patterns

patterns

Classes that support introducing ‘pre-defined’ spatiotemporal profiles as parameters. These patterns can be based on a callable or on data.

Classes

Name	Description
Pattern	Base class for evaluating a function as a function of an independent
StateVariable	This class is identical to Results, but named to something more generic,

Pattern

patterns.Pattern(
    equation,
    pattern_name=None,
    name=None,
    sim=None,
    module=None,
    debug=False,
    **kwargs,
)

Base class for evaluating a function as a function of an independent variable like age or time, or ‘agent-space’. Based on starsim’s Distribution class can be expanded in functionality.

Parameters

Name	Type	Description	Default
equation	str	required; a simple string the defines the pattern as a function of var	required
pattern_name	str	the name for this class of pattern (e.g. “temporal”)	None
name	str	the name for this particular pattern (e.g. “seasonal_modulation_exposure”)	None
sim	Sim	usually determined on initialization; the sim to use as input to callable parameters	None
module	Module	usually determined on initialization; the module to use as input to callable parameters	None
kwargs	dict	parameters of the pattern	{}

Examples:: pattern = typ.Pattern(“a + var * b”, pars={‘a’:2, ‘b’: 0}) res = pattern(var)

Methods

Name	Description
disp	Return full display of object
evaluate	Generate a discretized representation of the equation for the domain
generate_data	NOTE: The variable name of the actual independent variable
init	Calculate the starting seed and create the RNG
link_module	Shortcut for linking the module
link_sim	Shortcut for linking the sim, only overwriting an existing one if overwrite=True
plot	Plot the pattern
set	Set (change) the equation of this pattern, or one or more of its
show_state	Show the state of the object

disp

patterns.Pattern.disp()

Return full display of object

evaluate

patterns.Pattern.evaluate(var)

Generate a discretized representation of the equation for the domain represented by var.

The argument var can represent time, or age. It can be be a single number, a numpy.ndarray or pandas series???

generate_data

patterns.Pattern.generate_data(vmin=0, vmax=128, step=None)

NOTE: The variable name of the actual independent variable in the equation expression should be named var

init

patterns.Pattern.init(trace=None, module=None, sim=None, force=False)

Calculate the starting seed and create the RNG

link_module

patterns.Pattern.link_module(module=None, overwrite=False)

Shortcut for linking the module

link_sim

patterns.Pattern.link_sim(sim=None, overwrite=False)

Shortcut for linking the sim, only overwriting an existing one if overwrite=True

plot

patterns.Pattern.plot(data_kw=None, fig_kw=None, var_name=None)

Plot the pattern

set

patterns.Pattern.set(*args, equation=None, **kwargs)

Set (change) the equation of this pattern, or one or more of its parameters

show_state

patterns.Pattern.show_state(output=False)

Show the state of the object

StateVariable

patterns.StateVariable()

This class is identical to Results, but named to something more generic, so it can be used in different parts of the code and still make sense from a code-readability perspective. The concept of Results does not necessarily align with the internal-state of a model. A results array may have outputs that are a transformed version of the internal ‘state variables’ and as such ideally we don’t want them to be used directly. A Results array can also map one-to-one to one of the internal state variables, but this is not always guaranteed.

This is a structure that holds an internal state of our system (ie, module) prior to any transformations for output.