Contributing to FlowFuse

This guide will help you get setup to contribute to the FlowFuse project.

The core of the FlowFuse platform is available under the Apache-2.0 license and we welcome contributions from the community.

Software Requirements

This guide assumes you have a working development environment including:

  • Node.js 18/20
  • Platform build tools
    • Linux: apt-get install build-essential
    • MacOS: xcode-select --install
    • Windows: installed as part of the official node.js installer
      • ☑️ Automatically install the necessary tools must be checked
  • Git

Project Repositories

There are a number of repositories under the FlowFuse GitHub organisation that make up the platform.

Repository Description
FlowFuse This is the core of the platform.
installer The installer for the platform
driver-localfs The LocalFS driver. This deploys instances to the local system.
driver-docker The Docker driver. This deploys instances as containers in a Docker-managed environment.
driver-k8s The Kubernetes driver. This deploys instances as containers in a Kubernetes-managed environment.
nr-launcher The launcher application is used to start and monitor an individual instance of Node-RED in the FlowFuse platform. This includes a number of Node-RED plugins used to integrate with the FlowFuse platform.

Setting Up A Development Environment

With the project split across multiple repositories, setting up a development environment manually takes quite a lot of steps to ensure everything is checked out and configured properly.

To make it easier, you can use the FlowFuse Development Environment project to get set up.

The following steps will get your development environment setup in no time:

git clone https://github.com/FlowFuse/dev-env.git
cd dev-env
npm install
npm run init

This clones all of the main project repositories, installs their dependencies and builds the repositories that need it.

All of the repositories are cloned under the packages directory:

dev-env
└── packages
├── device-agent
├── docker-compose
├── driver-docker
├── driver-k8s
├── driver-localfs
├── file-server
├── flowfuse
├── helm
├── installer
├── nr-file-nodes
├── nr-launcher
└── nr-project-nodes

More details on using the FlowForge Development Environment are available in its documentation.

FlowFuse Code Structure

The FlowFuse/flowfuse repository is the core of the platform and where you'll likely want to begin.

.
├── bin
├── config - build config files
├── docs
├── etc - FlowFuse platform configuration files
├── forge - Platform core code
│ ├── config
│ ├── containers
│ ├── db
│ ├── ee
│ ├── lib
│ ├── licensing
│ ├── monitor
│ ├── postoffice
│ ├── routes
│ └── settings
├── frontend - Frontend code
│ ├── dist - build output - created by `npm run build`
│ ├── public - static assets
│ └── src - vue src
│ ├── api
│ ├── components
│ ├── pages
│ ├── routes
│ └── store
├── test - tests for FlowFuse
└── var - where the database and localfs project directories are created

Development Setup

  1. Create a Stack
  2. Running FlowFuse
  3. Configuring FlowFuse
  4. Mocking email
  5. Testing
  6. VSCode Tips
  7. Team Broker

Create a Stack

You will need to setup the version(s) of Node-RED you want to use in your stacks.

From the flowfuse directory run

npm run install-stack --vers=3.1.9

Where 3.1.9 is the version of Node-RED you want to use in the stack.

Working with Local Nodes

If you want to test local, in-development nodes in FlowFuse, you can create a dedicated stack for this purpose.

npm run install-stack --vers=3.1.9

Navigate to the stacks directory:

cd /var/stacks

Rename the directory to something more appropriate:

mv "3.1.9" "3.1.9-local"

Install your local repository directly into the stack:

cd 3.1.9-local
npm install /path/to/your/nodes-repo

With your stack created from the terminal, you can now add it via the FlowFuse Admin UI - see Managing Stacks.

Running FlowFuse

A number of npm tasks are defined in the package.json file of this repository. To get started from the flowfuse directory use:

npm run serve

This does a couple of things in parallel:

  • Starts the core FlowFuse application and watches the source code for any changes - triggering a restart if needed.
  • Builds the frontend application using WebPack and watches for any changes - triggering a rebuild as needed.

When running like this, the NODE_ENV environment variable gets set to development.

Note: if you have not used the FlowFuse Development Environment, then you will need to run npm run build to build the platform before you can use npm run serve.

Configuring FlowFuse

When running in development mode, the core app will use etc/flowforge.yml for its configuration. As you may want to have a local configuration that you don't want to commit back to git, you can create a file called etc/flowforge.local.yml and it will use that instead. That filename is set to be ignored by git so it won't be accidentally committed.

Mocking email

If you are developing locally and need to enable external email sending, you can either:

  • Setup a local test SMTP server. For example, the Nodemailer project provides a useful app that does the job: https://nodemailer.com/app/
  • Alternatively, set the email.debug option to true in your configuration file and the app will print all emails to its log.

Configuring billing

If you need to develop features covered by the Billing EE feature, you will need to configure the platform with a set of valid Stripe API keys and an EE license.

The development-only EE licence is provided in flowfuse/forge/licensing/index.js. This licence is not valid for production usage.

