a-b-street/abstreet
1
1
Fork 0
mirror of https://github.com/a-b-street/abstreet.git synced 2024-12-27 08:24:15 +03:00
Code Issues Projects Releases Wiki Activity
5,492 Commits 85 Branches 166 Tags 395 MiB
d799b20d93
Commit Graph

1340 Commits

Author SHA1 Message Date
Dustin Carlino
544dda9c5c When looking for turn conflicts, treat reserved turns (created before 
starting an uber-turn) the same as accepted turns, updating the
blocked-by graph and allowing a movement to start if a cycle is
detected.

No effect in montlake or lakeslice, and very slight progress at the
Krakow roundabout for #382 -- increasing a tiny 26 trips by 7:20 am.
 2020-12-03 10:56:42 -08:00
Dustin Carlino
5dba8f5896 Make the blocked-by graph used for debugging and the API detect problems 
from somebody about to start an uber-turn.
 2020-12-03 10:21:43 -08:00
Dustin Carlino
69b14e444f Simpler abstutil APIs: return string errors immediately, don't make callers transform. We don't need to distinguish different error types anywhere. 2020-12-02 08:44:07 -08:00
Michael Kirk
ddfaf73eae Improve lane changing vis-a-vis uber turns. 
1. Allow lane changing in an uber turn. Because of the way uber turns
   work, we lock in and commit to all the lane changes just before
   entering the uber turn.
2. Avoid overzealous lane changing by combining number-of-lanes-crossed
   and numer-of-vehicles-in-lane into a single cost, rather than always
   preferring the least  number-of-vehicles-in-lane.
3. Don't lane-change unless the candidate lane's cost is strictly better
   than the current lane cost.
 2020-12-01 19:20:53 -08:00
Dustin Carlino
30ce82bd3c Collapse get_pathfinding_request from TripSpec into start_trip. Avoids double-calculating paths (for driving trips, because of capping) and consolidates logic a bit more. #258 2020-12-01 16:42:36 -08:00
Dustin Carlino
14fd697282 Collapse/consolidate some logic involving DrivingGoal and TripEndpoint. #258 2020-12-01 16:04:48 -08:00
Dustin Carlino
d0b25d3530 Slightly refactor spawning cars during trip transitions #258 2020-12-01 15:30:17 -08:00
Dustin Carlino
4f764d6b87 Plumb options for delaying capped trips, instead of cancelling. Don't actually handle it yet. #325 2020-12-01 14:37:55 -08:00
Dustin Carlino
57963126fa Fold number of housing units into Building; ditch the extra file approach. 2020-12-01 12:43:19 -08:00
Dustin Carlino
2ebaf99e0d A fresh attempt at finding the root causes of gridlock map-wide 2020-12-01 10:50:05 -08:00
Dustin Carlino
b5ea263fc1 Expose the blocked-by graph through the API, and add root-cause analysis 
to the UI debugger. #392
 2020-11-30 17:01:15 -08:00
Dustin Carlino
8501ec70f9 Make the blocked-by graph more exploreable. Remove older debug code that 
did something similar. #392
 2020-11-30 15:15:13 -08:00
Dustin Carlino
6dd14c8bed Don't crash when jumping to nonexistent trip or when clicking 
follow/unfollow for a trip without an agent.
 2020-11-30 14:20:37 -08:00
Dustin Carlino
3eda17627b Handle binary or JSON scenarios 2020-11-30 14:00:19 -08:00
Dustin Carlino
72e54c68d8 Fix a bug with TripManager's spawn_ped -- it wasn't actually cancelling trips when pedestrians don't have a path for the second leg of their trip. #258 2020-11-25 10:34:53 -08:00
Dustin Carlino
903eab1bc4 Remove some old savestating optimizations from the simulation. The 
optimizations avoided serializing Analytics and paths of to-be-created
agents, to reduce the file size. The logic to manage all of this isn't
worth the complexity anymore, because:

1) We don't queue up a bunch of spawn commands anymore; we defer
pathfinding until the last minute anyway.

