deltakit.decode.analysis.RunAllAnalysisEngine#

class deltakit.decode.analysis.RunAllAnalysisEngine(experiment_name: str, decoder_managers: Iterable[DecoderManager], loop_condition: Callable[[EmpiricalDecodingErrorDistribution], bool] | None = None, data_directory: Path | None = None, num_parallel_processes: int = 16, lvl: int = 0, batch_size: int = 10000000, max_shots: int = 10000000)#

Bases: object

Class to run experiments on different codes and decoders, it will run all of the given sequence of decoder managers.

Parameters:

experiment_name (str) – Name of this experiment.
loop_condition (Callable[EmpiricalDecodingErrorDistribution, bool]) – For a given code and noise model, this loop condition controls how many evaluations of a particular code and noise model are run. It is given as a function that takes in the current empirical decoding error distribution and returns True if another shot should be performed.
data_directory (Optional[Path], optional) – Path to directory in which to store experiment results data. If not given, no data is output to CSV. By default None.
num_parallel_processes (int, optional) – Number of parallel processes to use. If <=1, process will be run sequentially. By default 16.
lvl (int, optional) – Logging level to use, by default logging.NOTSET.
batch_size (int, optional) – The number of decoding shots to run in a batch. Should beset as high as possible.
max_shots (int, optional) – The maximum number of shots to run.

Methods#

`RunAllAnalysisEngine.construct_file_path`	Return the file path to be used for the results data.
`RunAllAnalysisEngine.loop_until_failures`	Loop until there have been at least a specific number of failures, as defined by the decoder manager.
`RunAllAnalysisEngine.loop_until_observable_rse_below_threshold`	Loop until the maximum number of shots or relative standard error on the estimate of LEP go below the given amount.
`RunAllAnalysisEngine.run`	For all codes on all noise sources given to this instance, evaluate the given decoder and return the number of logical failures, non-zero syndromes and stabilisers after several shots in accordance with the given loop condition.
`RunAllAnalysisEngine.save_results`	Save the results to file and log that the file path was used.