Per @korikuzma was initially in place because of some tricky DrugBank data, but now that we're using a different distribution of the data, it might not be an issue. Maybe either remove the limit entirely or just make it accord with whatever ends up being the 400kb size limit set by Dynamo.