superneuroabm.core package

Submodules

superneuroabm.core.agent module

class superneuroabm.core.agent.AgentFactory

Bases: object

property breeds: List[Breed]

Returns the breeds registered in the model

Returns:

A list of currently registered breeds.

create_agent(breed: Breed, **kwargs) → int
Creates and agent of the given breed initialized with the properties given in

**kwargs.

Parameters:

  • breed – Breed definition of agent

  • **kwargs –

    named arguments of agent properties. Names much match properties already registered in breed.

Returns:

Agent ID

generate_agent_data_tensors(use_cuda=True) → List[List[Any]] | Dict[str, List[Any]]
get_agent_property_value(property_name: str, agent_id: int) → Any
Returns the value of the specified property_name of the agent with

agent_id

Parameters:

  • property_name – str name of property as registered in the breed.

  • agent_id – Agent’s id as returned by create_agent

Returns:

value of property_name property for agent of agent_id

get_agents_with(query: Callable) → Dict[int, List[Any]]
Returns an Dict, key: agent_id value: List of properties, of the agents that satisfy

the query. Query must be a callable that returns a boolean and accepts **kwargs where arguments may with breed property names may be accepted and used to form query logic.

Parameters:

query – Callable that takes agent data as dict and returns List of agent data

Returns:

Dict of agent_id: List of properties

property num_agents: int
Returns number of agents. Agents are not removed if they are killed at the

moment.

register_breed(breed: Breed) → None
Registered agent breed in the model so that agents can be created under

this definition.

Parameters:

breed – Breed definition of agent

set_agent_property_value(property_name: str, agent_id: int, value: Any, dims: List[int] | None = None) → None
Sets the property of property_name for the agent with agent_id with

value.

Parameters:

  • property_name – str name of property as registered in the breed.

  • agent_id – Agent’s id as returned by create_agent

  • value – New value for property

  • dims – Optional dimensions of property_value as List of ints. Not required, but improves performance if provied. E.g. [0] if value is a single number like 0, [4, 2] if it is a multi-dimensional list/array with dimensions 4 and 2.

update_agents_properties(equal_side_agent_data_tensors=typing.List[typing.List[typing.Any]], use_cuda=True) → None
class superneuroabm.core.agent.Breed(name: str)

Bases: object

property name: str
property properties: Dict[str, Any]
register_property(name: str, default: int | float | List = nan, max_dims: List[int] | None = None) → None
register_step_func(step_func: Callable, priority: int = 0)

What the agent is supposed to do during a simulation step.

property step_funcs: Dict[int, Callable]

superneuroabm.core.model module

SuperNeuroABM basic Model class

class superneuroabm.core.model.Model

Bases: object

THREADSPERBLOCK = 32
create_agent_of_breed(breed: Breed, **kwargs) → int
get_agent_property_value(id: int, property_name: str) → Any
get_agents_with(query: Callable) → Set[List[Any]]
get_synapse_weight(presynaptic_neuron_id: int, post_synaptic_neuron_id: int) → float

Return synaptic weight. This function is useful if STDP is turned on as weights will be updated during simulation.

Parameters:

  • presynaptic_neuron_id – int

  • presynaptic_neuron_id – int

Returns:

float, final weight of synapse

load(fpath: str) → Model

Loads model from pickle file.

Parameters:

fpath – file path to pickle file.

register_breed(breed: Breed) → None
save(app: Model, fpath: str) → None

Saves model. Must be overridden if additional data pertaining to application must be saved.

Parameters:

  • fpath – file path to save pickle file at

  • app_data – additional application data to be saved.

set_agent_property_value(id: int, property_name: str, value: Any, dims: List[int] | None = None) → None
setup(use_cuda: bool = True) → None

Must be called before first simulate call. Initializes model and resets ticks. Readies step functions and for breeds.

Parameters:

use_cuda – runs model in GPU mode.

simulate(ticks: int, update_data_ticks: int = 1, num_cpu_proc: int = 4) → None
superneuroabm.core.model.generate_gpu_func(n_properties, breed_idx_2_step_func_by_priority, ticks)
superneuroabm.core.model.smap(func_args)

superneuroabm.core.util module

superneuroabm.core.util.compress_tensor(arr: Iterable, level: int = 0)
superneuroabm.core.util.convert_to_equal_side_tensor(tensor: List[Any], max_dims: Tuple[int] | None = None) → array

Module contents