Parse the root from the document's URL properly

[mbland]

Given the https:// scheme used by https://18f.gsa.gov/, the '7' used in the original script was off-by-one, causing the error:

Failed to load resource: net::ERR_NAME_NOT_RESOLVED

in the JavaScript console for the computed value:

https://api/data/team.json

because the hostname was not corectly parsed. This change ensures the script works for both http:// and https:// schemes. Tested by hardcoding a value for var url with both http:// and https:// schemes for 18f.gsa.gov.

cc: @gboone @wslack @yozlet

[yozlet]

:+1: