Named Tuples

Named tuples used throughout the package to make code clearer.

class glompo.common.namedtuples.Result

Final result delivered by GloMPOManager.

Create new instance of Result(x, fx, stats, origin)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, x: Sequence[float], fx: float, stats: Dict[str, Any], origin: Dict[str, Any])

Create new instance of Result(x, fx, stats, origin)

__repr__()

Return a nicely formatted representation string

fx

The lowest function value found.

Type:float

origin

Optimizer name, settings, starting point and termination condition.

Type:Dict[str, Any]

stats

Dictionary of statistics surrounding the optimization.

Type:Dict[str, Any]

x

Location in input space producing the lowest function value.

Type:Sequence[float]

class glompo.common.namedtuples.Bound

Class of parameter bounds.

Create new instance of Bound(min, max)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, min: float, max: float)

Create new instance of Bound(min, max)

__repr__()

Return a nicely formatted representation string

max

Upper parameter bound.

Type:float

min

Lower parameter bound.

Type:float

class glompo.common.namedtuples.OptimizerPackage

Package of an initialized optimizer, its unique identifier and multiprocessing variables.

Create new instance of OptimizerPackage(opt_id, optimizer, call_kwargs, signal_pipe, allow_run_event, slots)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, opt_id: int, optimizer: Callable, call_kwargs: Dict[str, Any], signal_pipe: multiprocessing.connection.Connection, allow_run_event: multiprocessing.context.BaseContext.Event, slots: int)

Create new instance of OptimizerPackage(opt_id, optimizer, call_kwargs, signal_pipe, allow_run_event, slots)

__repr__()

Return a nicely formatted representation string

allow_run_event

Used by the manager to pause the optimizer.

Type:multiprocessing.Event

call_kwargs

Dictionary of kwargs send to BaseOptimizer.minimize()

Type:Dict[str, Any]

opt_id

Unique optimizer identification number.

Type:int

optimizer

Instance of BaseOptimizer

Type:Callable

signal_pipe

Used to send messages between the optimizer and manager.

Type:Connection

slots

Maximum number of thread/processes the optimizer may spawn.

Type:int

class glompo.common.namedtuples.IterationResult

Return type of each optimizer iteration.

Create new instance of IterationResult(opt_id, x, fx, extras)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, opt_id: int, x: Sequence[float], fx: float, extras: Sequence[Any])

Create new instance of IterationResult(opt_id, x, fx, extras)

__repr__()

Return a nicely formatted representation string

extras

Possible returns if detailed_call() is used.

Type:Sequence[Any]

fx

Corresponding function evaluation.

Type:float

opt_id

Unique optimizer identification number.

Type:int

x

Location in input space.

Type:Sequence[float]

class glompo.common.namedtuples.ProcessPackage

Package of a running process and its communication channels.

Create new instance of ProcessPackage(process, signal_pipe, allow_run_event, slots)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, process: Union[multiprocessing.context.Process, threading.Thread], signal_pipe: multiprocessing.connection.Connection, allow_run_event: multiprocessing.context.BaseContext.Event, slots: int)

Create new instance of ProcessPackage(process, signal_pipe, allow_run_event, slots)

__repr__()

Return a nicely formatted representation string

allow_run_event

Used by the manager to pause the optimizer.

Type:multiprocessing.Event

process

Process within which optimizer is run.

Type:Union[Process, Thread]

signal_pipe

Used to send messages between the optimizer and manager.

Type:Connection

slots

Maximum number of thread/processes the optimizer may spawn.

Type:int

class glompo.common.namedtuples.OptimizerCheckpoint

Information needed in the manager about initialized optimizers for checkpoint loading.

Create new instance of OptimizerCheckpoint(opt_type, slots)

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.

static __new__(_cls, opt_type: Type[BaseOptimizer], slots: int)

Create new instance of OptimizerCheckpoint(opt_type, slots)

__repr__()

Return a nicely formatted representation string

opt_type

Class of optimizer to be restarted.

Type:Type[BaseOptimizer]

slots

Maximum number of thread/processes the optimizer may spawn.

Type:int