Fix goldenspike pipeline and config files

[OliviaLynn]

Problem: The current pipeline files for goldenspike can no longer be run out of the box.

Solution: I'm currently troubleshooting this (and am making final checks on a working version), using both current files and outputs from goldenspike notebook.

Notes: Some problems with the current files:

• missing config entries for entire stages that presumably have been added since its creation
• reading in a pkl when tables_io only wants a pq

Some problems with pipeline file and config created by goldenspike notebook:

• missing at least one field in its output yml (site:name:local)
• refers to data files created only by running the notebook
• creates absolute paths that need to be made relative (and are mixed between goldenspike_examples/ and goldenspike_data/)
• requires a pretrained flow (I placed it goldenspike_data/data/pretrained_flow.pkl), but outputs goldenspike_data/output/goldenspike_data/data/pretrained_flow.pkl (maybe this is desired behavior? either way seems low priority for the moment)

[OliviaLynn]

Note to self to fix goldenspike outputting of pipeline so it can be run on its own in future

[aimalz]

I could make a separate issue for this, but I think it might be solveable along the way to addressing this one -- feel free to separate it out if it's too far off. The readme for the Golden Spike examples should describe the non-notebook version of the goldenspike pipeline and show how to run it at the command line. It would be nice if it also explained the roles of the two .yml files with nearly the same name (and/or renamed one to disambiguate).