Are you still scraping or did you migrate to their API?
I finally reworked my code to just download the html files using phantom.js and parse it with perl so that I can run this on a headless server, and this method hasn't broken lately. I find trying to use phantom.js is kind of crazy (mainly I have a hard time with the documentation) and javascript is not my native language. I was using curl, LWP, and wget in the past.
With regards to the anti-bot problem, this isn't running on any server, it's running on the user's computer. The number of requests they make per second doesn't seem to trigger any anti-bot measures. I'm just using plain Java URLConnections.