Skip to content

Latest commit

 

History

History
88 lines (66 loc) · 4.18 KB

README.md

File metadata and controls

88 lines (66 loc) · 4.18 KB
Raw

@fluid-example/app-integration-external-controller

This demo is a work-in-progress

Dice Roller is a basic example that has a die and a button. Clicking the button re-rolls the die and persists the value in the root SharedDirectory. The Fluid Container is defined in container/, the data object is defined in dataObject/.

This implementation demonstrates plugging that Container into a standalone application, rather than using the webpack-fluid-loader environment that most of our packages use. This implementation relies on Tinylicious, so there are a few extra steps to get started. We bring our own view that we will bind to the data in the container.

Getting Started

You can run this example using the following steps:

  1. Run npm install and npm run build:fast -- --nolint from the FluidFramework root directory.
    • For an even faster build, you can add the package name to the build command, like this: npm run build:fast -- --nolint @fluid-example/app-integration-external-controller
  2. In a separate terminal, start a Tinylicious server by following the instructions in Tinylicious.
  3. Run npm run start from this directory (azure/packages/external-controller) and open http://localhost:8080 in a web browser to see the app running.

Testing

    npm run test:jest

For in browser testing update ./jest-puppeteer.config.js to:

  launch: {
    dumpio: true, // output browser console to cmd line
    slowMo: 500,
    headless: false,
  },

Data model

Dice Roller uses the following distributed data structures:

  • SharedDirectory - root

Backed Locally and running with live Azure Fluid Relay service instance

We can connect to a live Azure Fluid Relay instance by passing in the tenant ID and discovery endpoint URL, or we can connect to a local Tinylicious server for development purposes by setting the type as "local".

To run the the AzureClient against our local Tinylicious instance, we set the type as "local" and make use of InsecureTokenProvider. For the latter, we pass in two values to its constructor: a key string, which can be anything since we are running it locally, and an object identifying the current user. For running the instance locally, the endpoint URL would point to the Tinylicious instance on the default values of http://localhost:7070.

To launch the local Tinylicious service instance, run npx tinylicious from your terminal window.

When running the live Azure Fluid Relay Instance, we would require the tenant ID and service discovery endpoint URL. We make use of AzureFunctionTokenProvider which takes in the Azure function URL and an object identifying the current user, thereby making an axios GET request call to the Azure Function. This axios call takes in the tenant ID, documentId and userID/userName as optional parameters. The Azure Function is responsible for mapping the tenantId to tenant key secret to generate and sign the token such that the service will accept it.

const connectionConfig: AzureConnectionConfig = useAzure
    ? {
          type: "remote",
          tenantId: "YOUR-TENANT-ID-HERE",
          tokenProvider: new AzureFunctionTokenProvider(
              "AZURE-FUNCTION-URL" + "/api/GetAzureToken",
              { userId: "test-user", userName: "Test User" },
          ),
          endpoint: "ENTER-DISCOVERY-ENDPOINT-URL-HERE",
      }
    : {
          type: "local",
          tokenProvider: new InsecureTokenProvider("fooBar", user),
          endpoint: "http://localhost:7070",
      };

In this way, we can toggle between remote and local mode using the same config format. We make use of AzureFunctionTokenProvider for running against live Azure Fluid Relay instance since it is more secured, without exposing the tenant secret key in the client-side code whereas while running the service locally for development purpose, we make use of InsecureTokenProvider.