WeLoop

WeLoop is a React/TypeScript client for ZenPub and a key component of ReflowOS

⚠️ We're migrating WeLoop from ZenPub to Bonfire to make it compatible with the latest version of Reflow OS 🛠️ Temporary use the following repository for the Bonfire compatible version https://github.com/reflow-project/weloop-bonfire/

About the project

This app is based on CommonsPub, a project to create a generic federated server, based on the ActivityPub and ActivityStreams web standards and originally developed as part of moodle.net.

Structure

High level folder structure:

Application source folder structure:

Please note that the project is undergoing some refactoring, and some of these may be changing...

Development Environment

In the project directory, you can run:

Before you start

We recommend using nvm to ensure you are running the right node version specified in .nvmrc. Install nvm here.

Then run nvm use every time before yarn to ensure you are using the right version.

You might also need to install yarn the first time by doing npm install --global yarn.

Customize your instance

Before we start we need to create an .env file pointing to the REFLOW OS GraphQL API: Example part of an .env file (see .env.example for the whole file):

REACT_APP_GRAPHQL_ENDPOINT=https://api.reflowproject.eu/api/graphql

We also need to add the app root end-point for the sign up email confirmation to work

REACT_APP_FRONTEND_HOSTNAME=https://your-weloop-instance.org

nvm use

Sets the right node version.

yarn

Installs the app dependencies.

yarn start

Runs the app in the development mode.

Open http://localhost:4000 to view it in the browser.

The page will reload if you make edits. You will also see any lint errors in the console.

yarn add-locale

Adds a locale for localisation, with lingui library.

yarn extract

Extracts new/updated strings from the codebase into JSON files for localisation (they need to be encapsulated with lingui library's ).

yarn compile

Compiles localisation files for production.

yarn build

Builds the app for production to the build folder. It correctly bundles React in production mode and optimizes the build for the best performance.

The build is minified and the filenames include the hashes.

Production deployment

Complete deployment of the full bundle will be described as part of the overall Reflow OS installation guide:

Self-hosted

Manually use yarn build to generate the ./build folder including the index.html file and all the static dependencies. Host the resulting website on any static web server such as NGINX or Apache Web Server.

Github Pages

WeLoop main repository uses Github Actions to automatically deploy the repository default branch as a public website using Github Pages every time changes are pushed. Anyone can fork the existing repository and deploy their own version quickly. Custom domains are defined via a ./build/CNAME following the standard Github pages workflow. Two Github Secrets can be defined to customize the app REACT_APP_GRAPHQL_ENDPOINT and REACT_APP_SENTRY_API_KEY. Learn more about them in the Customize your instance basic settings and Remote monitorings sections.

Remote monitoring

WeLoop integrates with Sentry, a cross-platform application monitoring. It allows you to remotely detect JavaScript performance issues before they become downtime. Sentry is fully open source and instances can be locally deployed together with the other Reflow OS tools.

Sentry can also be used as a SAAS from sentry.io without any installation required. To use it you need to provide a valid Sentry Key in your .env file under `REACT_APP_SENTRY_API_KEY. When deploying via Github Pages that can be set via Github Secrets.

Libraries

This section mentions notable libraries and tools used within the application and provides links to their documentation.

• React, UI framework (https://reactjs.org/docs/getting-started.html)
• TypeScript, programming language (https://www.typescriptlang.org)
• webpack, build tool (https://webpack.js.org)
• React Apollo, GraphQL client (https://www.apollographql.com/docs/react/)
• Phoenix.js, Phoenix channels JS client (https://hexdocs.pm/phoenix/js/index.html)

Browser testing

We're using BrowserStack for testing the front-end in various browsers.

Localisation

LinguiJS is the localisation library used for i18n.

Set up

• LinguiJS is configured in the .linguirc file in the root of the application.

• It comes with a provider component that sets up i18n within the application and makes components within the app able to consume the language strings of the user's chosen locale. The provider is configured in the App container (src/containers/App/App.tsx).

• The app uses React 16 Context to manage the chosen locale and maintain a state around this. The context (state) is also set up and handled within the aforementioned App container.

• Any component can "consume" the locale context by using the LocaleContext exported from the App container. This allows any component to access the API for changing the active locale. For example, the LanguageSelect component (/src/components/inputs/LanguageSelect/LanguageSelect.tsx) is wrapped in the LocaleContext.Consumer component, giving it the setLocale function:

    <LocaleContext.Consumer>
      {({ setLocale }) => (
        //...
      )})
    </LocaleContext.Consumer>

Usage

Wherever you write visible text, i.e. anything user-facing, the copy should be written using the LinguiJS components. The text used within the LinguiJS components can then be extracted using the CLI operations provided by the library, which are detailed in the Scripts section of this document.

Examples of using the LinguiJS library are given below.

Simple language strings

• First import the Trans component:

  import { Trans } from '@lingui/macro';

• Note: the Trans component is imported from the macro package, not the react package!

• Then consume the Trans component wherever text is used, like so:

  <Trans>Sign in using your social media account</Trans>

Language strings as reference

• Import the i18nMark function.

  import { i18nMark } from '@lingui/react';

• Define the language string however you like. It is usually the case that a file will contain more than one language string accessed via reference, in this case organise the strings within an object with properties that describe their purpose. For example, from the Login page:

  const tt = {
    //...
    validation: {
      email: i18nMark('The email field cannot be empty'),
      //...
    }
  }

• Note: the validation.email string is wrapped in a call to i18nMark. As the string is not passed to (as props or directly as children) to the Trans component it will not be picked up automatically by the LinguiJS extract script. In order to "mark" the string as a language string to be included in the compiled language files we must wrap it in a call to i18nMark.

• Then consume the strings. Again, for example, from the Login page:

  validation.push({
    field: ValidationField.email,
    type: ValidationType.error,
    message: tt.validation.email // <- notice the string reference here
  } as ValidationObject);

Plural language strings

LinguiJS has a Plural component, which is like the Trans component but used where the language contains pluralization.

component handles pluralization of words or phrases. Selected plural form depends on active language and value props.

The LinguiJS documentation is very comprehensive and should be referred to for usage of the Plural component:

https://lingui.js.org/ref/react.html#plural

Interpolated language string

It is very common to interpolate values into language strings. This can be done using the Trans and Plural components, where the interpolated string names are denoted with curly braces (but still within the actual string) and the component is given a key/value hash via a values prop, where a key of the hash is the name of a string to be interpolated. For example, from the Login page:

<Trans
  id="You don't need an account to browse {site_name}."
  values={{ site_name: 'MoodleNet' }}
/>

It is possible then to have site_name or any other interpolated string value produced dynamically and inserted during runtime. If interpolated values also require localisation then you would use a language string hash, as above in Language strings as reference, making sure to use the i18nMark function to mark them for extraction by the LinguiJS CLI.

Updating language files

Whenever updates are made to any language within the application you must run the LinguiJS extract script. This will pull out all the language strings and add or update them in the specific locale messages files, which live in locales.

All changes to the language within the application, including changes to the files within locales, should be committed alongside other changes.

Licensing

The present software is released under a copyleft license (GNU Affero General Public License v3.0) to promote use, customization and contribution inside and outside the consortium. The license conditions developers to make available the complete source code and any modifications. It fosters an open ecosystem and protects rights users by granting them access to the source code.