Update to graphserver-based bike mapper. Uses custom lua profiles based on existing bicycle model provided by OSRM.
Make sure you have XCode installed with command line tools (app store). Homebrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install.sh)"
brew install osmosis boost git cmake libzip libstxxl libxml2 lua tbb ccache postgis
brew cask install docker
mkdir -p data/elevation
curl -o ./data/california-latest.osm.bz2 https://download.geofabrik.de/north-america/us/california-latest.osm.bz2
gunzip ./data/california-latest.osm.bz2
osmosis --read-xml file=./data/california-latest.osm --bounding-box left=-123.404077 bottom=37.171696 top=38.619150 right=-121.674775 --write-pbf ./data/bay_area.osm.pbf
For testing sample that includes just the wiggle and golden gate bridge, use bounds (37.8343042,-122.487191), (37.764910, -122.414635).
For all of San Francisco: (37.70230166835287, -122.5241176570151), (37.84002900260014, -122.34455987162356)
Highest resolution is 1/3rd arc second. Adjust lat/lng as needed (values represent top-right corner of 1-degree bounding box. You can play with tile size (100x100) - this will impact loading vs. access time because index is on each cell. Place all the files you need to cover the area in a directory ./data/elevation. Note that the indexing is by top-right corner lat/lng. You can download from nationalmap.gov or directly if you know the lat/lon you are looking for.
For N38W122 to N39W123 (USGS 1/3rd arc second data):
for lat in 38 39; do
for lng in 122 123; do
curl "https://prd-tnm.s3.amazonaws.com/StagedProducts/Elevation/13/TIFF/n${lat}w${lng}/USGS_13_n${lat}w${lng}.tif" > "./data/elevation/USGS_13_n${lat}w${lng}.tif"
done;
done
Assuming all data are in correct folders, run:
virtualenv --python=python3.8 env
source env/bin/activate
pip3 install -r ./scripts/requirements.txt
export REDIS_URL=<VALUE FROM HEROKU>
python3 ./scripts/elevation_mapper.py --skip_redis <optional, does not write to redis> --skip_file <optional, does not write to elevations.csv> --input_file <optional, defaults to 'bay_area'>
Assuming you didn't disable it, this will write a file, ./data/elevation/elevation.csv, with a mapping from node_id to elevation in meters, and publish to redis (production). Any errors are recorded in errors.csv.
Choose processed input data and copy to docker file.
cp ./data/bay_area.osm.pbf ./docker/extract.osm.pbf
or
cp ./data/bay_area_sample.osm.pbf ./docker/extract.osm.pbf
cp ./data/elevation/elevation.csv ./docker/elevation.csv
docker build -t bike-mapper docker
This image can be posted to a repository to deploy on heroku, etc. To create custom profiles, pass argument to docker build and specify alternative name.
docker build -t bike-mapper-safe --build-arg profile=bicycle_extra_safe.lua docker
docker run -t -i -p 5000:5000 bike-mapper osrm-routed --algorithm mld ./data/extract.osrm
Ideally you would run alternate images on different ports, e.g.
docker run -t -i -p 5001:5001 bike-mapper-safe osrm-routed --algorithm mld ./data/extract.osrm
curl -s "http://127.0.0.1:5000/route/v1/driving/-122.424474,37.766237;-122.443049,37.775325?steps=false" | jq -r ".routes[0].geometry"
Should return }foeF``fjVjA`h@uS~@?h@iI`Ah@fI{Dd@h@hIiVtC~Bb^{Dd@j@vI
docker pull osrm/osrm-frontend
docker run -p 9966:9966 osrm/osrm-frontend
Note that you may need to up the standard 2 GB memory allocation in Docker (for Bay Area 16GB works okay). This is done using whatever platform you are developing on locally.
heroku git:remote -a bikesy-api
cd docker
heroku container:login
heroku container:push web
heroku container:release web
If needed, .
Install doctl and init repo (if needed).
brew install doctl
doctl auth init --context bikesy
doctl auth switch --context bikesy
Add new registry (one time)
doctl registry create bikesy
doctl registry login
Each profile is a single app in DO and require their own lua profile. Add as many as needed, push to docker repository, and deploy app. The convention is as followed:
h (hills): low -> low tolerance, will avoid slopes. high -> high tolerance, will take standard route. s (safety): low -> low safety, will consider most roads. high -> high safety, will prefer bike lanes and non-primary roads.
docker build -t bike-mapper-h-med-s-med --build-arg profile=bicycle-h-med-s-med.lua docker
docker tag bike-mapper-h-med-s-med registry.digitalocean.com/bikesy/bike-mapper-h-med-s-med
docker push registry.digitalocean.com/bikesy/bike-mapper-h-med-s-med
You can create postgis map of only bike lanes using the following SQL:
create temporary sequence tmp;
create table bikelanes as select nextval('tmp') as id, way_id, cycleway, st_makeline(geom) as geom from (select way_id, ways.tags -> 'cycleway' as cycleway, ways.tags -> 'geom, sequence_id from ways join way_nodes on (ways.id = way_nodes.way_id) join nodes on (nodes.id = way_nodes.node_id) where ways.tags -> 'cycleway' is not null order by way_id, sequence_id) foo group by way_id, cycleway;
Installed as submodule. Follow above link for any installation issues.
mkdir -p /usr/local/Cellar/osmosis/0.47/bin/plugins
jar cf /usr/local/Cellar/osmosis/0.47/bin/plugins/srtm.jar ./osmosis-srtm-plugin/src/main/java/de/locked/osmosis/srtmplugin/*.java
echo "de.locked.osmosis.srtmplugin.SrtmPlugin_loader" > /usr/local/Cellar/osmosis/0.47/libexec/config/osmosis-plugins.conf
psql -c "create database bikemapper;" -d postgres -U postgres
psql -c "create extension postgis; create extension hstore;" -d bikemapper -U postgres
psql -f /usr/local/Cellar/osmosis/0.47/libexec/script/pgsnapshot_schema_0.6.sql -d bikemapper -U postgres
osmosis --read-xml file=./data/bay_area.osm --write-pgsql host=localhost database=bikemapper user=postgres
Elevation data
psql -c "drop table if exists node_elevation; create table node_elevation (node_id bigint, elevation float);" -d bikemapper -U postgres
echo "\copy node_elevation from elevation.csv with csv delimiter as ','" | psql -d bikemapper -U postgres
psql -c "alter table node_elevation add primary key (node_id);" -d bikemapper -U postgres