This is a Node.js application used as a one-stop location to view the status of the NuPIC open source project.
You must have Github credentials for API calls set as environment variables.
export GH_USERNAME=<username>
export GH_PASSWORD=<password>
The server folder contains the JavaScript code that runs on the Node.js server. The client folder contains the JavaScript code that runs in the browser.
A monitor is a component that has a client component in client/js/monitors. If it needs it, it can also have a server component in server/monitors.
A monitor only needs a client component to be included on the status page. Let's walk through making a monitor called foo that has a client component that prints something to the status page.
client/js/monitors/foo.js
$(function() {
function initialize(id, config, server, template) {
template({
title: 'Foo monitor'
, message: 'Hello, world!'
});
};
window.WB.foo = initialize;
});
The client runtime expects your monitor to attach a function to the global WB namespace. This function is an initializer that will be called with a monitor id, the configuration of the runtime, an interface to interact with your monitor's server component (more on that later), and a template to populate and render data to the webpage. To render data to the screen, you simple have to call the template function with a data object that matches a template you've defined below:
client/js/monitors/foo.html
<h3>{{title}}</h3>
<p>{{message}}</p>
To include your monitor on the page, you must add it to conf/default.yaml in the monitors section. The key you use will be the id of the monitor, and you must specify a type, which in this case would be foo (matching the filename of the monitor in client/js/monitors). You may also specify options here, where you put any data you like. These options will be passed to your client monitor initializer (detailed above) as the config parameter.
configuration for foo monitor
monitors:
foo_1:
type: "foo"
options:
greeting_repeat: 4
Now you can use this greeting_repeat option to repeat your message:
client/js/monitors/foo.js
$(function() {
function initialize(id, config, server, template) {
var greetingRepeat = config.greeting_repeat
, initialMessage = 'Hello, world!'
, finalMessage = _.times(greetingRepeat, function() {
return initialMessage;
}).join('<br/>\n');
template({
title: 'Foo monitor'
, message: finalMessage
});
};
window.WB.foo = initialize;
});
You can place your monitor on the status page by adding an HTML element with the same id you specified in the configuration in client/index.html. You can add it wherever you like in the <body>.
<div id="foo_1"></div>
If you don't put this in the index file, it will be appended to the bottom of the body.
If you need to make an HTTP call that cannot be done from the browser, there is a simple HTTP proxying utility available to all client monitors through the global WB.utils.proxyHTTP function. Use it like this:
WB.utils.proxyHttp("http://example.com/whatever", function(err, response) {
});
Errors are contained in the first callback parameter, and the response body is contained in the 2nd.
There are only so many things you can do with only a client component to your monitor. You might need to take advantage of server-side APIs that can't be done from the browser. In that case, you'll need to add a server component that your client can interact with. This is simple. Add a file called server/monitors/foo.js.
Docs coming soon, but in the meantime, you can see how this works with the
travis_buildsortravis_latestmonitors.
npm install .
npm start
Then open http://localhost:8080.
- recent commits on github
- current contributor count
- mailing list statistics
- green / yellow / red state of doc build based upon date