SAFE Desktop App Tutorial

In this tutorial, we will create an Electron application. Electron allows you to create and build cross-platforms desktop applications using JavaScript.

To create this app, we will use the @maidsafe/safe-node-app npm package which exposes the safe-app-nodejs API that we need not only to interact with the SAFE Authenticator, but also to connect and interact with the SAFE Network.

We are going to use the SAFE App Electron quick start boilerplate, which is based on the Electron quick start boilerplate. If you would like additional information on creating Electron applications, visit the Electron tutorial site.

The boilerplate implements a simple single page application using Angular.js for rendering the UI components and the data. The application implements a trip planner, showing the list of trips planned by the user, allowing him/her to add or remove trips from the list. In this tutorial, we will be adding the implementation to store (and retrieve) the list of planned trips in the SAFE Network.

Pre-requisites

First you need to make sure you have the following tools installed to be able to work with this tutorial:

Git: to be able to clone the boilerplate code
Node.js v8.11.1 (which comes with npm v5.6.0) to be able to run the application since it's a Node.js application. All the steps in this tutorial are explained using npm, if you otherwise prefer to use yarn, please make sure you install v1.6.0. Note that the use of yarn is not required and totally optional
If you are using Ubuntu or Debian 9 as OS, libgconf-2-4 and/or build-essential dependencies might be missing. Please install them with Synaptic Package Mgr., or with apt from a shell console: $ sudo apt-get install libgconf-2-4 build-essential
If you are using Windows, run npm install --global --production windows-build-tools.
If you decide to use yarn and are using Windows, run yarn config set child-concurrency 1 because yarn attempts to build modules concurrently with multiple child processes, which causes intermittent timing issues on Windows.

Install a SAFE Browser

Since the application will be authorising with the Authenticator to get the credentials needed to then connect to the SAFE Network, we first need to have an instance of the SAFE Browser installed.

You can find the links to download the SAFE Browser package from the SAFE Network website, or directly from the SAFE Browser GitHub releases repository. It's recommended to always use the latest available version.

Note that there are packages for each of the supported platforms, i.e. Linux, Windows and macOS. Also note there are two type of packages for each of the supported platforms:

Peruse-<version>-<platform>-<arch>.zip: SAFE Browser package built to use the live SAFE Network
Peruse-<version>-<platform>-<arch>-dev.zip: SAFE Browser package built to use the mock routing. This will create a local temporary file and you won't need to connect to the live network.

In this tutorial we will be using the SAFE Browser package that is built to work with the mock network. So please go ahead and download the one corresponding for your platform, and unzip the package in your PC.

You can now launch the browser, please create an account from the Authenticator. You can enter any string when you are requested for the “Invitation token”.

After you finished creating your account, please keep the browser open and logged in your account before proceeding with next steps.

Create basic skeleton

We first clone the repo which contains the boilerplate using git onto a local folder named safe_examples:

$ git clone https://github.com/maidsafe/safe_examples safe_examples

And then install its dependencies:

$ cd safe_examples/safe_app_electron_quick_start
$ npm install

At this point we have an Electron application ready to be launched, let's run it:

$ npm start

You should see a "Hello SAFE Network!" message in our app's window and an empty list of trips. We are now ready to start creating the code to be able to store the planned trips into the SAFE Network.

Import the SAFE API

The application will interact with the SAFE Network using the safe-node-app package, we therefore need to add it as a dependency in our package.

As mentioned above in the pre-requisites section, it's recommended to use the browser built for mock routing for this tutorial, therefore we need to make sure to install the SAFE libraries required to connect to mock routing as well. We do this by setting the NODE_ENV environment variable specifically during installation of the safe-node-app package:

$ NODE_ENV=dev npm install @maidsafe/safe-node-app --save

Windows users in Command Prompt, will first need to run set NODE_ENV=dev, then run npm install @maidsafe/safe-node-app --save.
If using Windows PowerShell, run $env:NODE_ENV = "dev".
Note that this environment variable will only persist in your current terminal until it is closed.

