Add support for PostgreSQL

[Gerrrr]

Thank you for opensourcing hunter!

This PR adds support to using PostgreSQL as a data source for benchmark results. See the usage example below.

Configuration:

postgres:
  hostname: ${POSTGRES_HOST}
  port: ${POSTGRES_PORT}
  username: ${POSTGRES_USERNAME}
  password: ${POSTGRES_PASSWORD}
  database: ${POSTGRES_DATABASED}

templates:
  common:
    type: postgres
    time_column: ts
    attributes: [commit]
    metrics:
    ...

Test:

  # trunk
  trunk.aggregate_mem:
    inherit: [common]
    query: |
      SELECT e.*
      FROM experiments e
      INNER JOIN setups s ON e.setup_id = s.id
      WHERE e.exclude_from_analysis = false AND
            e.branch = 'trunk' AND
            s.benchmark = 'aggregate' AND
            s.scenario = 'singleNode' AND
            s.suppress = 'NONE' AND
            s.store = 'MEM' AND
            s.cache = true AND
            s.log = true AND
            s.sink = false AND
            s.keyspace = 10000 AND
            s.key_distribution = 'normal' AND
            s.partitions = 2 AND
            s.instance_type = 'ec2i3.large'
      ORDER BY e.ts ASC;

[pkolaczk]

Thank you for the contribution. Can you add more documentation on usage and about the expected database schema?

[Gerrrr]

Hey @pkolaczk @smccarthy788 ,

I've updated this branch. It now includes requested documentation as well as PostgresExporter that completes psql support.

Can you please take another look at this PR?