Skip to main content

Batch Processing

Bulk processing of large datasets is a built-in feature in MPS CLI.

Overview​

Instead of submitting VRS files one at a time, you can point the MPS CLI to a folder (or multiple folders) containing multiple VRS files. The CLI will automatically:

  • Discover all VRS files in the specified folder(s) recursively
  • Concurrenty process n (default 25) recordings at a time and queue the rest.

Usage​

Process all VRS files in a folder​

aria_mps single -i <path_to_folder>

The CLI will recursively find all VRS files in the specified folder and submit them for processing.

Process VRS files from multiple folders​

You can chain multiple input folders using the -i flag:

aria_mps single -i <path_to_folder_1> -i <path_to_folder_2> -i <path_to_folder_3>

Examples​

Single folder​

aria_mps single -i ~/recordings/session_01/

Multiple folders​

aria_mps single -i ~/recordings/day1/ -i ~/recordings/day2/ -i ~/recordings/day3/

Batch Concurrency​

The batch processing feature automatically manages concurrent jobs to stay within the service limits:

  • By default, a maximum of 25 concurrent processing jobs are allowed. This concurrency count can be updated using the concurrent_processing setting in CLI Settings. The max allowed concurrency is 100 at this time.
  • Jobs are queued and submitted automatically as capacity becomes available

This ensures reliable processing without overloading the system.

Output​

MPS outputs are saved next to each original VRS file, following the same folder structure as single file processing. See Getting Started for details on output folder structure.