pydistsim.algorithm.base_algorithm.BaseAlgorithm

class BaseAlgorithm(simulation: Simulation, **kwargs)

Bases: ObserverManagerMixin

Abstract base class for all algorithms.

Currently there are two main subclasses: * NodeAlgorithm used for distributed algorithms * NetworkAlgorithm used for centralized algorithms

When writing new algorithms make them subclass either of NodeAlgorithm or NetworkAlgorithm.

Every algorithm instance has a set of required and default params: * Required params must be given to algorithm initializer as a keyword

arguments.

• Default params can be given to algorithm initializer as a keyword

  arguments, if not their class defines default value.

Note: On algorithm initialization all params are converted to instance attributes.

For example:

class SomeAlgorithm(NodeAlgorithm):

required_params = (‘rp1’,) default_params = {‘dp1’: ‘dv1’,}

>>> net = DirectedNetwork()
>>> alg = SomeAlgorithm(net, rp1='rv1')
>>> alg.rp1
'rv1'
>>> alg.dp1
'dv1'

Params in algorithm subclasses are inherited from its base classes, that is, required params are extended and default are updated: * required_params are union of all required params of their ancestor

classes.

• default_params are updated so default values are overridden in subclasses

Methods

__init__
add_observers
apply_restrictions	Apply all applicable restrictions.
check_restrictions	Check if the restrictions are satisfied.
clear_observers
is_halted	Check if the distributed algorithm has come to an end or deadlock.
is_initialized
notify_observers
reset	Reset the algorithm to its initial state.
step

Attributes

algorithm_restrictions	Tuple of restrictions that must be satisfied for the algorithm to run.
default_params
network
required_params

algorithm_restrictions = ()

Tuple of restrictions that must be satisfied for the algorithm to run.

apply_restrictions()

Apply all applicable restrictions.

check_restrictions()

Check if the restrictions are satisfied. Does not apply ApplicableRestrictions.

is_halted()

Check if the distributed algorithm has come to an end or deadlock.

reset()

Reset the algorithm to its initial state.