This speeds up scenario instantiation (because picking a bus to use can
be spread out over time) and is a step towards simplifying the spawning
code. Starting downtown goes from 12.8s to 2.2s.
All vehicles spawning at a border now regress to using the 1st valid
lane, instead of random. Now that the choice is made when the trip
starts, this could later be improved to pick the least loaded lane.
Now regenerate everything.
- stretch central polygon a bit to avoid crash when clipping
https://www.openstreetmap.org/way/511767781
- rename polygons ("center" to "huge", and removing the "leeds_" prefix
from the others)
- generate a region overview from the huge map
- only import/match collision data on the huge map
a hard error when they become out-of-date going forward.
Better heuristics make some of these unnecessary. And now the the JSON
files are in this repo, updating files manually when pulling down new
OSM data becomes less tedious.
Previously, dual carriageways (pairs of one-way roads in opposite
directions) mostly didn't get any signal templates successfully applied.
This change ignores outbound-only roads when applying the templates. In
one fell swoop, lots of previously broken signals along places like
Aurora Ave suddenly work reasonably.
in the interior of a big intersection. #255, #114
- No sidewalks or parking on it
- Automatically try to merge it
Bring in fresh Seattle OSM with a few places on Aurora tagged, for
further experimentation.
Also, there's some bug in the importer; Seattle maps didn't get
regenerated last change. Picking up the diffs now.
For the (still disabled) cases of merging short roads, this helps
immensely. It doesn't affect most other maps visibly. Makes a few
already broken things in Krakow and London slightly worse, but don't
care, because they didn't look sane before either.
data/ directory, because they're statically bundled; they're not
actually read from the filesystem. #253
Two SVGs get duplicated between widgetry and abst assets -- dont_walk
and arrow_drop_down.
for all the custom extra import data. #326
Verified there are no changes when importing the affected cities (except
for renaming the original OSM input file for Leeds to match the
geofabrik source)
Woops, and fix updater uploading with the compressed_size_bytes change.
- Tune colors in the experiment
- Regenerate maps again -- I forgot that the parcel data gets joined in
a later importer step.
- Add some of the new maps as levels
degrees to 30 degrees. It works around the issue in #428, but it doesn't
solve the root cause there, so the unit test is also adjusted to provide
a way to solve the harder problem.
Regenerated all maps accordingly. Many traffic signals tended to
improve, with a straight turn marked protected, instead of permitted as
a "right turn."
saves lots of callers from cloning the request and separately plumbing
around the requested start/end distance. Also a step towards exposing
more granular distance crossed in a path for #392.
Still a few more places to simplify, but will do in a separate, smaller
change.
processing code in widgetry doesn't handle paths nested under groups
with transforms. To workaround, preprocess the SVGs:
1) Open in Inkscape
2) Click the problematic group
3) Menu "Path > Object to path"
4) Menu "Extensions > Modify path > Apply transform"
And scale down the two bike sprites
space on maps with "private area around houses".
And a few tweaks to the KML viewer to make it more useful:
- optionally save the clipped file
- click an object to view all attributes in a scrollable popup
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.
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!
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
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!
struct. Whatever choices we make next about naming cities hierarchially
or not can be managed in just one place. #326
This is a pretty huge change, but the compiler gives reasonable
confidence it's correct. More bugs are likely to crop up in the next
step, when filenames start being namespaced by the city too.
Uploading files individually to Dropbox is flaky, and the unique URL per
file complicates data/MANIFEST.
While we're at it, start carving aside a directory structure for keeping
data for the last few releases.
[rebuild] to make sure github actions can download everything
intersection polygon in Krakow that has really bad geometry, and this
improves it. The extra check absolutely shouldn't be necessary, but of
course, all the core line intersection code is quite suspect! #161
Also gets rid of some annoying warnings about roads with missing names.
I could continue to skip the warning for more situations, but I think
this sort of data quality check could be done better in the OSM viewer.
showed this is faster than downloading the uncompressed files, but of
course it probably depends on network speed vs CPU. #326
Also cut over the updater to use gzipped files using flate, instead of
single-file .zip's. [rebuild]
imported, because they referenced way IDs from before the service road
import. That happened after a bad Cargo.lock merge undid the effects of
pinning to the latest seattle_traffic_signals.
only allow the leftmost source lane to turn to any destination lane. As
a future improvment, need to handle multiple explicitly tagged left turn
lanes, but this gets closer to reality, particularly helping some crazy
maneuvers along Mercer in downtown.
*or right
Also had to update lanes along Madison and fiddle a bit to keep
lakeslice running. Spotted some major traffic signal bottlenecks due to
stage generation falling back, will iterate on that separately.
raw string" stuff from the previous commit.
Add tests of a few interesting intersections. The results right now
aren't ideal, but this sets things up for fast iteraton.
lanes all lead to a single lane via left/right turn, and just keep the
left/rightmost lane.
Sanity checked at Rainier / S Massachusetts, and 23rd / S Massachusetts.
- Temporarily workaround snap_cycleways crash in Xi'an
- Fix interpretation of blank turn restrictions. https://www.openstreetmap.org/way/739621435 was missing a right turn, which was causing vehicles to do this crazy loop to go from Madison EB to Lake Wash SB.
- Ignore turn restrictions when they don't match the number of lanes. https://www.openstreetmap.org/way/428090702 and similar need some updating.
Regenerate all data, but give up on lakeslice running fully. Going to
sacrifice that one for a bit to get new roads imported.
