dynadojo.wrappers.SystemChecker#

class dynadojo.wrappers.SystemChecker#

Bases: object

Wrapper class for systems that ensures proper input and output handling.

Methods

`__init__`(system)	Initialize the SystemChecker object.
`calc_control_cost`(control)	Wraps calc_control_cost.
`calc_error`(x, y)	Checks that calc_error is called with properly-shaped x and y.
`make_data`(init_conds[, control, timesteps, ...])	Checks that trajectories tensor has the proper shape.
`make_init_conds`(n[, in_dist])	Verifies initial condition matrix is the right shape.

Attributes

`embed_dim`	The embedded dimension of the underlying system.
`latent_dim`	The latent dimension of the underlying system.
`seed`	The seed for the underlying system.

__init__(system)#

Initialize the SystemChecker object.

Example

>>> from dynadojo.systems.lds import LDSystem
>>> SystemChecker(LDSystem())
<SystemChecker<LDSystem>>

calc_control_cost(control)#

Wraps calc_control_cost.

Parameters:: control (np.ndarray) – (n, timesteps, embed_dim) control tensor
Returns:: (n,) control costs vector
Return type:: np.ndarray

calc_error(x, y)#

Checks that calc_error is called with properly-shaped x and y.

Parameters:

Returns:

Error between x and y.

Return type:

float

make_data(init_conds, control=None, timesteps=1, noisy=False)#

Checks that trajectories tensor has the proper shape.

Parameters:

init_conds (np.ndarray) – (n, embed_dim) initial conditions matrix
control (np.ndarray) – (n, embed_dim) initial conditions matrix
timesteps (int) – timesteps per training trajectory (per action horizon)
noisy (bool) – If True, add noise to trajectories. Defaults to False. If False, no noise is added.

Returns:

(n, timesteps, embed_dim) trajectories tensor

Return type:

np.ndarray

make_init_conds(n, in_dist=True)#

Verifies initial condition matrix is the right shape.

Parameters:

n (int) – Number of initial conditions.
in_dist (bool) – If True, generate in-distribution initial conditions. Defaults to True. If False, generate out-of-distribution initial conditions.

Returns:

(n, embed_dim) Initial conditions matrix.

Return type:

np.ndarray