This is a browser-based client for interfacing with the New York Public Radio digital infrastructure. This Ember application renders all of www.wqxr.org using the preexisting Django server as a backend data source, consuming HTML documents and JSON payloads to present a dynamic interface. This gives us the benefits of a persistent in-browser application while still using the existing content and infrastructure.
This ember app is a client, which means its primary purpose is to provide an interface to data source. On its own, this application will not do much without some kind of source from which it can retrieve content to display. To this end, you will need to run a checkout of publisher as a server.
publisher is a Django app with its own environment and dependencies, but is configured to work in concert with wqxr-web-client. In development mode, publisher will make requests to http://localhost:4200/assets/ for Ember assets (see puppy/util/ember_config.py). So you will need to set up a separate checkout of publisher to do any substantial development work on wqxr-web-client.
Fortunately, publisher has its own set of easy-to-follow getting started instructions.
You will need the following software properly installed on your computer. Please follow all the directions as written on the websites of these projects and do not neglect to install their dependencies.
As noted above, you must spin up a checkout of publisher at a web address against which this ember app can make requests. By default, the app will use https://wqxr.demo2.wnyc.net as its destination.
This value is controlled by a .env file you will create by following the directions below.
We use modernizr to detect for certain browser features. Rather than include a full build in the app, there is a single Grunt task included in this project which scans all the .js and .scss files for mentions of the modernizr API. Step 9 below will generate a modernizr build; do not skip it or you may have mysterious errors.
$ git clone git@github.com:nypublicradio/wqxr-web-client.git && cd wqxr-web-client$ git checkout <working branch>$ cp .env.sample .env- Edit
.envwith yourpublisherapp location if you don't want to use the demo environment $ yarn$ grunt modernizr:dist
The publisher service is not set up as a strictly data-only API server and will return HTML with embedded script tags that request against the domain root. To resolve this issue, you should run your local server with the proxy command:
$ ember serve --proxy https://wqxr.demo2.wnyc.netWhen we fetch pages from the Django app, we include the X-WNYC-Ember HTTP header. This allows Django to suppress certain things (like redundant Javascript libraries) that are unnecessary when the Ember app is managing the client side.
module_wrapper.py provides a very slim module system around all the legacy Javascript files. See app/services/legacy-loader.js for more detail.
See docs/prototype-report.html for the original writeup when this application was created.
We use ember-cli-mirage to provide mock server output against which to develop. It's turned off by default in dev environments, controlled by the USE_MIRAGE envvar in your .env file. You can turn it on by setting USE_MIRAGE to true.
Logging by ember-cli-mirage is turned off by default, due to how noisy the console becomes if it's turned on. This setting can be changed in mirage/config.js. Look for this line:
this.logging = false;
Set that to true and you'll see a line for all the requests that mirage handles, but be warned: there are many.
wqxr-web-client parses publisher's HTML output for <script> and <link> tags, which it then loads according to matching and ordering rules found in app/services/legacy-loader.js. By default it will log when it has decided whether or not to load a legacy script. This setting can be turned off by commenting out the following line in config/environment.js:
ENV.LOG_LEGACY_LOADER = true;
We use ember-metrics to proxy analytics tracking to two backends, our Data Warehouse and Google Analytics. Debug output for the Data Warehouse adapter can be turned on by adding debug to the url's query string. GA debugging is handled using a chrome extension.
FYI - we hid the non WQXR streams from the schedule page with CSS in the _schedule.scss file.