The system in structured in a single Cargo workspace and can be built with cargo while standing in the src directory.

cargo build

The following build dependencies are required:

• OpenSSL
• protobuf
• cargo

NVIDIA container toolkit is also required for local deployment and testing.

For local development with real data the engine container can be started and will scrape all datasources and store them in your local postgres instance. The only data missing will be AIS position data which only exists in our azure database. The engine will use quite abit of time on a fresh run, consult the docker logs to view its progress.

Commands to run a local instance:

docker-compose up -d postgres
docker-compose -d --build engine

For a local api instance you will also need to start the duckdb service:

docker-compose up -d postgres
docker-compose up -d duckdb
docker-compose up -d api

The api will then be availabe at http://localhost:8080, swagger at http://localhost:8080/swagger-ui/.

We use sqlx for all database interactions and migrations. To be able to build the system you will need a running and migrated database, a dedicated migration db service exists in the compose file in this repository. Run the following commands to start and migrate your local database:

cd src/postgres
docker-compose up -d migration-db
cargo sqlx migrate run

See the sqlx-cli crate documentation for more information on how it operates.

The following build dependencies are required:

• sqlx-cli

We use Dockertest as our test library which will manage container lifecycles in our tests. All tests will reuse the same postgres container across multiple cargo test invocations and create their own isolated database within that instance. Prior to running tests remember to build the test-db service if you've added any migrations or if its the first run:

docker-compose build test-db

By default not all tests are run, to enable cache tests and db migrations run the following (you will need the master db state image):

echo <GITHUB_TOKEN> | docker login ghcr.io/orcalabs/kyogre -u <USERNAME> --password-stdin
docker pull ghcr.io/orcalabs/kyogre/test-postgres:master
cargo test --features all-tests

If you experience Too many open files errors from tests either increase your max amount of file descriptors (ulimit -n NUM, linux) or run cargo test with limited test threads cargo test -- --test-threads=NUM

The following build dependencies are required:

• docker
• docker buildkit
• docker-compose (to simplify image building)
• python 3.9 with the packages specified here

To access the private Orca Labs packages you will need to be invited to the Orca Labs cloudsmith organization, then add your credentials to your credentials helper:

git config --global credential.helper store

echo "https://USERNAME:API-KEY@dl.cloudsmith.io" > ~/.git-credentials

To view the code documentation run the following from the src directory (the docs will popup in your browser):

cargo doc --open --no-deps --workspace --document-private-items