We were drawing "pop-up" style buttons in several places, using
copy/pasted logic - which was simple enough when using a unicode glyph.
But we want to use a different glyph which isn't in our font. Further,
using an svg like this gives us finer grained control around the layout
around the button "handle" than we could get with just space characters.
clipping boundary. I don't remember why the more complicated thing came
about. This fixes a weird incoming border in SLU into an unreachable
intersection (which is more accurate) and doesn't seem to cause any
problems to normal or oneshot maps, with or without explicit clips.
it for the prolet robot model. Expose it in the trip info panel.
Total scenario size from 385MB to 412MB, but that's not so bad, and this
seems worth it.
uber-turns (sequences of turns through a few intersections) due to OSM
turn restrictions, we have to be a little careful how we sum up the cost
for the entire sequence, only rounding at the end.
multiple points. This was already handled when the roads went between
the same intersections, but I found a case in Ballard where the roads
are just really close to each other. Screenshot diff empty in Krakow,
but still related to #243
- Cut off the one-way markings before the end of the road, to stop stomping over turn markings
- Draw turn arrows to every road, not each lane
- Only draw turn arrows when a lane is restricted from going to some
outbound lane. At most intersections, all turns are legal, so don't draw
anything.
[rebuild]
data is regenerated. (Ideally screenshots would also be automated, but
that's a little trickier.)
_NOW_ regenerate all data! The only diff anywhere is the binary map
format, so there's confidence the last few commits haven't changed
anything.
the API (#245), and beef up the Python example.
Impact to prebaked file size is tiny -- for lakeslice, the original
intersection_thruput is 2MB and the new traffic_signal_thruput is 435KB.
[rebuild]
a zone per hour. This is part of support for some kind of congestion
charging experiments. This step just rearranges the data to define the
cap and makes a UI to edit it. Not enforcing the cap yet.
one of the outbound roads. This was motivated by one particular example
in downtown, and per screenshot tests, doesn't have regressions
elsewhere. Also improves a few cases in montlake.
- wire up the flag to skip building contraction hierarchies in one-shot
importer. 406s to import london without, 230s by skipping CH
- lazily render zoomed parking lot details. 72s and laggy X11 mouse
before, 42s and no GPU melting after
- add my script for stress-testing the importer
when allowing a car to start a turn. It causes
https://github.com/dabreegster/abstreet/pull/276#discussion_r470269394
and also the lakeslice scenario to gridlock (a regression that began a
few weeks ago). But keep the flag on for now, to keep the montlake
scenaro running at least.
https://dabreegster.github.io/abstreet/trafficsim/gridlock.html has
notes about the many different causes and in-progress fixes for
gridlock. This experiment hasn't been explained very well yet, but
roughly it treats a cluster of traffic signals as one, so that once a
vehicle gains access through the first light, they guarantee immediate
access through the entire sequence. This interacts with the "don't block
the box" behavior (don't start a turn if you might get stuck in the
intersection) strangely.
While attempting to get this rollback to work, I also had to manually
redraw the traffic lights for a few manually specified intersections.
They became out-of-date a few weeks ago when I cleaned up the OSM
geometry upstream and the referenced IDs changed, and I hadn't bothered
to re-time the signals. Luckily, with the new multi-signal editor,
redrawing the timing was much easier than originally!
Regenerated all data and lots of bus routes vanished. Plan to get back
to that project soon.
extend short borders, but it can't be done in convert_osm clip. has to
happen later when trimming and making intersection geometry. too much of
a headache right now.
west seattle and krakow still broken because of short borders.
also grab fresh osm
ignoring parking spots if the parking lane is in the opposite direction.
workaround is to reverse the lane. dealing with that bug next though.
part of #64 and #176. fixes#109
connections to driving/biking. just store the immutable stuff -- whether
there's parking in the building, the connection to the sidewalk, and the
physical driveway line. compute all the rest dynamically, so it responds
to edits without effort.
shouldn't be major behavior changes yet (besides maybe fixing some bugs
involving edits)
parking on tiny roads. fixes some geometry in krakow. #230
(unrelated: add a debug layer to show parking blackholes, to work on
bike connections)
(also unrelated: better error message for #256)
skip unreferenced objects, make the output extremely easy to consume.
also fix timers in importer. this helps with #212, although the actual
problem there is weird.
regen everything. some parking lots change a bit, but not horribly, and
that code needs some attention anyway.
access restrictions. makes old town in krakow look much better and
brings in stay healthy streets in seattle. also commit some code related
to footways that isn't exercised yet.
- just generate every combo of turns from incoming->outgoing
- stop doing all of the weird bike->bike restrictions. that gets handled
anyway in later layers like pathfinding costs, opportunistic LCing,
and picking valid starting lanes
- rely on OSM for filtering out left/right turns from multiple lanes
- why was it useful at all to distinguish Straight from LCing turns?
scrap that
has the effect of eliminating a class of bugs where a driving lane is
sandwiched between lots of bus lanes and wasn't winding up with an
in/outbound turn
gridlock near the huge parking lots downtown from this
and fix a bug with spawning bike trips. never OK to silently not
schedule a trip, because it breaks the schedule continuity
- interpret tags to make 3rd Ave bus-only. cars technically allowed
sometimes, but bus-only is more accurate for now.
- also remove austin from bundled maps. maintaining it has a cost, and
the point was just to have at least one non-seattle city kept up to
date. krakow now satisfies that.
