitude) by number of depth_in_km
- projslices(projection, variable, depth, resolution=0, realization=0)[source]#
Projection matrix multiplied by model ensemble. Choses the nearest depth available for the projection.
- reparameterize(model3d, resolution=0, realization=0, interpolated=False, tree=None, nearest=1, dbs_path='/home/pm5113/Github/avni-private/avni/files')[source]#
Inverts for new coefficients in self.data from the coefficients in model3d class
write : output the reparamterized model as a text file
interpolated: assume that it won’t be a simple interpolation
- to_profiles(grid=10.0, type='pixel', resolution=0, realization=0, model_dir='.', interpolated=True)[source]#
converts a model3d class to profiles class
grid: either a grid size of a grid file
interpolant: interpolant type either in pixel or nearest
model_dir: directory to find reference 1D model
- get_resolution(rescovfile=None, LU2symmetric=True, resolution=0, realization=0)[source]#
Reads Resolution or Covariance matrix created by invwdata_pm64 with option -r. R=inv(ATA+DTD)ATA and the name of file is typically outmodel.Resolution.bin First read the matrix and model file, perform checks and create a sparse matrix.