Any interaction with the SAFE Network is made thru the API imported from the safe-node-app package, we do this by adding a require statement at the top of the safenetwork.js file:

const safeNodeApp = require('@maidsafe/safe-node-app');

Send authorisation request to the Authenticator

A SAFE application needs to get an authorisation from the user before being able to connect to the network, this is achieved by sending an authorisation request to the Authenticator.

We first need to generate a SAFEApp instance by calling the initialiseApp function of the API, providing information about the application (this information is displayed to the user when requesting the authorisation):

const appInfo = {
  name: 'Hello SAFE Network',
  id: 'net.maidsafe.tutorials.nodejs',
  version: '0.1.0',
  vendor: 'MaidSafe.net Ltd.',
  bundle: 'com.github.electron',
  customExecPath
};

const opts = {
  forceUseMock: true
};

let safeApp = await safeNodeApp.initialiseApp(appInfo, null, opts);

Notice the use of forceUseMock, which signals to our application that we want to use the mock routing binaries that were additionally installed with @maidsafe/safe-node-app.

We are using await to call the initialiseApp function since it's asynchronous as most of the functions exposed by the safe-app-nodejs API. You can also use JavaScript Promises if you prefer.

When the Authenticator sends an authorisation response using the system URI mechanism, the system will match the URI scheme to our application, and it will launch our application passing the URI as an argument. Since we are running our application using the electron executable rather than with our app's executable (which we are not generating in this tutorial), we need make sure the system registers the URI scheme with the electron executable's path, and this is what we do by passing the customExecPath attribute as part of the appInfo object. Please note this variable is defined some lines above as part of the boilerplate's code, so make sure you keep the following code:

const app = require('electron').remote.app;

const customExecPath = [process.execPath, app.getAppPath()];

For OSX we need a bundle that matches the electron app (as we're not registering a packaged app), so we should add the bundle com.github.electron to the app info, otherwise the registered protocol won't function for this OS.

Once the safeApp instance is initialised, we can use it to generate the authorisation request and send it to the Authenticator:

const authUri = await safeApp.auth.genAuthUri({});
await safeApp.auth.openUri(authUri);

Note that we are passing an empty object as argument to the genAuthUri function, this object can contain a list of permissions to access different containers, but for the sake of simplicity we are not doing it here with this application.

Let's make all the code for these steps to be the body of the function called sendAuthReq in the safenetwork.js file, it should now look like this:

const safeNodeApp = require('@maidsafe/safe-node-app');
const app = require('electron').remote.app;

const customExecPath = [process.execPath, app.getAppPath()];
let safeApp;

async function sendAuthRequest() {
  console.log('Authorising SAFE application...');
  const appInfo = {
    // User-facing name of our app. It will be shown
    // in the Authenticator user's interface.
    name: 'Hello SAFE Network',
    // This is a unique ID of our app
    id: 'net.maidsafe.tutorials.desktop-app',
    version: '0.1.0',
    vendor: 'MaidSafe.net Ltd.',
    bundle: 'com.github.electron',
    customExecPath
  };

  const opts = {
    forceUseMock: true
  };

  safeApp = await safeNodeApp.initialiseApp(appInfo, null, opts);
  const authUri = await safeApp.auth.genAuthUri({});
  await safeApp.auth.openUri(authUri);
}
...

As you can see, we declare the safeApp variable outside the function since we will be using this same SAFEApp instance to access the API from other functions.

The sendAuthReq function is invoked when the application's window is loaded, this is part of the code we inherited with the boilerplate. You can look at the code in controller.js if you are interested in it.

We can now launch our application again to verify that now it's able to generate the authorisation request and send it to the Authenticator. Make sure you launch the SAFE Browser and log in using the Authenticator (see the pre-requisites section) before running the application:

$ npm start

You should have had the Authenticator to show a pop-up with the authorisation request with the information of the application.

