Parameters

Parameters to define multi-simulation sweeps.

Parameter

class onscale.Parameter(name: str, start_val: float | None = None, end_val: float | None = None, steps: int = 1, use_log_scale: bool = False, inverse_log: bool = False, min_val: float | None = None, max_val: float | None = None)

A parameter range that will be used to generate simulation sweeps.

Examples

>>> import onscale as on
>>> import numpy as np
>>>
>>> # Testing
>>> with on.Simulation('example') as sim:
...     on.Parameter('test-param', 0, 100, 10)

Parameters:

name – Name for this parameter.
start_val – Start value for the sweep.
end_val – End value for the sweep.
steps – The number of sweep steps between start_val and end_val.
use_log_scale – if true, the logarithmic scale should be used. default to False.
inverse_log – if true, the logarithmic scale is inverse. default to False.

static from_dict(data: Dict[str, Any], sim: Simulation | None = None) → Node: Construct this node from serialized dictionary data

sim_count(sim=None) → int: Defines how many swept simulations this variable implies

to_json() → str: Convert this node into JSON representation

to_list() → List[float]: Return enumerated values of this parameter

property uuid: str: Get the UUID associated with this node.

CustomParameter

class onscale.CustomParameter(name: str, values: List[float] | ndarray)

User-defined array of values to use for parametric sweep.

Parameters:

name – Name for this parameter.
values – A 1D array-like object enumerating values to sweep.

static from_dict(data: Dict[str, Any], sim: Simulation | None = None) → Node: Construct this node from serialized dictionary data

sim_count(sim=None) → int: Defines how many swept simulations this variable implies

to_json() → str: Convert this node into JSON representation

to_list() → List[float]: Return enumerated values of this parameter

property uuid: str: Get the UUID associated with this node.

ParameterGroup

class onscale.ParameterGroup(alias: str | None = None)

A group of parameter values. Parameters can only belong to 1 group.

The purpose of a ParameterGroup is to iterate over multiple parameters together, instead of taking all possible combinations of them.

All parameters have a “sim_count”, indicating how many simulations will be generated from them. If multiple parameters are defined in a simulation, the resulting sweep will have the product of all parameter “sim_count” number of simulations. If multiple parameter are defined in a ParameterGroup, then all of those parameters must have the same “sim_count”, and the resulting sweep will have that “sim_count” number of simulations.

Parameters:: alias – Optional alias for this node.

static from_dict(data: Dict[str, Any], sim: Simulation | None = None) → Node: Construct this node from serialized dictionary data

to_json() → str: Convert this node into JSON representation

property uuid: str: Get the UUID associated with this node.

onscale.parametrize(*args: List[float] | ndarray) → List[CustomParameter]

Create a series of grouped parameters from arrays of values.

Examples

>>> import numpy as np
>>> import onscale as on
>>>
>>> # Sweep a force vector 360deg around the z-axis over 20 sims
>>> with on.Simulation('example') as sim:
...     theta = np.linspace(0, 2 * np.pi, 20)
...     x = np.cos(theta)
...     y = np.sin(theta)
...     z = np.zeros(len(x))
...     x, y, z = on.parametrize(x, y, z)
...     force = on.loads.Force(250, [x, y, z])

Parameters:

args – A series of array-like objects. All of these arrays must have the same length.

Returns:

A list of CustomParameter objects, all grouped together. Returns: one CustomParameter object for each array given.