At the moment these caches are simple disk caches with expectation of them being lost on reboot/redeploy of heroku dyno.
For better performance we should really try to leverage this caching in a more permanent way by building up a large database of pre-computed calculations, thus avoiding ever calculating the same thing twice.
At the moment these caches are simple disk caches with expectation of them being lost on reboot/redeploy of heroku dyno.
For better performance we should really try to leverage this caching in a more permanent way by building up a large database of pre-computed calculations, thus avoiding ever calculating the same thing twice.