2) We're not using savestates except for occasional manual debugging.
Previously, there was an idea to quickly preview prebaked traffic
mid-day. That idea was never fleshed out.
 2020-11-24 15:38:20 -08:00
Dustin Carlino
ab88010ed0
Refactor PathRequest creators that go between two buildings. #176 (#404) 2020-11-24 13:21:44 -08:00
Dustin Carlino
0dd246c83f Remove the old --num_agents flag, used for debugging spawning agents 2020-11-22 18:27:25 -08:00
Dustin Carlino
47d09d61f0 Plumb the intended TripMode for cancelled trips; instead make the trip duration optional. Breaking change for anyone using the API. Makes way more sense. [rebuild] [release] 2020-11-22 11:21:44 -08:00
Dustin Carlino
9b018e42a1 Make the blocked-by graph also understand target queue being full. Still 
not handling all cases yet. #392
 2020-11-20 14:48:56 -08:00
Dustin Carlino
e43f551af6 New query and debug viewer to get the full graph of agents waiting on others. #392 2020-11-20 14:19:19 -08:00
Dustin Carlino
4690ea36cd Reduce visibility of some sim crate stuff. #258 2020-11-19 14:53:27 -08:00
Dustin Carlino
3905131fe4 Track and expose total trip distance in the API. #392 2020-11-19 10:40:00 -08:00
Dustin Carlino
cb4b84873b Refactor code into person_finished_trip for all trip transitions. #258 2020-11-18 19:05:44 -08:00
Dustin Carlino
37ae02352e Just reorganize TripManager code a bit. #258 2020-11-18 18:30:03 -08:00
Dustin Carlino
e6e7ec42ff Collapse two methods in TripSpawner. While we validate a TripSpec, also produce the TripLegs. Also get rid of the stateful TripSpawner entirely. #258 2020-11-18 15:08:06 -08:00
Dustin Carlino
915e12ff7e Remove PersonID from Scenario; the simulation will assign IDs sequentially. They're redundant. #258 
Slightly reduces scenario file size, and simpler code to construct
scenarios.
 2020-11-18 14:13:25 -08:00
Dustin Carlino
4ab98b1477 Compact the schedule in Scenario by just storing each TripEndpoint. #258 
The schedule validation changes slightly. No-op trips between the same
origin/destination are now an error and get filtered out.

huge_seattle scenario goes from 129MB to 110MB with the redundant
endpoints removed.
 2020-11-18 11:13:09 -08:00
Dustin Carlino
739e33ec3e We do need to fill out trip legs for TripSpec::SpawningFailure; smoke tests in other maps hit it. #258 2020-11-17 18:02:48 -08:00
Dustin Carlino
82c7cd1e8f At last, collapse the OD -> SpawnTrip -> TripSpec function into one thing. #258 2020-11-17 17:55:29 -08:00
Dustin Carlino
7741ea79b5 Refactor starting trips -- caller passes in TripInfo. Stop inferring 
TripEndpoints from TripLegs. #258

NOW regenerate scenarios. I'm confident this sweeping refactor didn't
break behavior, because prebaked data didn't budge. huge_seattle
scenario went from 147MB to 129MB. Not bad!
 2020-11-17 16:59:37 -08:00
Dustin Carlino
8d80668a56 The big cutover: a Scenario's trips are now just start/endpoints and mode, not the old SpawnTrip structure with a bunch of different cases and details picked out. #258 
Not regenerating scenarios yet...
 2020-11-17 16:28:52 -08:00
Dustin Carlino
8d28274736 Simplify the old ScenarioGenerator, mostly by removing OriginDestination 
in favor of Option<TripEndpoint>. The bike/car contention tutorial stage
has to be tweaked manually, since there's no longer a nice way to spawn
vehicles at a non-border intersection and force them to use a certain
DirectedRoad. #258
 2020-11-17 15:51:30 -08:00
Dustin Carlino
0599d4b47b Whittle down places constructing SpawnTrip manually. #258 2020-11-17 14:10:47 -08:00
Dustin Carlino
7bb1d5b59c Switch the UI's spawn_agents_around intersection to go through Scenarios instead of reaching into internal sim spawning gunk. Lock down visibility of some sim spawning things. #258 2020-11-17 11:13:42 -08:00
Dustin Carlino
2abeebc55e Add a third TripEndpoint case to appear at an exact Position (lane + distance). This is a necessary step before simplifying Scenario's IndividTrip to use TripEndpoints and mode instead of SpawnTrip. #258 2020-11-17 10:50:42 -08:00
Dustin Carlino
5ea3c2ba6f Remove support for remote trips, and stop plumbing off-map locations. A 
remote trip goes between two locations off-map, specified just by a GPS
coordinate. The trips aren't simulated at all. They were originally
added to support Orestis's pandemic model, to handle transmission
off-map in shared buildings. This work has died off, there are no other
anticipated use cases for remote trips, and they complicate bigger
refactorings. #258

