[Symphony]: Symphony logic final (final) version

[lvachon1]

Scope

Asana Ticket: 🐬 Symphony logic final (final) version

Implementation

Changed the business logic for stale? alerts, now they're only for the symphony work. Updated alerts_test tests to suit the new logic.

Screenshots

Symphony alert hidden on Subway Status

Symphony alert hidden on Green-E line

Symphony alert shown in Green-E alerts

Symphony alert hidden in Green line

Symphony alert shown in Green alerts

How to test

http://localhost:4001/ & http://localhost:4001/alerts/subway - Confirm the symphony alert isn't shown in Subway Status
http://localhost:4001/schedules/Green-E/line & http://localhost:4001/schedules/Green/line - Confirm the alert doesn't show on the line diagram but does on the Alerts tab

Depending on the order in which these PRs are merged you may want to test the departures page as well.

http://localhost:4001/departures?route_id=Green&direction_id=1&stop_id=place-symcl - Confirm the alert shows
http://localhost:4001/departures?route_id=Green&direction_id=1&stop_id=place-hymnl - Confirm the alert doesn't show

[joshlarson]

Looks good! I appreciate that this changes the definition of stale? without changing anything about its interaction with the outside world, which means that it's fully, 100% independent of your other stale-related PR 🎉

---

I feel compelled to flag that technically, this would match an alert that was created to close Symphony and another stop, so if, say, a single alert closed both Symphony and Prudential, then that alert would also get marked as stale, and wouldn't show up.

I still consider this non-blocking, both because I think the scenario above is exceedingly unlikely, and also surprisingly difficult to solve for (since individual platforms are typically informed entities during station closures as well, so it's not as simple as saying "are all of the informed stop entities place-symcl?"), but I also think it is worth noting... data can do surprising things!