Commit Graph

123 Commits

Author SHA1 Message Date
Dustin Carlino
638e93c36a Stop re-exporting stuff from synthpop in sim 2022-01-28 13:13:43 +00:00
Dustin Carlino
9f86365cf6 Upgrade to rust 1.56 and cutover to rust edition 2021 2021-10-21 09:38:06 -07:00
Dustin Carlino
1d7149c18b Don't crash in the activity model generators if a map has no borders. This currently happens with Overpass imports. #717 2021-07-25 10:30:06 -07:00
Rémi Lauzier
1a4b997818 Fix some clippy warnings 2021-06-16 16:21:35 -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
9fc79b1089 Use explicit start/endpoints for trips in scenarios. This is necessary
for people that leave one border, then re-enter a different one. #664

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
Vinzent Steinberg
a1899e29f8
Update dependencies (#656)
* Update dependencies

* Use github for polylabel, to make all dependencies use geo 0.18, not a mix of 0.18 and 0.17

* Downgrade lyon to avoid a crash

Co-authored-by: Dustin Carlino <dabreegster@gmail.com>
2021-05-22 16:23:06 -07:00
Vinzent Steinberg
9d17faa878 Prefer slices over reference to owned type 2021-05-15 14:24:08 -07:00
Dustin Carlino
1edbca6509
Refactor most of the places extracting polygons from geojson files (#644) 2021-05-15 09:05:28 -07:00
Vinzent Steinberg
fd3b0e2a14
Fix compilation failures and most clippy warnings (#642)
* abstutil: Fix compilation failure

* map_gui: Fix compilation

* traffic_signal_data: Fix compilation failure

* map_model: Fix compilation failure

* abstutil: Fix doctests

* abstio: Fix most clippy warnings

* abstutil: Fix most clippy warnings

* collisions: Fix clippy warning

* convert_osm: Fix clippy warnings

* sim: Fix most clippy warnings

* geom: Fix clippy warnings

* kml: Fix clippy warnings

* map_model: Fix most clippy warnings

* fifteen_min: Fix clippy warnings

* game: Fix many clippy warnings

* Disable some noisy clippy warnings

* headless: Fix clippy warnings

* importer: Fix clippy warnings

* map_editor: Fix clippy warnings

* map_gui: Fix clippy warnings

* osm_viewer: Fix clippy warnings

* parking_mapper: Fix most clippy warnings

* popdat: Fix clippy warnings

* santa: Fix clippy warnings

* sumo: Fix clippy warnings

* traffic_seitan: Fix clippy warning

* updater: Fix clippy warnings

* widgetry: Fix clippy warnings

* tests: Fix some clippy warnings

* Fix compilation on stable Rust

* Simplify unwrapping

* Make use of `Entry` more readable

* Fix formatting

* Fix code that was broken in the refactoring

* Apply cargo +stable fmt

* Fix code that was broken in the refactoring, second try

* Remove `Default` impls that are equivalent to `new`

* Remove obsolete clippy wrapper

* Avoid turbofish

* Prefer `unwrap_or_else` over allowing `clippy::or_fun_call`

* Remove redundant `into_iter`

* Fix typo

* Prefer `&& false` over commenting code out

* Fix some clippy warnings

Co-authored-by: Dustin Carlino <dabreegster@gmail.com>
2021-05-14 08:32:56 -07:00
Dustin Carlino
0357ae0a23 Tidy up the new UK census gen code. #556 2021-03-05 16:51:02 -08:00
Dustin Carlino
8ffd249558 Get the new UK scenario generation working everywhere. #556
If a study area exists for the map, make a copy of the base/active
scenarios with the background traffic mixed in. Also remove people
living in the site, since they're redundant.

Ran it like this: for city in `ls data/system/gb/`; do ./import.sh
--scenario --city=gb/$city || break; done
2021-03-05 16:26:29 -08:00
Dustin Carlino
14ccddf15e Create people that live or work (or both) off-map. #556 2021-03-05 15:50:56 -08:00
Dustin Carlino
9610b69de4 Remove zones without homes or shops. Found one in London. This can
happen when the overlap with the map is tiny. #556
2021-03-05 14:53:29 -08:00
Dustin Carlino
d2b25f9340 Use choose_weighted to sample homes/workplaces proportionally, instead of worrying about how many employees per store there might be. #556 2021-03-05 12:33:04 -08:00
Dustin Carlino
89f317c79d Add in normal distributions for trip departure times. #556 2021-03-05 12:24:17 -08:00
Dustin Carlino
c7042eda06 Create simple people for UK scenarios. #556
Can't quite debug yet, because for Poundbury, the matching zones (like
E02004277) have no entries in the census data...
2021-03-05 11:59:20 -08:00
Dustin Carlino
1f33916502 Match houses/workplaces to zones. #556 2021-03-05 11:27:25 -08:00
Dustin Carlino
aa59677ad2 Transform the UK OD and zone data from raw input into a clean structure. #556 2021-03-05 10:47:48 -08:00
Michael Kirk
4b665c871d
update geo related deps. (#508)
geo-types to 0.7
geo to 0.17 which removes redundant old geo used by polylabel
geojson to 0.22
2021-02-08 16:34:56 -08:00
Michael Kirk
7cd4329866
fix census driven activity model (#507)
It was failing because the version of flatgeobuf we were using was not
compatible with the recent tokio update. Now we point to the unreleased
flatgeobuf which is compatible.
2021-02-08 14:15:58 -08:00
Dustin Carlino
73e405e382 Upgrade most dependencies 2021-02-03 18:13:06 -08:00
Dustin Carlino
d49c6c5d7d Refactor the logic of distributing N people to K buildings. This was
duplicated between the Berlin importer and the new census-based popdat
crate, and I suspect some form of it might get used for the actdev
integration. #424
2021-01-13 11:18:51 -08:00
Dustin Carlino
e99def9124 Switch all error handling to anyhow. Nothing consumes our code as a
library yet, so this is a step forward from string errors. Manually
tested how errors show up (including the nice with_context bit).
2021-01-05 11:35:03 -08:00
Dustin Carlino
d529ae2e5c Upgrade usvg and rand 2021-01-04 09:19:54 -08:00
Dustin Carlino
965d999199 Upgrade misc other dependencies 2021-01-03 20:01:27 -08:00
Dustin Carlino
2945913fb3 Upgrade rand 2021-01-03 19:48:03 -08:00
Michael Kirk
147c36d8fe
remove broken popdat main (#441)
I broke this in acadad8bd

