1
1
mirror of https://github.com/a-b-street/abstreet.git synced 2024-12-29 09:25:15 +03:00
Commit Graph

6575 Commits

Author SHA1 Message Date
Dustin Carlino
96aa09edd0 Small steps towards matching more GMNS signals:
- toss a loading screen around it
- allow jump to ID in edit mode
- break down failure types more
2021-06-13 11:59:37 -07:00
Dustin Carlino
59597708ec Draw polylines as arrows in the KML viewer; it's pretty much always useful to see the direction of the raw data 2021-06-13 11:32:02 -07:00
Dustin Carlino
500f14869f Le Fournil special: don't redraw KML text constantly; it was sluggish for signal timing. 2021-06-13 11:26:08 -07:00
Dustin Carlino
76aba75082 Update links to docs after some reorg in the other repo 2021-06-12 13:38:38 -07:00
Dustin Carlino
43d6ba80c5 Fix all_walking_costs_from by properly accounting for the distance between sidewalk endpoints and building paths. 2021-06-11 10:10:41 -07:00
Trevor Nederlof
4e150dcab0
Generalized connectivity to use Spot and added BorderIsochrone () 2021-06-11 10:06:53 -07:00
Dustin Carlino
972f5b0187 Lower the priority of service roads when placing stop signs
Regenerate all data. Wallingford started gridlocking due to this change;
plowing ahead and will follow up and fix it.
2021-06-11 09:40:44 -07:00
Dustin Carlino
18ff0eefc0 Store Turns per Intersection, instead of a massive BTreeMap.
Minimal file size savings, but substantial runtime improvement!
2021-06-11 09:40:44 -07:00
Dustin Carlino
cabee219db Fix some crashing spinners; the rust compiler inferred the wrong numeric type. 2021-06-10 18:21:24 -07:00
Dustin Carlino
fc4f0a8c8e Some Tempe fixes:
- Stop importing rail in Tempe. Not simulating anything on it yet, and it complicates gridlock. 
- Update the GMNS timing.csv import based on slight format change. 
2021-06-09 16:35:46 -07:00
Dustin Carlino
d827673ace Periodically free memory in the middle of screenshot test, to avoid
running out on my current machine. Fixes .

Finally regenerate screenshots for the first time in ages... just
blindly accepting everything, because the slightly different screen size
means everything was slightly shifted down.
2021-06-09 14:31:40 -07:00
Dustin Carlino
60352747bf Get importer and game crates to share dependencies again, by enabling
some features for transitive deps in each crate.
2021-06-09 10:21:25 -07:00
Dustin Carlino
889c500eb1 Force geojson to use geo-types feature everywhere. This lets different
binaries properly share workspace dependencies.

See
https://www.reddit.com/r/rust/comments/nvd6y7/common_crates_in_cargo_workspace_recompiled_due/
for the full story behind this mess.