This also has the nice side effect of substantially reducing scenario
size -- huge_seattle from 177MB to 147MB. That unused metadata was
expensive!
 2020-11-17 10:14:23 -08:00
Dustin Carlino
32dea68f5d Measure how many simulation events actually result in something 
changing. #368
 2020-11-16 11:51:24 -08:00
Dustin Carlino
3433fd99e0 Smarter traffic recorder. Capture the exit point when the vehicle crosses the entry point. Otherwise, in the Krakow scenario of interest, the vehicles piling up and making poor LCing decisions actually get blocked by other gridlock and never make it to the exit. #382 2020-11-16 09:59:43 -08:00
Dustin Carlino
754749a578 Parallelize part of scenario spawning. One of the code paths calls 
biking_connection, which does a slow graph floodfill.

Prolet robot in Krakow: from 22.6s to 10.1s

Because Krakow now uses separate sidewalks and they're not properly
connected to the road graph yet, biking_connection has to search very
far. The speedup isn't noticeable in other cases.
 2020-11-16 09:36:48 -08:00
Dustin Carlino
b9ff94f275 Fix crash in krakow with a biking trip that should get converted to a walking trip, but isn't 2020-11-15 19:59:31 -08:00
Dustin Carlino
24e1a48d75 Generalize TripSpec::NoRoomToSpawn to handle trips that're impossible to instantiate. Hitting this with 'random traffic' in krakow. #258 2020-11-15 17:52:54 -08:00
Dustin Carlino
d3aa81d2d2 new tool to record a scenario consisting of all traffic around some intersections, to help with producing a better test case for #382 2020-11-15 16:57:33 -08:00
Dustin Carlino
22cef74aec Always allow blocking a degenerate intersection. Visually inspected on 
the lane-changing test map. #382

Also abandon sim.run_until_done; the only caller was the LC test, and it
uses 30s steps inappropriately anyway.
 2020-11-12 11:34:17 -08:00
Dustin Carlino
62135ffc26 Allow blocking the intersection at another Green lake case, and trigger a [rebuild] 2020-11-12 10:37:21 -08:00
Dustin Carlino
3583d068a6 Fix a bug introduced in e3b9c42506 on the proposal saver, and allow blocking-the-box on another Ballard intersection 2020-11-11 13:20:56 -08:00
Dustin Carlino
e5ff76e00c Working on gridlock in the Ballard map. Allow blocking-the-box near the 
famous Green Lake intersection, and add a little debug tool to see
current queue reservation length.
 2020-11-11 12:43:28 -08:00
Dustin Carlino
07ac288e72 Refactor: rename Pt2D::from_gps to just gps.to_pt(bounds) 2020-11-10 12:00:43 -08:00
Dustin Carlino
e15efa9002 Organize raw maps by city. With this, data/input/ is totally namespaced by city! Nothing prevents us from having a map named "downtown" in multiple cities now. #326 2020-11-05 14:58:28 -08:00
Dustin Carlino
cda586dfc9 Cleanup by moving some methods to belong to MapName. #326 2020-11-05 14:18:13 -08:00