Create new tables with JSONB columns to store statistics.
One table each for user, artist, release and recordings. All
these tables are inside a new schema `statistics`

Also minor changes in `db/testing.py` to bring the SQL there
out of the code and into the `admin/sql` directory.

First drop schemas with cascade and then drop the rest of the
tables

Also, some refactoring to make the `update_last_login` function
use PostgreSQL's NOW() method, considering we never change
the `last_login` value to anything else.

First run an SQL query to update the last login value to
unix epoch. Then call the function and assert that the
value now is greater than 0.

LB-167: Create schema for storing statistics in PostgreSQL

The stats module is gonna be similar to the db module. All the
queries we make to BigQuery should be contained here.

run_query takes a query and its parameters and then returns the
results of the that query.

get_top_tracks is an example query that gets the top tracks of
the user over a particular interval of time.

Also, add a non-parametrized query for getting the sitewide
total artist count to check if it works.

Add release_mbid and recording_mbid to docstrings

LB-176: Create a stats module and add functions to run queries on Google BigQuery

LB-187: Add NOT NULL constraint to the last_login field