process_flush_classes makes bioportal unresponsive when it deletes large ontologies

[alexskr]

process_flush_classes runs every Tuesday at 10PM PST to purge archived ontology submissions and it usually takes Bioportal down until class graphs are deleted. Apparently, 4store completely blocks all subsequent queries until the delete class graph query is complete.

This is problematic when large frequently updated ontologies such as GO or DDPHENO are purged since each graph can take 5 minutes to delete and there are usually several submissions that are deleted one after another and site can be down for more than 1 hour