Create instructions for how to initiate a pipeline run programmatically

We need to document how to programmatically run pipelines. The distinction between just running the pipeline vs running + automatically saving/registering an asset in the aind-open-data bucket should be made clear. The latter requires some special credentials and somehow involves the automation capsule?

All of this is very mysterious to me and I'd like a reference for myself and for others on how to do this (in a way that doesn't require going to code ocean and clicking buttons!)