feat: add a script to bulk mark users as spam retroactively

[tefkah]

• feat: add bulk spam labeling script
• refactor: write analyze results incrementally

Issue(s) Resolved

Still so many spam users!

This PR adds a script that scans every. single. user in PubPub and determines whether they are likely spam.

It does this by checking for a few metrics, which increase the spam score

• Do all the users comments contain links? +2
• Is the user not associated to any Community (are a member, have attribution) and have a link in their profile? +2
• Do they post comments with links, and don't have any memberships? (this alone qualifies someone as spam) +4
• Do their comments contain common spam phrases ("I like your blog!") and contain a link? +2
• Do they have a URL in their bio? +2
• Have they immediately added their website to their profile? +3
• Do they have a website in their profile, and no memberships? +2
• Does their profile have some spam phrases (BUY PILLS!!) (depends on the phrase how severe)

By default the script considers you spam if you score more than 4.
I think running it with a threshold of 6 should weed out almost everyone, but can be increased.

Test Plan

Screenshots (if applicable)

Optional

Notes/Context/Gotchas

Supporting Docs

[tefkah]

wild that we were still on such an old one