Examples Python#

Details about mesh and input units can be found in Quick Start (C++ & Python).

The use of the Python interface is pretty straight-forward since there is only one method: evaluate(..) or the alternative class GravityEvaluable caching the Polyhedron and intermediate results. Have a look at GravityEvaluable vs. Evaluate Function for further details about the difference. If you strive for maximal performance, use the the class GravityEvaluable. If you quickly want to try out something, use evaluate(..). The polyhedral source can either be a tuple of vertices and faces, or a list of polyhedral mesh files (see Supported Polyhedron Source Files).

The method calls follow the same pattern as the C++ interface. You always define a Polyhedron and then evaluate the polyhedral gravity model using either the GravityEvaluable or evaluate(..).

Free function#

Example 1: Evaluating the gravity model for a given constant density polyhedron defined from within source code for a single point.

from polyhedral_gravity import Polyhedron, evaluate

# Defining every input parameter in the source code
vertices = ...          # (N-3)-array-like of type float
faces = ...             # (N-3)-array-like of type int
density = ...           # float
computation_point = ... # (3)-array-like

# Create a constant density polyhedron & Evaluate the gravity model
# Notice that the third argument could also be a list of points
# Returns a tuple of potential, acceleration and tensor
# If computation_point would be a (N,3)-array, the output would be list of triplets!
polyhedron = Polyhedron((vertices, faces), density)
potential, acceleration, tensor = evaluate(polyhedron, computation_point, parallel=True)

Example 2a: Evaluating the gravity model for a given constant density polyhedron in some source files for several points. Of course, you can also use keyword arguments for the parameters. We also check that the polyhedron’s plane unit normals are actually outwards pointing. This will raise a ValueError if at least on plane unit normal is inwards pointing.

from polyhedral_gravity import Polyhedron, evaluate, PolyhedronIntegrity, NormalOrientation

    # Reading the vertices and files from a .node and .face file
    file_vertices = '___.node'      # str, path to file
    file_nodes = '___.face'         # str, path to file
    density = ...                   # float
    computation_points = ...        # (N,3)-array-like

    # Evaluate the gravity model
    # Notice that the last argument could also be a list of points
    # Returns a list of tuple of potential, acceleration and tensor
    polyhedron = Polyhedron(
        polyhedral_source=[file_vertices, file_nodes],
        density=density,
        normal_orientation=NormalOrientation.OUTWARDS,
        integrity_check=PolyhedronIntegrity.VERIFY,
    )
    results = evaluate(
        polyhedron=polyhedron,
        computation_points=computation_points,
        parallel=True,
    )

Example 2b: Evaluating the gravity model for a given constant density polyhedron in some source files for a specific point. We also check that the polyhedron’s plane unit normals are actually outwards pointing. We don’t specify this here, as OUTWARDS is the default. It will raise a ValueError if at least on plane unit normal is inwards pointing.

from polyhedral_gravity import Polyhedron, evaluate, PolyhedronIntegrity

    # Reading the vertices and files from a single .mesh file
    file = '___.mesh'       # str, path to file
    density = ...           # float
    computation_point = ... # (3)-array-like

    # Evaluate the gravity model
    # Notice that the last argument could also be a list of points
    # Returns a tuple of potential, acceleration and tensor
    # If computation_point would be a (N,3)-array, the output would be list of triplets!
    polyhedron = Polyhedron(
        polyhedral_source=[file],
        density=density,
        integrity_check=PolyhedronIntegrity.VERIFY,
    )
    potential, acceleration, tensor = evaluate(polyhedron, computation_point)

For example 2a and 2b, refer to Supported Polyhedron Source Files to view the available options for polyhedral input.

Example 3a: Here explicitly disable the security check. We won’t get an exception if the plane unit normals are not oriented as specified, but we also don’t pay for the check with quadratic runtime complexity!

from polyhedral_gravity import Polyhedron, evaluate, PolyhedronIntegrity, NormalOrientation

# Defining every input parameter in the source code
vertices = ...          # (N-3)-array-like of type float
faces = ...             # (N-3)-array-like of type int
density = ...           # float
computation_point = ... # (3)-array-like

# Evaluate the gravity model
# Returns a tuple of potential, acceleration and tensor
# If computation_point would be a (N,3)-array, the output would be list of triplets!
polyhedron = Polyhedron(
    polyhedral_source=(vertices, faces),
    density=density,
    normal_orientation=NormalOrientation.OUTWARDS,
    integrity_check=PolyhedronIntegrity.DISABLE,
)
potential, acceleration, tensor = evaluate(polyhedron, computation_point)

Example 3b: Here we use the HEAL option. This guarantees a valid polyhedron. But the ordering of the faces array and the normal_orientation might differ. And we also need to pay the additional quadratic runtime for the checking algorithmus.

from polyhedral_gravity import Polyhedron, evaluate, PolyhedronIntegrity, NormalOrientation

# Defining every input parameter in the source code
vertices = ...          # (N-3)-array-like of type float
faces = ...             # (N-3)-array-like of type int
density = ...           # float
computation_point = ... # (3)-array-like

# Actually, the normal_orientation doesn't matter! We could the argument
# as HEAL guarantees a valid polyhedron
# but the polyhedron might different properties polyhedron.faces
# and polyhedron.normal_orientation than specified
polyhedron = Polyhedron(
    polyhedral_source=(vertices, faces),
    density=density,
    normal_orientation=NormalOrientation.OUTWARDS,
    integrity_check=PolyhedronIntegrity.HEAL,
)
potential, acceleration, tensor = evaluate(polyhedron, computation_point)

GravityEvaluable#

Use the GravityEvaluable class to cache the polyhedron data over multiple calls. This drastically improves the performance, as the polyhedral data is “stored” on the C++ side, rather than being converted from Python to C++ for every call.

This approach is especially useful one wants to calculate multiple points for the same polyhedron, but the points are not known in advance (e.g. when propagating a spacecraft). Have a look at the example below to see how to use the GravityEvaluable class.

from polyhedral_gravity import Polyhedron, GravityEvaluable, evaluate, PolyhedronIntegrity

# Defining every input parameter in the source code
vertices = ...           # (N-3)-array-like of type float
faces = ...              # (N-3)-array-like of type int
density = ...            # float
computation_points = ... # (N,3)-array-like

# Definition of the Polyhedron in previous examples
polyhedron = Polyhedron(
    polyhedral_source=(vertices, faces),
    density=density,
    integrity_check=PolyhedronIntegrity.HEAL,
)

# Create the evaluable object
evaluable = GravityEvaluable(polyhedron, density)

for point in computation_points:
    # Evaluate the gravity model for single points (3)-array-like
    potential, acceleration, tensor = evaluable(point, parallel=True)

# Due to the GravityEvaluable's caching the above for-loop is nearly
# as fast as the following (find the runtime details below), which returns
# a list of triplets comprising potential, acceleration, tensor
results = evaluable(computation_points, parallel=True)

Examples Python

Contents

Examples Python#

Free function#

GravityEvaluable#