feat: rule aria-hidden-focus

[jeeyyy]

Rule: Aria Hidden Focus.
Spec for rule - Spec: auto-wcag.github.io/auto-wcag/rules/SC4-1-2-aria-hidden-focus.html

Closes issue:

• Rule: SC4-1-2-aria-hidden-focus (PR#1166) #1150

Reviewer checks

Required fields, to be filled out by PR reviewer(s)

• Follows the commit message policy, appropriate for next version
• Has documentation updated, a DU ticket, or requires no documentation change
• Includes new tests, or was unnecessary
• Code is reviewed for security by: @WilcoFiers

[jeeyyy]

On hold, until issue #1208 is resolved.

[marcysutton]

I'm so excited for this rule!!!

Updated. Review again.

Updated. Please review again.

[jeeyyy]

Performance analysis

Page	Measure rule_aria-hidden-focus took
http://qateam.dequecloud.com/attest/p-as-heading.html (Aria Hidden on Body)	22.399999999947795ms
http://qateam.dequecloud.com/attest/skiplink_first.html (Aria Hidden on Body)	41.69999999999163ms
https://fontawesome.com/v4.7.0/cheatsheet/	239.00000000003274ms
https://fontawesome.com/v4.7.0/cheatsheet/ (Aria Hidden on Body)	256.0000000000855ms
https://www.tensorflow.org/api_docs/python/tf/InteractiveSession	161.7999999998574ms
BIG PAGE https://www.tensorflow.org/api_docs/python/tf/InteractiveSession (Aria Hidden on Body)	1056.0999999997875ms

[jeeyyy]

@dylanb @WilcoFiers updated with perf measurements.

[WilcoFiers]

See if you can get this to 500ms for the large page.

[dylanb]

@JKODU I think the performance impact on large pages is significant enough that its worth refactoring to try to achieve performance improvements. I would like to see what the theoretical minimum would be (if we used all native code for selecting, did one pass only for the entire rule instead of two and did the filtering at the front instead of the back.

One more question, how did you measure those performance numbers?

[jeeyyy]

One more question, how did you measure those performance numbers?

@dylanb

• I took a snapshot of each of those pages (view source > save).
• I added axe.js to those pages and executed axe.run with ruleOnly > aria-hidden-focus and setting the performanceTimer flag to true.
• In some cases, I set aria-hidden=true on the body of the page for test reasons.

The numbers reported above are an average of 3 run's for each page, that was measured/ reported on the console for each of those pages.

[jeeyyy]

@JKODU I think the performance impact on large pages is significant enough that its worth refactoring to try to achieve performance improvements. I would like to see what the theoretical minimum would be (if we used all native code for selecting, did one pass only for the entire rule instead of two and did the filtering at the front instead of the back.

@dylanb - Sure, I will attempt some refactoring as suggested.

[jeeyyy]

For BIG PAGE https://www.tensorflow.org/api_docs/python/tf/InteractiveSession with aria-hidden=true, caching isFocusable on the virtualNode improved the run time of rule aria-hidden-focus to circa ~850ms - Measure runchecks_aria-hidden-focus took 848.2999999980093ms, and caching tabbableElements on the virtualNode improved the run time of rule aria-hidden-focus to circa ~550ms - Measure runchecks_aria-hidden-focus took 554.5000000020082ms.

I have cached isFocusable and tabbableElements on the virtualNode.

Also, had to update certain tests involving virtualNode as above functions were introduced in the object structure, and .deepEqual assertion fails when checking for referencing equality, see - chaijs/chai#896

Performance analysis (after caching results)

Page	Measure rule_aria-hidden-focus took	Previous Measure (non-caching)
http://qateam.dequecloud.com/attest/p-as-heading.html (Aria Hidden on Body)	4.300000000512227ms	22.399999999947795ms
http://qateam.dequecloud.com/attest/skiplink_first.html (Aria Hidden on Body)	3.6000000036437996ms	41.69999999999163ms
https://fontawesome.com/v4.7.0/cheatsheet/	84.60000000079162ms	239.00000000003274ms
https://fontawesome.com/v4.7.0/cheatsheet/ (Aria Hidden on Body)	91.8999999994412ms	256.0000000000855ms
https://www.tensorflow.org/api_docs/python/tf/InteractiveSession	40.80000000249129ms	161.7999999998574ms
BIG PAGE https://www.tensorflow.org/api_docs/python/tf/InteractiveSession (Aria Hidden on Body)	554.5000000020082ms	1056.0999999997875ms

Updated.