Firebase Node.js Quickstarts

To run the Realtime Database quickstart, follow these steps to configure your environment and service account:

1. Firebase Project: Create a project in the Firebase Console.
2. Service Account: Create a service account (refer to Adding Firebase to your Server) and place the JSON credentials file in the project directory.
3. Configuration: Open index.js and update the following placeholders:
   • <PROJECT_ID>: Your Firebase Project ID.
   • <PATH_TO_SERVICE_ACCOUNT_CREDENTIAL_FILE>: The file path to your service account JSON.
   • Email Transport: Configure your email transport settings within index.js to enable notifications.
4. Installation: Run npm install to install dependencies.
5. Execution: Run node index.js to start the server.

To see the app in action, you must also run a client-side quickstart (Web, iOS, or Android) and use it to publish posts. The Node.js server will then automatically update star counts and send email notifications.

This quickstart provides a CLI interface via index.js to manage your Remote Config templates. Navigate to the config directory to run these commands.

Get the active template

Retrieve the current template from the server. This command saves the template to a local file named config.json and prints the current ETag to the console.

node index.js get

Update the template via a file

To update the template using a local JSON file (like the config.json generated by the get command):

1. Modify the values in your JSON file. If the template is empty, ensure you include conditions and parameters.
2. Run the publish command:

node index.js publish

Update the template in memory

Retrieve the current template and apply updates programmatically in memory without using a local file. This specific command sets a condition (android_en) and a corresponding parameter.

node index.js update

# Get template
node index.js get

# Publish template from file
node index.js publish

# Update template in memory
node index.js update

The manage-ml.js script provides a CLI interface to manage your Firebase ML models.

Available Commands and Usage

• List models: View existing models, their IDs, and tags.

  ./manage-ml.js list

• Create a new model: Upload a new .tflite model file and optionally assign tags.

  ./manage-ml.js new <model_name> -f <path_to_tflite_file> --tags <tag1>,<tag2>

  Example: ./manage-ml.js new yak_detector -f ~/yak.tflite --tags vision,experimental

• Create a model from an existing resource: Upload a model using a specific resource path.

  ./manage-ml.js new <model_name> -a <resource_path>

  Example: ./manage-ml.js new emu_detector -a projects/12345/locations/us-central1/models/ICN12345

• Update a model: Modify an existing model (e.g., removing tags).

  ./manage-ml.js update <model_id> --remove_tags <tag_to_remove>

  Example: ./manage-ml.js update 8717019 --remove_tags experimental

• Delete a model: Remove a model by its ID.

  ./manage-ml.js delete <model_id>

# Example session workflow
./manage-ml.js list
./manage-ml.js new yak_detector -f ~/yak.tflite --tags vision,experimental
./manage-ml.js update 8717019 --remove_tags experimental
./manage-ml.js delete 8716959

The application requires two configuration files in the ./auth-sessions directory:

1. Web Configuration (config.js)

Create a config.js file containing your Firebase Web SDK configuration. You can obtain this by clicking the "Web setup" button in the Firebase Console Authentication page.

const config = {
  apiKey: "...",
  authDomain: "...",
  databaseURL: "...",
  storageBucket: "...",
  messagingSenderId: "..."
};
module.exports = config;

2. Admin SDK Credentials (serviceAccountKeys.json)

Since the app uses the Firebase Admin SDK, you must provide a service account private key. Save your generated private key as serviceAccountKeys.json in the ./auth-sessions folder. Warning: Never expose this file publicly.

Authentication Setup

In the Firebase Console, navigate to Authentication > SIGN-IN METHOD and enable:

• Google
• Email/Password

Ensure your authorized domain (e.g., localhost) is whitelisted in the Firebase Console.