✨ Modify and/or bundle data as downsampled/truncated for pipeline testing

[sgiavasis]

Related problem
Pipelines can be very long, especially with many fork points. It would be good to allow users and developers to test the connection AND execution of a pipeline config very quickly.

Proposed feature
The pipeline can finish processing downsampled and truncated BOLD data very rapidly. It would not allow for testing quality or correlations of output analyses, but instead of infrastructure, and to make sure the pipeline reaches the end. Either or both:

• A feature that modifies a single participant of the user's data, where it truncates and downsamples the data appropriately, but not so much that it causes crashes (too few TRs for certain tools, or too few TRs for GLMs to complete, etc.).
• Bundle some sample public datasets that have been downsampled and truncated, for easy one-click test runs.

[shnizzedy]

This feature would go a long way toward #1455 (comment)