generate_replay_data

src.lookoutequipment.dataset.generate_replay_data(dataset_name, replay_start_timestamp, upload_frequency, replay_days=1, inference_timezone='UTC')

Generates inference input data from the training data to test a scheduler that would be configured for a model trained with this dataset. The data will be output in an S3 location next to your training data S3 location.

Parameters
  • dataset_name (string) – Lookout for Equipment dataset_name containing the training data for replaying.

  • replay_start_date (string) – Point in time in the training data from which to begin generating replay data. Example: “2020-10-01 00:00:00”

  • upload_frequency (string) – How often replay data is uploaded to the S3 bucket for the inference input data. Valid Values are PT5M, PT10M, PT15M, PT30M, or PT1H.

  • replay_days (integer) – Duration of the replay data in days (default: 1)

  • inference_timezone (string) – Indicates the timezone for the inference replay dataset. (default: ‘UTC’)

Returns

(boolean)

True if no problem detected, otherwise a list of sequences that could not be generated (which will trigger a failed scheduler execution)