Quickstart

This page gets you from zero to a locally running, tested change. The commands below are taken from the README of each repository; if they disagree, the README of the repository is the source of truth.

If you do not know yet which repository you need, read the Repository map first.

1. Prerequisites

You only need the tools for the repositories you work on:

Repository	Requirements
Frontends	Node.js 22 LTS (or another version supported by the Angular version used) and npm
Learning-APIs	Java 25 and Docker (Quarkus Dev Services start the databases and the Orchestration API). The Maven wrapper is included
Orchestration-API	Java 25 and a running Docker daemon
Federated-Learning-Communication-API	Go (version in go.mod), Docker for the images, Python 3 for the end-to-end tests
Tool-Build-Pipeline	Docker (the container needs access to the Docker socket), Python for the tests
Python-Tool-API	Python
Documentation	Node.js LTS and Git LFS

2. Getting the repositories

All repositories are in the FedLearnNet GitHub organization. Clone the ones you need, for example:

git clone https://github.com/FedLearnNet/Learning-APIs.git
git clone https://github.com/FedLearnNet/Frontends.git

3. Setting up the development environment

Frontends

npm install
npm run start-local-fl-net     # local (Client) frontend on http://localhost:4200
npm run start-global-fl-net    # global (Platform) frontend on http://localhost:4201

See Frontend local setup for brands, environment files, and required backend services.

Learning-APIs

./mvnw clean install                # build all modules and run the tests

cd local-learning-api
cp .env.example .env                # fill in the secrets
cp orch_secrets.env.example orch_secrets.env
./mvnw compile quarkus:dev          # live reload, Dev UI at /q/dev/

The same applies to global-learning-api. Development ports: global-learning-api 8080, local-learning-api 8081, datamodeler-api 8086. The secrets are described in Environment secrets.

Orchestration-API

cp .env.example .env          # fill in the secrets
./mvnw compile quarkus:dev    # http://localhost:8082 · Dev UI /q/dev/ · Swagger UI /q/swagger-ui

Alternatively, docker compose up --build runs the API together with PostgreSQL (API on port 8093).

Federated-Learning-Communication-API

go build -o relay ./cmd/relay
go build -o controller ./cmd/controller

./relay --mode dev --tls-mode self-signed --domain localhost
./controller --address-tcp localhost:9141 --mode dev --tls-mode self-signed

Alternatively, docker compose up starts a relay and a controller from the published staging images.

Tool-Build-Pipeline

cp .env.example .env          # fill in the values
docker compose up --build

Documentation

See Documentation workflow.

4. Making a change

1. Find or open an issue in the repository you are changing.
2. Create a branch from the develop branch of that repository. If the develop branch does not exist, create a branch from main. Use a descriptive name for the branch, if you work based on an image you can get the branch name from the issue, e.g. 5-fix-date-to-iso8601.
3. Make the smallest change that solves the issue. The documentation of each repository has further information on this.
4. Run the checks below, then open a pull request.

The full process is described in Development workflow.

IMPORTANT

If you contribute also update the documentation if your change affects:

• user-facing behavior, e.g. a new feature, a changed UI flow, or a changed API changing how the frontend behaves
• If your change is internal please check the documentation of the service you changed. If your changes affect whats documented there, update the documentation accordingly. For example, if you change the learning flow in the local-learning-api, update the federated learning flow page.

This helps keep the system usable over time!

5. Testing and building

Run at least the checks that the CI of the repository runs:

Repository	Commands	Run by CI
Frontends	npm run lint, npm test, npm run cy:local / npm run cy:global	npm run lint
Learning-APIs	./mvnw verify (or mvn test inside one module)	mvnw verify including Checkstyle and SpotBugs
Orchestration-API	./mvnw verify	mvnw verify
Federated-Learning-Communication-API	gofmt -l . (must print nothing), go test ./..., bash e2etests/e2e_test.sh after pip install -r e2etests/requirements.txt	formatting, unit tests, end-to-end tests, golangci-lint, govulncheck
Tool-Build-Pipeline	pip install -r requirements.test.txt, then pytest -q	pytest
Python-Tool-API	pip install -r requirements.txt pytest, then pytest tests/	pytest, package build
Documentation	npm run typecheck, npm run build	typecheck and build

For frontend-specific guidance on which checks to run, see Testing and review.

Where to go next

Continue with the Development workflow, or go back to the Welcome page for the recommended reading order.