Welcome to the repository for the MapRoulette back-end server code. The MapRoulette back-end exposes the MapRoulette API, which the MapRoulette front-end web application depends on. The source code for the web application is in a separate repository.
If you just want to deploy the MapRoulette back-end, we have a 🚢 Docker image 🚢 for that. This is especially useful if you want to contribute to the MapRoulette front-end and don't intend to touch the back-end.
MapRoulette depends on several technologies for building and running the project:
- Java 17 JDK
- PostgreSQL with PostGIS
- Scala Build Tool
MapRoulette is a complex tool and depends on several other tools (Mapillary, Overpass API, OpenStreetMap, PostgreSQL, and others) that need to be correctly configured within the application's HOCON configuration files.
The initial setup is not trivial, so the documentation is split to specific steps where each step can be easily validated. The validation at each step is a helpful sanity check so please, regardless of your experience, take the time to verify along the way.
To get started you'll need to install Docker, JDK 17, and sbt.
Docker is a virtualization tool and feel free to use Docker Desktop, or Podman, or even Rancher Desktop.
sdkman is a great tool to install a specific build of the JDK to keep your environment as similar to production as possible. It also handles fetching x8664 and aarch64 builds automatically. Follow the installation steps and install the JDK and sbt using a command similar to:
sdk install java 17.0.12-tem
sdk install sbt 1.10.1
Within a terminal check that docker, javac, and sbt are working.
docker run hello-world
- Verify docker works
javac -version
- Verify javac is JDK 17
- Apple Silicon: Verify that the JDK is an aarch64 build and not x8664.
sbt -version
MapRoulette development assumes a database is running on the local system within a docker container.
Below is a sample command to run a PostGIS database within a container and sets necessary ports/credentials.
- NOTE: Apple Silicon: Use
ghcr.io/baosystems/postgis:13-3.3
docker image since postGIS does not yet publish aarch64 images. - NOTE: No volume mount is used so the database's data will be deleted when the container is deleted.
If you'd like to keep the data external of the container, be sure to add
--volume "/some/path/here/postgres-data":/var/lib/postgresql/data
to the docker call.
docker run \
-d \
-p 5432:5432 \
--name maproulette-postgis \
--restart unless-stopped \
--shm-size=512MB \
-e POSTGRES_DB=maproulette-db \
-e POSTGRES_USER=maproulette-db-user \
-e POSTGRES_PASSWORD=maproulette-db-pass \
postgis/postgis:13-3.3
- NOTE: If there's a port conflict, you probably have another pg instance running. Check with
docker ps
. - NOTE: It is helpful to see logs with
docker logs -f maproulette-postgis
- NOTE: To stop the container, that'd be
docker stop maproulette-postgis
. Then you can start it again usingdocker start maproulette-postgis
.
Clone the maproulette-backend repository and cd
to that directory, and create conf/dev.conf
using the example file:
cp conf/dev.conf.example conf/dev.conf
Edit conf/dev.conf
and set db.default based on the previous step's POSTGRES_DB, POSTGRES_USER,
and POSTGRES_PASSWORD values:
db.default {
url="jdbc:postgresql://localhost:5432/maproulette-db"
username="maproulette-db-user"
password="maproulette-db-pass"
}
Alternatively, you can configure the backend using environment variables. The
variables MR_DATABASE_URL
, MR_DATABASE_USERNAME
and MR_DATABASE_PASSWORD
control the same values as the config parameters shown above. Look in conf/ application.conf
for what else can be overridden via environment variables.
Any pattern like ${?FOO}
in that file will be replaced with the value of the
environment variable FOO
at runtime.
Now start the MapRoulette server! Run this command in a terminal, not within Intellij/vscode:
sbt -J-Xms4G -J-Xmx4G -J-Dconfig.file=./conf/dev.conf -J-Dlogger.resource=logback-dev.xml run
There should be some output that looks like this:
--- (Running the application, auto-reloading is enabled) ---
[info] p.c.s.AkkaHttpServer - Listening for HTTP on /0:0:0:0:0:0:0:0:9000
(Server started, use Enter to stop and go back to the console...)
