Decorators¶
action¶
- action(*args, **kwargs)[source]¶
Decorator for action methods in
Batchclasses- Parameters
use_lock (bool or str) – whether to lock an action when a pipeline is executed. It can be bool or a lock name. A pipeline variable with a lock is created in the pipeline during the execution.
no_eval (str or a sequence of str) – parameters to skip from named expression evaluation. A parameter should be passed as a named argument only.
Examples
@action def some_action(self, arg1, arg2): ... @action(no_eval='dst') def calc_offset(self, src, dst=None): ... @action(use_lock=True) def critical_section(self, some_arg, another_arg): ... @action(use_lock='lock_name') def another_critical_section(self, some_arg, another_arg): ...
inbatch_parallel¶
parallel¶
apply_parallel¶
- apply_parallel(*args, **kwargs)[source]¶
Mark class method for transform in its metaclass.
Decorator writes kwargs to the method attribute apply_kwargs, so they can be extracted and used in metaclass.
- Parameters
args – other parameters passed to apply_parallel method of the class where this decorator is being used
kwargs – other parameters passed to apply_parallel method of the class where this decorator is being used
Notes
Redefine the attribute apply_defaults <.Batch.apply_defaults> in the batch class. This is proposed solely for the purposes of brevity — in order to avoid repeated heavily loaded class methods decoration, e.g. @apply_parallel(src=’images’, target=’for’) which in most cases is actually equivalent to simple @apply_parallel assuming that the defaults are redefined for the class whose methods are being transformed.
Note, that if no defaults redefined those from the nearest parent class will be used in
MethodsTransformingMeta.