Setup OSRM and Serve and Query Data

Source

Converted from https://github.com/Project-OSRM/osrm-backend/pkgs/container/osrm-backend

Prerequisite: Apptainer Installed

Please see the Apptainer - Installing Apptainer guide for details. You do not need suid for OSRM, you do need fuse.

Linux Installation

Assuming Ubuntu (including Windows WSL2 Ubuntu), follow these instructions.

Instructions for other sets of pre-built Linux packages can be found here.

MacOS Installation

More soon!

Clone the Repository

A one-time task.

cd some_path
git clone git@github.com:upenn/OSRM_simple.git
cd OSRM_simple

Build the apptainer Image

A one-time task.

cd path/to/OSRM_simple
apptainer build OSRM.sif docker:osrm/osrm-backend

Prepare the OSM Data for Serving

A one-time task per geo area + profile to get and process the osm.pbf.

New York (the state) in this example. Details for the URLs at https://download.geofabrik.de/

cd path/to/OSRM_simple
./OSRM_setup.sh PBF_FILE_URL PROFILE
# example (new-york, car):
./OSRM_setup.sh https://download.geofabrik.de/north-america/us/new-york-latest.osm.pbf car

Start the OSRM Server

The OSRM_start.sh script file makes starting the OSRM server with your location and profile easy.

cd path/to/OSRM_simple
./OSRM_start LOCATION PROFILE
# example (new-york, car):
./OSRM_start.sh new-york car

Stop the OSRM Server

Once you're done with the server, stop it:

cd path/to/OSRM_simple
# get the name of your server instance:
apptainer instance list
apptainer -v instance stop INSTANCE_NAME
# example (new-york):
apptainer instance stop OSRM-new-york-car-5000

Do Work With R (Example)

Set up and use R with OSRM.

This is some toy work in R ... take a look at this post for some R code using the R osrm package.

Prerequisites - Ubuntu

Including Windows WSL2 Ubuntu ... install some required packages:

sudo apt update
sudo apt install libudunits2-dev gdal-bin gdal-data gdal-plugins libgdal-dev

Setting Up R

If you're using R, you'll want to do this one-time tasks for the R package osrm.

Rscript --no-save -e 'install.packages("osrm", dependencies=TRUE)'

This can take some time! Grab a ☕︎. Once complete, you're all set to use R with OSRM!

Run the Example(s)

Build a complete trip matrix between all points in the NYC example points data (11 points):

Rscript --no-save trip_table.R input/osrm_example_points_nyc.csv 5000

Process all trip routes in the NYC example trip data (5 trips):

Rscript --no-save trip_route.R input/osrm_example_trip_nyc.csv 5000

Results in output.