immunity

immunity

Variants and immunity for Covasim on the Starsim base.

cv.variant adds a co-circulating variant to a sim. Since there is a single cv.COVID module with an internal variant axis, cv.variant is not a module: as in v3, it registers the variant’s parameters with the COVID module and seeds infections with it on the days it is introduced. cv.Sim(variants=[...]) registers each into the module before the states are allocated. The cross-immunity matrix builder (build_immunity_matrix) and the neutralizing antibody (NAb) waning functions live here too, and are used by cv.COVID and cv.CrossImmunity (connectors.py). As in v3 unless noted.

Classes

Name	Description
variant	Add a new variant to the sim.

variant

immunity.variant(variant, days, label=None, n_imports=1, rescale=True)

Add a new variant to the sim.

Parameters

Name	Type	Description	Default
variant	str / dict	name of a predefined variant ('alpha', 'delta', …), or a dict of the per-variant parameters (rel_beta/rel_symp_prob/ rel_severe_prob/rel_crit_prob/rel_death_prob).	required
days	int / str / list	day(s) on which the variant is introduced, as day indices or dates (e.g. '2020-03-15').	required
label	str	if variant is a dict, the variant’s name (dict key).	None
n_imports	int	number of infections to import on each introduction day.	1
rescale	bool	whether to scale n_imports down by pop_scale (so the number of real introductions is preserved under population scaling).	True

Examples::

alpha = cv.variant('alpha', days=10, n_imports=20)
delta = cv.variant('delta', days='2020-04-01', n_imports=20)
sim   = cv.Sim(variants=[alpha, delta]).run()

my_var = cv.variant(variant={'rel_beta': 2.5}, label='My variant', days=20) # A custom variant

Methods

Name	Description
apply	On a matched introduction day, seed n_imports susceptibles with this variant.
initialize	Register this variant into the single COVID module’s variant axis (grow nv).
parse	Unpack the variant info, given as either a predefined-name string or a pars dict.
process_days	Convert the days (day indices or dates) to day indices, as for interventions; called by cv.COVID.init_pre()

apply

immunity.variant.apply(covid)

On a matched introduction day, seed n_imports susceptibles with this variant.

initialize

immunity.variant.initialize(covid)

Register this variant into the single COVID module’s variant axis (grow nv).

parse

immunity.variant.parse(variant=None, label=None)

Unpack the variant info, given as either a predefined-name string or a pars dict.

process_days

immunity.variant.process_days(sim)

Convert the days (day indices or dates) to day indices, as for interventions; called by cv.COVID.init_pre()

Functions

Name	Description
build_immunity_matrix	Build the asymmetric nv x nv cross-immunity matrix (v3 init_immunity).

build_immunity_matrix

immunity.build_immunity_matrix(variant_map, override=None)

Build the asymmetric nv x nv cross-immunity matrix (v3 init_immunity).

matrix[target, source] is the protection a prior source infection confers against a target challenge (diagonal 1.0 = full homologous protection). Start from np.ones((nv,nv)) and overwrite known pairs from get_cross_immunity() (a dict-of-dicts keyed by variant label).

Parameters

Name	Type	Description	Default
variant_map	dict	{index: label} for the variants in the sim.	required
override	array	an explicit nv x nv matrix to use instead of the defaults.	None

Returns

Name	Type	Description
		an nv x nv float ndarray.