integrator.distributed_integration module#

integrator.distributed_integration.actor_process_stack(start_idx, end_idx)#
integrator.distributed_integration.actor_get_attribute(attr_name)#
integrator.distributed_integration.actor_generic_method(function_name, *args, **kwargs)#
integrator.distributed_integration.get_actors()#
class integrator.distributed_integration.DistributedStackIntegrator(ai_config, resources, dataset=None, output_file=None, n_images_per_integrator='auto', logger=None, extra_options=None, **integrator_kwargs)#

Bases: integrator.base.DistributedIntegrator

A class for distributing StackIntegrator over SLURM.

Initialize a DistributedStackIntegrator.

Parameters

  • ai_config (AIConfiguration) – Azimuthal Integration configuration

  • processing_resources (any) – Data structure describing computational resources

  • dataset (DatasetParser, optional) – XRD dataset information object. If not provided, set_new_dataset() will have to be called prior to integrate_dataset().

  • output_file (str, optional) – Path where the integrated data will be stored (hdf5 file)

  • n_images_per_integrator (int, optional) – Number of images to process at each stack. By default it is automatically inferred by inspecting the dataset file

  • logger (Logger) – Logger object

  • extra_options (dict, optional) –

    Dictionary of advanced options. Current values are:
    • ”create_h5_subfolders”: True

      Whether to create a sub-directory to store files that contain the result of each integrated stack

    • ”scan_num_as_h5_entry”: False

      Whether to use the current scan number as HDF5 entry.

  • class. (The other named arguments (**kwargs) are passed to the StackIntegrator) –

get_workers()#
workers_are_alive()#

Return True if the computation resources are still available, False otherwise.

set_new_dataset(dataset, output_file, timeout=None)#
distribute_integration_tasks(n_images=None, check_workers=False)#
integrate_dataset(timeout=None, retry_failed_tasks=0)#
get_eta()#
create_output_files()#
class integrator.distributed_integration.LocalDistributedStackIntegrator(ai_config, resources, dataset=None, output_file=None, n_images_per_integrator='auto', logger=None, extra_options=None, **integrator_kwargs)#

Bases: integrator.distributed_integration.DistributedStackIntegrator

A class for distributing stack integration on the local machine

Initialize a DistributedStackIntegrator.

Parameters

  • ai_config (AIConfiguration) – Azimuthal Integration configuration

  • processing_resources (any) – Data structure describing computational resources

  • dataset (DatasetParser, optional) – XRD dataset information object. If not provided, set_new_dataset() will have to be called prior to integrate_dataset().

  • output_file (str, optional) – Path where the integrated data will be stored (hdf5 file)

  • n_images_per_integrator (int, optional) – Number of images to process at each stack. By default it is automatically inferred by inspecting the dataset file

  • logger (Logger) – Logger object

  • extra_options (dict, optional) –

    Dictionary of advanced options. Current values are:
    • ”create_h5_subfolders”: True

      Whether to create a sub-directory to store files that contain the result of each integrated stack

    • ”scan_num_as_h5_entry”: False

      Whether to use the current scan number as HDF5 entry.

  • class. (The other named arguments (**kwargs) are passed to the StackIntegrator) –