We could fix this by making main async and fetching CensusAreas and
passing them into the generate function... but it doesn't seem like
we're using this any longer anyway.
2021-01-01 21:37:45 -08:00
Michael Kirk
acadad8bdf
fetch US census data from a hosted FlatGeoBuf (#432) 2020-12-31 16:20:32 -08:00
Dustin Carlino
32a3159d19 Express AmenityType as a proper enum. 2020-12-28 12:18:29 -08:00
Dustin Carlino
2b4b92664e Much needed refactor: map.pathfind() returns a Result, stating the failed request. Many call sites were cloning the request and doing this anyway. 2020-12-16 10:06:12 -08:00
Michael Kirk
375e0caeb8
integrate census scenario (#426)
#424
2020-12-14 20:27:57 -08:00
Michael Kirk
8b9b3ce2b1
census areas import (#425)
#424
2020-12-14 17:48:47 -08:00
Michael Kirk
5c09e24225
Correlate some more activities with OSM tags. (#423)
Thanks to Tom Skinner (@trskinner) for the patch!
2020-12-10 21:55:56 -08:00
Dustin Carlino
634b6ad375
Get more pieces of popdat working: (#421)
- adapt the Berlin logic for disaggregating people to buildings
- start some person -> activity -> schedule mappings
- clean up all the unused var warnings
2020-12-10 16:44:42 -08:00
Dustin Carlino
fb9635e8c8
Copy some pieces of the proletariat robot model into the new popdat pipeline: (#419)
- mode choice
- going off-map when there's no building on-map to satisfy a particular
  activity
2020-12-10 14:51:36 -08:00
Dustin Carlino
7e376de03f Start transforming an abstract schedule into a specific list of trips 2020-12-10 12:28:12 -08:00
Michael Kirk
d445ca386a
fix typo (#418) 2020-12-10 09:41:14 -08:00
Dustin Carlino
77a37f897b Wire together a UI to generate census-based scenarios 2020-12-10 09:35:50 -08:00
Dustin Carlino
f23dacbde7 Start some skeleton code for tomorrow's hackathon 2020-12-09 22:11:05 -08:00
Dustin Carlino
8be319300b fold the popdat crate into importer; nothing else should have any dependencies on it 2020-04-25 13:14:48 -07:00
Dustin Carlino
dd1c9f936d plumb through exact off-map locations from popdat->scenario 2020-04-24 22:05:33 -07:00
Dustin Carlino
a25751ea43 add more sanity checks for schedules, because every other map is
crashing

and fix the problem -- SpawnTrip -> TripSpec MUST succeed. auto-abort
some trips, to still warp the person to the right spot.
2020-04-23 13:20:25 -07:00
Dustin Carlino
6ef3a180a3 start to simplify the dynamic nabbing of parked cars 2020-04-22 14:39:02 -07:00
Dustin Carlino
2a98f4ea27 Person now has a fixed pedestrian, car, and bike ID. <start massive breakage> 2020-04-22 09:43:03 -07:00
Dustin Carlino
a39d19f36a remove the tiny fraction of people with non-continuous schedules from
the PSRC scenarios
2020-04-20 17:55:26 -07:00
Dustin Carlino
0574d4316c add some stuff to match people in scenarios with the raw input 2020-04-20 16:52:54 -07:00
Dustin Carlino
8b975f5083 validation was wrong, and also panicking is bad right now 2020-04-20 16:12:09 -07:00
Dustin Carlino
d80513235d defer choosing a starting lane for borders in scenarios, to respond to lane edits 2020-04-19 10:51:23 -07:00
Dustin Carlino
35c193eebc associated parked cars with people, not buildings
pt1: scenario layer
2020-04-15 15:01:49 -07:00