That's the expected output. And when you need, like it says, use Enter to stop the server.
Open a new terminal so that the MapRoulette server is not stopped. Verify:
- Check that the database is running,
docker inspect -f '{{.State.Running}}' maproulette-postgis
- Start the MapRoulette server if it's not already running (see the previous section for the
sbt
command) - Verify that the local Swagger docs load in a web browser
- Do this GET call to get the list of challenges:
curl http://127.0.0.1:9000/api/v2/challenges
- The result may be an empty array or some data, as long as it's not a failed call
- Tip: the response can be formatted by piping to
jq
- Open the terminal that has
sbt
running the server. Verify that there are log messages printed.
If there's a failure at this point READ THE LOG MESSAGES, checking for simple setup issues.
A few known setup misconfigurations could be:
- Does the server log that it can't communicate with the database? "HikariPool Connection is not available, request timed out after 30000ms"
- Check the
conf/dev.conf
MapRoulette server configuration file for typos, accidental setting overrides, etc - Check that the MapRoulette server conf and postgis are using the same database, user, password.
- Check the
- Recursive loading of Guice injected object: 'java.lang.IllegalStateException: Recursive load of: org.maproulette.framework.service.ServiceManager.()'
- We've seen this with Apple Silicon aarch64 systems. Possibly related to mixing aarch64 and x8664 JDKs. Intellij 2022.3 pre-release seems to work fine
- Having some other issue?
- Please file a GitHub issue with a detailed description on what was tried and include the error message.
With the database and MapRoulette server working together, it's now time to get the front-end login workflow working!
High level the user visits to MapRoulette, clicks 'login' and is sent to OpenStreepMap to login then, after authentication, MapRoulette is able to make OSM changes of behalf of the user. This is done by registering the local dev instance of MapRoulette as an OSM OAuth 2.0 Application to create keys, and then the those keys are pasted into the MapRoulette server configuration. More details on how OSM uses OAuth is on the OpenStreetMap OAuth documentation.
Within OpenStreetMap, MapRoulette needs to be setup as an "OAuth 2 Application" so that changes in MapRoulette are associated with a user's OSM account.
- Login to the development OpenStreetMap server (make an account if needed)
- In the top right, click the user and go to
My Settings
- Then click
OAuth 2 applications
tab - At the bottom of the OAuth2 settings page, click
Register new application
- Register using these settings:
Name
: maprouletteDevLocalhost (any name is fine)Redirect URIs
: http://127.0.0.1:9000Confidential application
: Keep this checkedPermissions
: SelectRead user preferences
Modify user preferences
(MapRoulette saves the user's MR API key in OSM)Modify the map
(MapRoulette tasks edit the map)
Click Register
and note the displayed Client ID and the Client Secret. Copy and paste those keys into the MapRoulette
conf/dev.conf
file's osm section where it has "CHANGE_ME":
osm {
# The OSM OAuth 2.0 Client ID and Client Secret, respectively
consumerKey="CHANGE_ME"
consumerSecret="CHANGE_ME"
}
Stop and start the MapRoulette server to load the updated configuration.
Clone the maproulette3 repository and follow the steps in the DEVELOPMENT.md "Run the UI from Docker" section.
Be sure to execute the docker run
step that starts the UI.
- Open http://localhost:3000/ and attempt to log in
- This should redirect to the development OpenStreetMap server and redirect back to MapRoulette
- The username should show in the top-right corner
- Create a new Project (this is the easiest way to verify), or a new challenge, or update a project/challenge/task, or do some other work that causes various request types to the MapRoulette server
Known setup issues:
- Exceptions with 'java.lang.NumberFormatException: For input string: "CHANGE_ME"'. Edit the
conf/dev.conf
and verify that these are not "CHANGE_ME":osm.consumerKey
osm.consumerSecret
maproulette.super.key
maproulette.super.accounts
The server is working end-to-end on command line! Time to import into IntelliJ. Stop the running MapRoulette server from the previous step, if it's still running.