I'm going to keep repeating this process until I get game and importer
to share dependencies again.
2021-06-09 09:48:53 -07:00
Dustin Carlino
ecce96fd75 Grab new Tempe OSM data after lots of fixes upstream to 13th/Mill. 2021-06-07 19:10:02 -07:00
Dustin Carlino
f39a50a7e5 new release 2021-06-06 15:45:40 -07:00
Dustin Carlino
710e9152cd Just trigger [rebuild] [release] 2021-06-06 14:32:59 -07:00
Dustin Carlino
5192847e62 Dramatically speed up initial seeding for infinite parking. 2021-06-04 11:37:47 -07:00
Dustin Carlino
a38125ad2c Adjust multi road editor more -- custom cursor icons for select/unselect 2021-06-03 15:23:55 -07:00
Dustin Carlino
8721c4f6d9 When the desktop version is old, have a button to go directly to the releases page to grab a new one 2021-06-03 09:14:17 -07:00
Dustin Carlino
bf2e4be2a2 Adjust road editor based on real live coffeeshop feedback:
- "x/y similar segments" to relate to explanatory text about lane
  configurations below
- "edit multiple" -> "apply to multiple", since you have to change the
  road first
- start with all candidates selected
2021-06-03 09:02:12 -07:00
Dustin Carlino
b1150c9524 Export raw CSV data about problems per trip, for a data analyst to explore possible dataviz in their more familiar tools 2021-06-02 19:23:48 -07:00
Dustin Carlino
c9e4f754e1 Fix a crash in the problem map layer, caused by checkboxes getting out of sync 2021-06-02 16:26:04 -07:00
Dustin Carlino
ef4ea01c71 Add a button to import all traffic signals map-wide from a timing.csv file, reporting the total number of successes and failures. 2021-06-02 12:18:24 -07:00
Dustin Carlino
d2ecab44a7 Be sure to validate signals before committing the results, since GMNS imports aren't validated upfront. 2021-06-02 11:40:54 -07:00
Dustin Carlino
cd66d7209e Automatically fill out crosswalks to the GMNS signal import. 2021-06-02 11:13:30 -07:00
Dustin Carlino
5127a82b54 Import Tehran 2021-06-01 10:19:06 -07:00
Dustin Carlino
d014098f59 Fix an extremely naive implementation of calculating average angle,
which was causing bizarre turn lane arrows in Tempe.
2021-06-01 09:52:40 -07:00
Dustin Carlino
00c3614301 new release 2021-05-31 12:12:13 -07:00
Dustin Carlino
864155f998 Try again, fix github builder [rebuild] [release] 2021-05-31 11:01:43 -07:00
Dustin Carlino
3bd219fd20 Just triggering [rebuild] [release] a whole day late 2021-05-31 10:57:45 -07:00
Dustin Carlino
dc1ce15c30 Heavily penalize GMNS snapping if the bearing doesn't match the movement.
University/Mill looks great now.
2021-05-28 15:56:04 -07:00
Dustin Carlino
30a317cf94 Make the KML viewer be able to inspect multiple overlapping objects. Useful for GMNS timing.csv files. 2021-05-28 15:21:07 -07:00
Dustin Carlino
4227371bbe Rewrite the GMNS signal movement snapping algorithm. Instead of
independently picking the best match for "from" and "to", it considers
all vehicle movements, and minimizes "turn_type_cost * (from_cost +
to_cost)", using a hint about turn type to correct for geometry
mismatches at consolidated intersections. 

Gets better results in Tempe at University and Mill, at least.
2021-05-28 14:55:53 -07:00
Dustin Carlino
5aa00955f9 Clean up the timer.parallelize API, getting rid of the weird enum to
toggle how many CPUs to thrash. We always use all of them, except for
one case, where a separately named method is more clear. Also make that
variation use all but 1 CPU, instead of just half.
2021-05-27 17:47:22 -07:00
Dustin Carlino
cf6b2f6db7 Rewrite the heuristics for stop signs. Similar to traffic signals, only
count incoming roads when figuring out if an intersection is degenerate.
Also make link roads (on/off ramps) lower priority than the main part of
the road.

Regenerated everything.

(and fixing up the cloud scripts)
2021-05-27 15:37:46 -07:00
Dustin Carlino
8ceaf296eb Allow u-turns at merged intersections when the turn lane is explicitly tagged in OSM. This makes the infamous U-turn from Montlake to 520 WB work! The inferred signal timing isn't great, but ah well.
Not regenerating yet
2021-05-27 10:09:22 -07:00
Dustin Carlino
83bc768e28 Optimize costs of importing in the cloud.
Every time I run an import, 10 GCE workers download about 20GB of
data/input. The S3 outbound charges are uncomfortably high.

Instead, use GCP's S3->GCS transfer tool manually before each run, and
make the GCE VMs read from that instead.

I haven't tested these changes yet, but will soon with the next import.
2021-05-27 08:09:03 -07:00
Dustin Carlino
53430319b1 Filter out trips between the same place. Regenerate all Britain background scenarios. Fixes 2021-05-27 07:57:25 -07:00
Dustin Carlino
455399b90e Collapse degenerate intersections between cycle+footpaths. At various
times in the past, I've also tried doing this for other roads, but wound
up reverting for reasons only git remembers. But it seems like an
obvious win for bike paths; especially around Seattle, the ways are
split because of all of this raised curb data we're ignoring anyway.