At this point you can allow or deny the request, and the Authenticator will send the response back to the application. The boilerplate code already contains the code to receive the response, the uponAuthResponse function in the safenetowrk.js file will be invoked with the authorisation URI the Authenticator sent:

async function uponAuthResponse(resAuthUri) {
}

Connecting to the SAFE Network

We can now use the authorisation URI we received from the Authenticator to connect to the SAFE Network. In order to do this we simply call the loginFromUri API function:

await safeApp.auth.loginFromUri(resAuthUri);

This function will decode the authorisation URI and create a connection with the SAFE Network using the credentials obtained from it.

Create a public MutableData

One of the native data types of the SAFE Network is the MutableData. A MutableData is a key-value store which can be created at either a specific address on the network, or just at a random address, and it can be publicly available (a public MutableData) or otherwise have all its content encrypted (private MutableData). It also has a type associated to it (type tag) which is a number that can be chosen at the moment of creating the MutableData.

We are not going to go into the other aspects of the MutableData here, we will just create MutableData in the network to store the data of our application. Please refer to the Discover page to learn more about the MutableData type as well as the other types of data available in the SAFE Network.

In this tutorial we are going to create a public MutableData at a random address. Each piece of data stored on the network has its own unique 256 bits address in the network (you can read more about XOR addresses of the SAFE Network in the MaidSafe's blog), we will request the API to generate a random address for our new public MutableData:

const typeTag = 15000;
const md = await safeApp.mutableData.newRandomPublic(typeTag);

The type tag we are choosing is just a random number here, although you must know there is a range of reserved numbers for the type tags, any MutableData stored with any of this reserved type tags will have a special treatment by the network.

At this point we have a MutableData object which was not committed to the network yet, so we can now request the API to send the corresponding request to the SAFE Network to store it:

const initialData = {
  "random_key_1": JSON.stringify({
      text: 'Scotland to try Scotch whisky',
      made: false
    }),
  "random_key_2": JSON.stringify({
      text: 'Patagonia before I\'m too old',
      made: false
    })
};
await md.quickSetup(initialData);

We use the quickSetup function which allows us to (optionally) provide an initial set of key-value entries that the MutableData shall be populated with when storing it on the network.

As you can see the key-value pairs we are storing as the initial set of data contain a "random" key, and a serialised object as the value. The application will use the key to identify each of the trips (e.g. when it needs to remove any of them), and it will also generate a random key for new trips added to th list. The value is expected to contain the information about the trip itself, we are storing the description of the trip (called text) and a boolean value (called made) which indicates if the trip has been made by the user. Since the entry's value in a MutableData needs to be a string of bytes, we serialise with JSON.stringify before storing it.

Let's have our app to create this MutableData right after it connects to the network:

let md;

async function uponAuthResponse(resAuthUri) {
  console.log("Authorisation response received");

  await safeApp.auth.loginFromUri(resAuthUri);
  console.log("Application connected to the network");

  const typeTag = 15000;
  md = await safeApp.mutableData.newRandomPublic(typeTag);
  const initialData = {
    "random_key_1": JSON.stringify({
        text: 'Scotland to try Scotch whisky',
        made: false
      }),
    "random_key_2": JSON.stringify({
        text: 'Patagonia before I\'m too old',
        made: false
      })
  };
  await md.quickSetup(initialData);
}

We are declaring the md variable outside the function so we can then access this same MutableData from other functions.

If we run the application again now, it should successfully connect to the network after an authorisation was given, and it will create a random MutableData with some initial data, although we won't see it on the UI yet, so let's now add the code to retrieve the values from the MutableData to render it in the UI.

