Automate additional data sources using GHA

[e-belfer]

Overview

Partially addresses #276

What problem does this address?
Adds all archivers currently running out of the box to GHA, and sorts file_changes to improve the legibility of reported changes.

What did you change in this PR?

• Adding remaining data sources to the run-archiver method to prevent stale archives, and adding some quality of life improvements to the summary.
• Ran mshamines locally with refresh-metadata flag to update keywords that were causing an error, but did not resolve previous_version error so removed this archive from the list and moved into Fix broken archivers #285.
• Manually approved new archives for mshamines and epacamd_eia to fix previous_version errors.
• Updated mshamines partition name to form from dataset to avoid downstream errors in extraction

Out of scope:

• Any additional required debugging is catalogued in Fix broken archivers #285

Testing

How did you make sure this worked? How can a reviewer verify this?
Run run-archiver in GHA. All datasets should pass.

To-do list

Tasks

Beta Give feedback

1. Review the PR yourself and call out any questions or issues you have

[zaneselvans]

Do we still expect the FERC XBRL archives will always appear completely new, even if nothing has changed in their contents?

IIRC they were doing something like autogenerating new IDs for every post in the RSS feed every time the feed was read, rather than using persistent unique IDs for each post. Did we find some way around that?

[e-belfer]

FERC archivers are certainly not working at present, but I'm tracking this in #285. Maybe this is a known and intended failure and I'm missing something, in which case these archivers shouldn't be candidates for automation.

[zaneselvans]

Hmm, #285 looks like real brokenness in the FERC archivers.

If we haven't addressed the unique ID thing, we'd just see that all the FERC XBRL archives get updated every time the archiver is run, but I think we'd still be able to get an idea of how much new data there is from the change in the size of the archives, and saving interim data doesn't seem like a bad idea given how flaky FERC's data curation is!