Tested manually around Montlake.

Finally regenerating all data... Only Phinney breaks. One for tomorrow.
2021-05-26 22:27:29 -07:00
Dustin Carlino
3a68e5677f Allow all vehicles to use bus-only lanes that also happen to be turn lanes.
There are some caveats with the implementation commented inside.
Verified manually in Montlake -- down to just 122 cancelled turns!

Still not regenerating...
2021-05-26 19:13:14 -07:00
Dustin Carlino
6f0e01c511 Fix a critical bug with pathfinding v2: when calculating allowed
movements, make sure that a vehicle can access BOTH the source and
destination lane of a turn. 

Without this fix, a turn from a bus lane to a driving lane is allowed
for cars at the pathfinding v2 level, but then it fails when converting
into v1 lanes.

(Next step: actually allow this for combination bus/turn lanes)

Not regenerating data yet
2021-05-26 15:39:34 -07:00
Dustin Carlino
e4a95b604e Fix Rainier Valley gridlock, fallout from .
https://www.openstreetmap.org/changeset/105381427 mapping a turn lane is
the fix, preventing crazy U-turns using the service road. Not sure how I
missed that lane when looking here before. I made the edit to .osm
locally instead of grabbing fresh data for all of Seattle.
2021-05-26 15:02:20 -07:00
Dustin Carlino
9fc79b1089 Use explicit start/endpoints for trips in scenarios. This is necessary
for people that leave one border, then re-enter a different one. 

Alternative considered: insert a dummy remote trip between the two
borders. We used to do something like this at non-trivial code
complexity expense and having to explain the trip in the UI.

Regenerated all scenarios and prebaked data.

- Modest file size increase, as expected. Montlake scenario goes from
  1.3MB to 1.5, downtown from 37MB to 43MB, all Seattle scenarios from
  280MB to 327MB
- Eyeballed a few of the previously broken trips; they work now!
- Montlake goes from 3127 cancelled trips to just 302!
- No new gridlock, except in Rainier Valley -- disabling that for now
- I discovered missing validation in Poundbury for no-op trips between
  the same building. I'll filter those out and restore prebaked results
  there in a followup.
2021-05-26 14:30:01 -07:00
Dustin Carlino
3b375078b1 Prebake data until the simulation is done, not just until midnight. That
way, we have trip stats for people starting near the end of the day, and
we stop incorrectly showing failed trips when comparing.

Prebaked data: no change in size (245MB)
Montlake: 3135 "cancelled" trips -> 3127
Lakeslice: 6766 "cancelled" trips -> 6647
2021-05-26 10:22:09 -07:00
Dustin Carlino
4bd79cabbf Upload regenerated map data and address turn restriction PR feedback 2021-05-25 20:24:11 -07:00
Dustin Carlino
2dd611fd73 Refine the connectivity check for bus and bike lanes
Down to 12 connectivity problems in Montlake! Now that the dust has
settled, manually check the 3 unit tests and confirm the goldenfile
diffs.
2021-05-25 20:24:11 -07:00
Dustin Carlino
7af86a9e59 Interpreting turn:lanes tagging more carefully.
And fix an absolutely infurriatingly bad typo from the previous commit.

Now both the multiple_left_turn_lanes and divided_highway_split test
maps look great!
2021-05-25 20:24:11 -07:00
Dustin Carlino
c95a0b0f64 Handle another case for creating multiple left/right turns at an intersection. Use the new validation for connectivity at an intersection and give up on using OSM turn restrictions in some cases.
39 problems in Montlake down to 33, but regressions on some of the LCing unit tests...
2021-05-25 20:24:11 -07:00
Dustin Carlino
62af2fef8c Introduce a new method to check for connectivity problems between vehicle lanes at an intersection, and start tracking the number of problems across a map. 2021-05-25 20:24:11 -07:00
Dustin Carlino
f586ca262d Upload new prebaked data after . This diff shows the file size
increase
2021-05-25 16:09:35 -07:00