Browser-based accessibility testing tools and plugins require manually testing each page, one at a time. Tools that can crawl a website can only scan pages that do not require login credentials, and that are not behind a firewall. Instead of developing, testing, and using a separate accessibility test suite, you can now integrate accessibility testing into your existing automation test suite using AATT.
AATT tests web applications regarding conformance to the Web Content Accessibility Guidelines (WCAG) 2.0. Find a list of the WCAG 2.0 rules checked by HTMLCS Engine on the HTML CodeSniffer WCAG Standard Summary page and Chrome Engine on the Google Chrome Developer Audit rules. AATT provides an accessibility API and custom web application for HTML CodeSniffer, Axe and Chrome developer tool. Using the AATT web application, you can configure test server configurations inside the firewall, and test individual pages.
AATT includes HTML CodeSniffer, Axe and Chrome developer tool with Express and PhantomJS, which runs on Node.
For example, it can be used to test Java web applications using SeLion automation test frameworks.
For node applications, it can be integrated into NemoJS testing framework to run accessibility testing during automated unit testing .For Nemo framework use [Nemo-Accessibility plugin] (https://github.com/paypal/nemo-accessibility)
Have docker installed
$ docker run -d esteinborn/docker-aatt
You can now access the running instance of AATT from http://{{dockerip}}:3000
AATT provides an API for evaluating HTML Source code from other servers. The API EndPoint is: https://{{dockerip}}:3000/evaluate
-
Accepts the following parameters:
- "source" to send the HTML source of the page. Can be a whole page or partial page source
- "engine" E.g. engine=htmlcs. This is the engine which will scan the code. It accepts a single value of "axe", chrome" or "htmlcs".
- "ouput" to get the jsonified string. E.g. output=json. If this parameter is not set or left empty, it will return a string with table data that can be parsed or appended directly into your page. Default to "htmlcs"
- "errLevel" Error level like Error, Warning or Notices . Mapped to 1, 2 and 3 respectively. E.g. "1,2,3"
- "level" This option applies only for the default htmlcs evaluation engine. Options can be either of the following WCAG2AA, WCAG2A, WCAG2AAA, Section508 . Defaults to "WCAG2A"
-
Set the Request Header Content-type as application/x-www-form-urlencoded
Here is a sample ajax script which would initiate the request:
var xmlhttp = new XMLHttpRequest();
xmlhttp.open("POST","http://{{dockerip}}:3000/evaluate",true);
xmlhttp.setRequestHeader("Content-type","application/x-www-form-urlencoded");
xmlhttp.send("source=" + document.getElementById('source').value + "&priority=" + document.getElementById('priority').value);
Nemo is a node.js based automation framework for browser automation. It's plugin-architecture helps switch on/off different capabilities. The nemo-accessibility plugin performs accessibility scanning while running browser automation using Nemo framework.
nemo-accessibility
plugin uses the AATT accessibility API to evaluate HTML source. Therefore you must specify the API url under as a plugin argument like below.
"nemo-accessibility":{
"module":"nemo-accessibility",
"arguments": ["https://{{dockerip}}:3000/evaluate"]
}
Once nemo-accessibility
plugin is registered, you should now have nemo.accessibility
namespace available in your tests. nemo.accessibility
exposes a method called scan
to help you run accessibility evaluation against your page/element. scan
method takes an optional object like below,
var options = {
'element': driver.findElement(wd.tagName('iframe')), //default is entire page
'engine': 'axe', 'htmlcs' or 'chrome', //default is htmlcs
'output': 'html' or 'json', //default is html,
'level': 'WCAG2AA' or 'WCAG2A' or 'WCAG2AAA' //option applies to htmlcs only and default to WCAG2AA
'errLevel' : '1,2,3' // for htmlcs only, 1 means Error, 2 means Warning, 3 means Notice default:1,2,3
}
scan
method returns a promise with resulting HTML
or json
response from AATT api when fulfilled. You can then write the HTML to a file or parse JSON response for later reporting. For example,
nemo.driver.get('http://www.yahoo.com');
nemo.accessibility.scan().then(function (result) {
fs.writeFile('report/accessibilityResult.html', result, function (err) {
done();
});
});
You could also run accessibility scan on a certain element like below. This is useful when lets say you scanned an entire page already, and subsequently a certain automated test interaction opened a dialog box; you can now only scan newly opened dialog box since you already scanned the rest of the page before.
Here is a example, (note this example uses excellent nemo-view plugin for finding elements)
it('will run scan on an element', function (done) {
nemo.driver.get('http://www.paypal.com');
var options = {
'source': 'btnDonate',
'engine' : 'htmlcs',
'errLevel': '1,2,3'
};
nemo.accessibility.scan(options).then(function (result) {
fs.writeFile('report/entirePage.html',result,function (err) {
done();
});
});
var welcomePage = nemo.view.welcomePage;
welcomePage.buttonThatOpensAPopup().click();
var element = welcomePage.popup(),
options = {
'element': element,
'engine' : 'axe',
'errLevel': '1,2,3'
};
nemo.accessibility.scan(options).then(function (result) {
fs.writeFile('report/scanAnElement.html', result, function (err) {
done();
});
});
});
For more details, please refer to: nemo-accessibility plugin
Nightwatch JS is another UI automated testing framework powered by Node.js and uses the Selenium WebDriver API. To call AATT, you need to use the request module. NightwatchJs has call back functions like before and after hooks that would be called before or after executing a test case. Request to AATT API should be done in after hook passing the source code of the page to the API. Here is an example commit on how to do this with Nightwatch.
The AATT web application can be used to test HTML code snippets or pages. To test logged in pages on a test server, first configure the login credentials in AATT which creates a cookied experience. Then, enter the url for the page you want to test. Results are displayed as a table that can be exported as a CSV file. Results include:
- WCAG 2.0 principle: Perceivable, Operable, Understandable or Robust
- Error description
- code snippet
- WCAG 2.0 techniques to help developers fix the issue.
Optionally, you can configure the tool to save a screensnap of the html page you test. You can also configure the tool to display only errors, or also to include warnings and notices. Warning and notices require manual inspection to determine the severity of the warning or notice.
Copyright 2016, PayPal under the BSD license.
- Prem Nawaz Khan, developer || https://github.com/mpnkhan || @mpnkhan
- Cathy O'Connor, design || @cagocon
- Nilesh Kulkarni , developer for nemo accessibility plugin ||https://github.com/nikulkarni
- Srinivasu Chakravarthula, user interaction, testing || @csrinivasu
We welcome your feedback. Please file issues and/or enhancement requests.