- Open IntelliJ
- Click
Plugins
and verify that the Scala plugin is installed - Click
Open
and navigate to the MapRoulette source directory andOpen as Project
- The project will load
- Create a new
sbt Task
runtime configuration-
Name it
MapRoulette Server
or similar -
Set Tasks to
run
-
Uncheck 'Use sbt shell`
-
Use these
VM parameters
-Xms4G -Xmx4G -Dconfig.file=./conf/dev.conf -Dlogger.resource=logback-dev.xml
-
Open the front-end UI http://localhost:3000/ and attempt to log in. It should function.
- Open
dev.conf
in a text editor and change at least the following entries:super.key
: a randomly chosen API key for superuser accesssuper.accounts
: a comma-separated list of OSM accound IDs whose corresponding MapRoulette users will have superuser access. Can be an empty string.mapillary.clientId
: a Mapillary Client ID, needed if you want to use any of the Mapillary integrations.
During development, please set -Dlogger.resource=logback-dev.xml
to have the best experience. The logback dev file sets
many items to the devel level and logs all HTTP request paths and response times.
Any changes to the conf/logback-dev.xml
will be loaded within about 5 seconds and a service restart is not needed.
To have all request headers sent by the client logged, update the conf/logback-dev.xml
"org.maproulette.filters" entry to TRACE.
Windows is not officially supported. However, there is an unofficial setup guide.
MapRoulette now supports transmission of emails, for example to inform users
when they receive new in-app notifications. You will need access to an SMTP
server to send emails, and will also need to add SMTP configuration settings to
your configuration file (or whatever configuration mechanism you're using).
play.mailer.host
is the only required SMTP setting, but most SMTP servers
will also want a username and password.
play.mailer.host = "smtp.server.com"
play.mailer.user = "smtpusername"
play.mailer.password = "secret"
Many additional SMTP configuration options are available -- see play-mailer for a full list.
Note: If you're doing development, you can set
play.mailer.mock = yes
to simply log emails instead of transmitting them
It's also important that you set the emailFrom
setting to the email address
you wish emails to come from, and that you ensure publicOrigin
is set so that
links in emails will properly point to your server.
emailFrom = "[email protected]"
publicOrigin = "https://www.yourserver.com"
By default, notification emails that are to be sent immediately are processed by a background job every 1 minute. Both the frequency and max number of emails to process in a single run can be controlled.
notifications.immediateEmail.interval = "1 minute"
notifications.immediateEmail.batchSize = 10 # max emails per run
Notification emails that are to be sent as a digest are initially processed at 8pm local server time by default, and then every 24 hours thereafter (i.e. daily digests). Both of these settings can be customized. There is not currently a maximum limit to the number of emails for digest emails.
notifications.digestEmail.startTime = "20:00:00" # 8pm local server time
notifications.digestEmail.interval = "24 hours" # once daily
The wiki for this repo has some information on creating challenges.
Challenge API has further information about creating challenges through the API.
See also the Swagger API documentation. You can view the documentation by going to the URL /docs/swagger-ui/index.html
on any MapRoulette instance.
Please fork the project and submit a pull request. See Postman Docs for information on API Testing. The project is integrated with Travis-CI, so PR's will only be accepted once the build compiles successfully. MapRoulette also uses Scalafmt as it's code formatter. This is too keep the code style consistent across all developers. The check will be run first in Travis for the build, so if there are any code style issues it will fail the build immediately. IntelliJ should pick up the formatter and use Scalafmt automatically, however you can also use sbt scalafmt
to format any and all code for you.
Bug and feature requests are best left as an issue right here on Github. For other things, contact [email protected]
MapRoulette now also has a channel #maproulette on the OSM US Slack community.