With the nested bucket aggregations, I am seeing a discrepancy between the facet count and search results. See https://github.com/CDRH/habeascorpus/issues/108. This seems to be related to how Orchid is constructing API queries, rather than an issue with individual projects