`equivalence-nodes` query is very slow, https://github.com/histograph/stats/blob/master/queries/equivalence-nodes.cypher prob because I compute `:=` node size, could store this while constructing https://github.com/graphmalizer/graphmalizer-core/blob/master/queries.yaml#L53