fluidimage.executors

Executors of computational topologies

The executors are used to execute a topology.

From the user point of view, the executor is chosen from the method fluidimage.topologies.base.TopologyBase.compute(). The default executor is fluidimage.executors.multi_exec_async.MultiExecutorAsync.

There are many executors with different computational strategies. Depending on the computational topology and the hardware, it can be more efficient to chose an executor compared to another.

base	Base class for executors
exec_sequential	Execute a topology sequentially
exec_async	Executor async/await
exec_async_sequential	Executor async/await sequential
multi_exec_async	Multi executors async
multi_exec_sync	Multi executors sync
multi_exec_subproc	Multi executor based on subprocesses
multi_exec_subproc_sync	Multi executor based on subprocesses using exec_seq_for_multi
exec_async_seq_for_multi	Sequential executor for multi executor
exec_seq_for_multi	Sequential executor for multi executor
exec_async_multiproc	Executor async/await + multiprocessing
exec_async_servers	Executor async/await using servers
servers	Servers for exec_async_servers (fluidimage.executors.servers)

fluidimage.executors.get_entry_points(reload=False)

Discover the executors installed

fluidimage.executors.get_executor_names()

Get available executor names

fluidimage.executors.import_executor_class(name)

Import an executor class.

Parameters:

namestr

Executor name.

Returns:

The corresponding executor class.

Functions

afterfork()
get_entry_points([reload])	Discover the executors installed
get_executor_names()	Get available executor names
import_executor_class(name)	Import an executor class.