Parallel Processing in Sidle

Sidle is built to be run in parallel using dask. This system allows for flexible implementations. All sidle commands except reconstruct taxonomy allow parallel processing. There are three ways to manage parallel processing:

• --debug which turns off all parallel processing. This is primarily implemented for testing
• --p-n-workers will create a new dask cluster with the specified number of workers and will use the avalaible resources
• --p-client-address allows you to pass in a pre-configured cluster for processing. You can learn more about setting up dask clients in their documentation.

This hopefully provides a flexible interface for parallel procesisng of your samples