Extract email preferences into contact_rollups_raw

[bencodeorg]

Basic logic to extract email preferences into contact_rollups_raw by executing bin/cron/build_contact_rollups_v2.

Testing story

Tested locally and on a newly created ad hoc connected to a production clone. email_preferences has close to a million rows in production -- some interesting findings:

• using truncate in raw SQL (in this PR) instead of delete_all in ActiveRecord (not in this PR) takes .3 seconds instead of 14 seconds.
• changing find_in_batches batch size from 1000 to 10000 has minimal impact on execution time (actually was 7 seconds slower, 337 vs 330)
• using a raw SQL query for inserts (not in this PR) instead of batched active record queries (in this PR) significantly reduces run time (17 seconds vs. 5.5 minutes) -- here is SQL query that was used:

  insert into contact_rollups_raw(email, sources, data, data_updated_at, created_at, updated_at)
  select email, 'dashboard.email_preferences', json_object('opt_in', opt_in), updated_at, now(), now() 
  from email_preferences

General finding is raw SQL dramatically outperforming ActiveRecord.

Reviewer Checklist:

• Tests provide adequate coverage
• Code is well-commented
• New features are translatable or updates will not break translations
• Relevant documentation has been added or updated
• User impact is well-understood and desirable
• Pull Request is labeled appropriately
• Follow-up work items (including potential tech debt) are tracked and linked

[hacodeorg]

You can add contact-rollups label to this PR. It will be easier to search for all related PRs using this link

[hacodeorg]

You also want to do ContactRollupsRaw.delete_all to assure the table is empty before inserting to it.

[bencodeorg]

Isn't that an assumption of our test framework (cleanup/teardown of database before/after every test)?

[hacodeorg]

Good question. I don't know what is our current test cleaning strategy and how it behaves in local dev machine, in Drone and in test machine. Would be very helpful for me and the team if you could do some research and share what you find.

[hacodeorg]

LGTM. Thanks, Ben!