For FlowForge Inc. employees the configuration is provided in 1Password as 'Stripe Testing Configuration'.

license: ***

billing:
stripe:
key: ***
wh_secret: ***
team_price: ***
team_product: ***
project_price: ***
project_product: ***
device_price: ***
device_product: ***
deviceCost: 10
new_customer_free_credit: 1000
teams:
starter:
price: ***
product: ***
userCost: 0

You will also need to install the Stripe CLI to handle webhook callbacks properly. Install the CLI following their documentation, then run the following command, with the API key using the value of billing.stripe.key from above.

stripe listen --forward-to localhost:3000/ee/billing/callback --api-key ***

Note that due to the way Stripe works, you will receive events for all activity in the configured Stripe account. That means if someone else is actively developing with billing enabled on the same account, you will see their events arrive.

Free Trials

Free trials are implemented as a Stripe Credit that is applied when a FlowFuse user creates their first team and completes billing sign up.

To enable trials, set the billing.stripe.new_customer_free_credit value to a credit amount in cents. For a totally free trial, this amount should match the cost of the Stripe product for the project type to be trialed to be trialed.

The Stripe webhook forwarder must be running as the credit is handled as part of the webhook handling.

Testing

Our testing philosophy follows the principle of:

Write tests. Not too many. Mostly integration [1]

We create both unit tests and system level tests. The former is suitable for well-contained components that need to provide a stable api and behavior to the rest of the code base. The latter is for testing the external behavior of the platform as a whole with as little internal mocking as possible.

We use code coverage reporting as one aspect of assessing our testing coverage. We do not treat 100% coverage as an imperative goal - that can often lead to busy work writing tests that don't provide any real value in understanding the overall quality of the system.

Unit tests should provide sufficient coverage to give us confidence that a component's behavior does not unexpectedly change.

Running tests

To run the tests for the project, you can use the following npm tasks:

  • npm run test - runs the whole test suite, covering code linting, unit and systems tests.
  • npm run lint - runs the linting tests
  • npm run test:unit - runs the unit tests
  • npm run test:system - runs the system tests
  • npm run test:docs - checks the validity of links in the documentation
Testing against PostgreSQL

By default, the tests use an in-memory SQLite database to test against. This is the most self-contained way of testing the platform. But it is also necessary to test against PostgreSQL. To enable the use of PostgreSQL in the tests:

  1. Ensure you have an instance of PostgreSQL running locally. For example, via docker:

       docker run -it -p 5432:5432 --name ff-postgres -e POSTGRES_PASSWORD=secret postgres:14
  2. Enable PostgreSQL mode by setting the following environment variable:

       export FF_TEST_DB_POSTGRES=true

    The database connection can be set using the following env vars (default values shown)

       export FF_TEST_DB_POSTGRES_HOST=localhost
    export FF_TEST_DB_POSTGRES_PORT=5432
    export FF_TEST_DB_POSTGRES_USER=postgres
    export FF_TEST_DB_POSTGRES_PASSWORD=secret
    export FF_TEST_DB_POSTGRES_DATABASE=flowforge_test

Reporting code coverage

The test:* tasks have corresponding code coverage tasks. These tasks run the tests using nyc to generate code coverage information.

  • npm run cover - runs the whole test suite (excluding linting) with code coverage enabled and generates a report (via the cover:report task)
  • npm run cover:unit - runs the unit tests with code coverage enabled. It does not generate the report.
  • npm run cover:system - runs the system tests with code coverage enabled. It does not generate the report.
  • npm run cover:report - generates a report of the code coverage. This is printed to the console and generates a browsable HTML copy under coverage/index.html

VSCode Tips

To step debug in VSCode

  1. Open launch.json config and enter the JavaScript below
  2. Choose Start-Watch from the "Run and Debug" menu
  3. Press ▶️ or F5 to start debugging

There are 2 other "Run and Debug" entries in the menu...

  • "Attach by Process ID" - this will allow you to attach to a launched driver
  • "Debug Current Test" - this will enable you to step debug a test (starts debugging the currently open test file)

{
// Use IntelliSense to learn about possible attributes.
// Hover to view descriptions of existing attributes.
// For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
"version": "0.2.0",
"configurations": [
{
"command": "npm run start-watch",
"name": "Start-Watch",
"request": "launch",
"type": "node-terminal",
"env": {
"NODE_ENV": "development"
}
},
{
"name": "Attach by Process ID",
"processId": "${command:PickProcess}",
"request": "attach",
"skipFiles": [
"<node_internals>/**"
],
"type": "node"
},
{
"type": "node",
"request": "launch",
"name": "Debug Current Test",
"program": "${workspaceFolder}/node_modules/mocha/bin/_mocha",
"args": [
"--no-warnings",
"-u",
"bdd",// set to bdd, not tdd
"--timeout",
"999999",
"--colors",
"${file}"
],
"env": {
"NODE_ENV": "development"
},
"internalConsoleOptions": "openOnSessionStart"
}
]
}

  1. https://kentcdodds.com/blog/write-tests ↩︎