optimize total values for inscriptions endpoint

[rafaelcr]

When calling /ordinals/inscriptions, the "total" JSON key returns the number of results for the given filter set. When this number is very large, the SQL query will take a very long time to finish because Postgres has to count every single row before returning the total value.

To optimize this, we should keep a pre-calculated count of common filters such as mime_type for the most popular types. We currently only do this kind of caching for unfiltered queries (and it made a huge difference in performance).

[rafaelcr]

A fix for this issue will also help with errors such as

2023-04-30 19:30:07.220 GMT [1207312] ERROR:  could not write to file "base/pgsql_tmp/pgsql_tmp1207312.0": No space left on device

which occur when postgres needs to allocate huge amounts of temporary space in order to count affected rows

[rafaelcr]

Closed by #55