fix: always show route_pattern_prefix_overrides trips

[nlwstein]

Summary of changes

Asana Ticket: [extra] 🐛🍎 API: Ashland stop disappeared on the Worcester Line

I lost my mind debugging this, but we were missing a case where a valid "canonical override" trip was falsely being flagged as non-canonical because it ended up in a list that was not checked against the configuration value.

This logic (and perhaps the approach overall) for assigning stops to routes seems like it may be overcomplicated for what we are trying to do. It may even make sense to have a hard coded source of truth that we conditionally override vs. reverse engineering stops from trips in the future.

Asana Ticket: 🍎 Update merge_ids logic for StopsOnRoute

For this (Boat-F6), I am suggesting that we just whack the hardcoded configuration - from my understanding of the code, the output we are currently seeing is correct. For reference, locally I am seeing:

NOT SPECIFIED: Boat-Aquarium, Boat-Fan, Boat-Quincy, Boat-Winthrop, Boat-Logan

INBOUND: Boat-Winthrop, Boat-Quincy, Boat-Logan, Boat-Fan, Boat-Aquarium

OUTBOUND: Boat-Aquarium, Boat-Fan, Boat-Quincy, Boat-Winthrop

According to the PDF schedule and my reading of the static schedule in GTFS, that is correct: https://cdn.mbta.com/sites/default/files/media/route_pdfs/2023-10-02-winthrop-ferry-weekday-schedule-effective-through-november-30-2023.pdf

I have removed the overrides as they state Logan is in both directions, which conflicts with this paper schedule.

Some extra context can be found here, though honestly stepping through this code with a debugger gave me a better understanding than anything else 💫

[github-actions]

Coverage of commit 8933b61

Summary coverage rate:
  lines......: 89.2% (4101 of 4596 lines)
  functions..: 70.9% (2224 of 3138 functions)
  branches...: no data found

Files changed coverage rate:
                                                                        |Lines       |Functions  |Branches    
  Filename                                                              |Rate     Num|Rate    Num|Rate     Num
  ============================================================================================================
  apps/state/lib/state/helpers.ex                                       |93.8%     16| 100%     3|    -      0
  apps/state/lib/state/stops_on_route.ex                                |94.9%     99|96.7%    30|    -      0

Download coverage report

[nlwstein]

@bfauble I updated this branch to not remove the override to avoid the unordered Logan. I suggest that we close the investigation ticket and either make a follow up ticket to support overriding global stop order (IE: when pulling stops without a direction_Id specified) or leave as-is and hopefully refactor the entire approach at some point.

[github-actions]

Coverage of commit 289aef0

Summary coverage rate:
  lines......: 89.2% (4104 of 4600 lines)
  functions..: 70.9% (2225 of 3139 functions)
  branches...: no data found

Files changed coverage rate:
                                                                        |Lines       |Functions  |Branches    
  Filename                                                              |Rate     Num|Rate    Num|Rate     Num
  ============================================================================================================
  apps/state/lib/state/helpers.ex                                       |93.8%     16| 100%     3|    -      0

Download coverage report