Read the MutableData entries

We now have our MutableData stored on the network with an initial set of key-value entries, thus we can now retrieve them using the md variable we declared global. Let's create the body for getItems function in our safenetwork.js file:

async function getItems() {
  const entries = await md.getEntries();
  let entriesList = await entries.listEntries();
  let items = [];
  entriesList.forEach((entry) => {
    const value = entry.value;
    if (value.buf.length == 0) return;
    const parsedValue = JSON.parse(value.buf);
    items.push({ key: entry.key, value: parsedValue, version: value.version });
  });
  return items;
};

Note we are expecting the value of the entry to be a serialised JSON object, since that's how we stored them when we called the quickSetup function before, so we need to de-serialise it with JSON.parse before returning it.

We can now run it again and we should be able to see the list of trips we initially stored on the SAFE Network.

Add more entries to our MutableData

It's time now to allow the user to add new trips to the list by entering them in the form on the UI.

The MutableData's key-value entries can be mutated by creating a mutation transaction object where we set all the mutation actions we want to apply to the entries:

const mutations = await safeApp.mutableData.newMutation();

We can set three different type of mutation actions on the mutation transaction: insert, update, or remove. Let's go ahead and add an "insert" action to add the new entry:

await mutations.insert(key, JSON.stringify(value));

Note we are expecting the value argument to be a JSON object, so we serialise it with JSON.stringify before adding it to the mutation transaction.

Now we just need to apply this mutation transaction to our MutableData by calling the applyEntriesMutation method, let's put all this code into the body of the insertItem function:

async function insertItem(key, value) {
  const mutations = await safeApp.mutableData.newMutation();
  await mutations.insert(key, JSON.stringify(value));
  await md.applyEntriesMutation(mutations);
};

Let's now run the application and try to add a new trip to the list, the application will insert it in the MutableData and the list will be automatically refreshed on the UI afterwards.

Note that since we are creating a MutableData at a random location each time the application loads, any new items/trips the user inserts won't be displayed after restarting the application. This can obviously be changed by storing the MutableData at a custom location that can be found each time the application loads. We are leaving this out of the scope of this tutorial for the sake of simplicity.

Update and remove entries

As we saw above, to update or remove entries we just need to create a mutation transaction, with "update" and/or "remove" actions, and apply the mutations to the MutableData. Let's fill up the body of the updateItem and deleteItems functions to respectively perform these mutations on our MutableData:

async function updateItem(key, value, version) {
  const mutations = await safeApp.mutableData.newMutation();
  await mutations.update(key, JSON.stringify(value), version + 1);
  await md.applyEntriesMutation(mutations);
};

async function deleteItems(items) {
  const mutations = await safeApp.mutableData.newMutation();
  items.forEach(async (item) => {
    await mutations.delete(item.key, item.version + 1);
  });
  await md.applyEntriesMutation(mutations);
};

Note that there is a versioning involved in such mutations, as opposed to "insert" mutations which don't have any. Each entry in a MutableData has a numeric version associated to it. When you insert a new entry it's inserted with version 0, and every time a mutation is performed you need to specify the subsequent version the entry is being bumped to. This is used by the network to ensure that only one mutation is applied when simultaneous mutations requests were received for the same version of an entry, making sure the state change of such an entry is effectively what the originator of a request was intending to do.

Also bear in mind that when you remove and entry it is never deleted from the MutableData, but its value is just cleared, so you cannot insert a new entry with same key but update it. This is the reason that in our implementation of the getItems function we filter the deleted entries with the following condition when iterating thru them:

if (value.buf.length == 0) return;

The deleteItems is invoked when the user selects some of the trips from the list and then clicks on "remove trips already made". As you can see we receive a list of items to be removed and we are able to add a "remove" action for each of them into the mutation transaction before we actually send the mutation request to the network when invoking applyEntriesMutation. This is to reduce the network traffic needed to perform several mutations on a single MutableData.

Note that the boilerplate code doesn't have the implementation in the UI to be able to update trips, but we jut added the implementation for updating the items on the MutableData entries, so go ahead and try to add the UI components to allow the user to do this ;)