- Conditional updates in python code was moved to SQL.

Merge branch 'dk-issue-3-concurrent-block-query-rebase' of gitlab.syncad.com:hive/hivemind into dk-issue-3-concurrent-block-query-rebase

- Merged/rebased current work of @wrona and @dkedzierski onto current
  develop,
- follow.py requires check by @wrona,
- rebase require carefull check and tests.

Note: Not tested!!! Testing in progress!!!

[JES] Add a limit on the get_account_posts call to only return 2000. There are some posts out there with thousands upon thousands of comments and those seem to freeze up the hivemind server. This should help avoid that

Gandalf authored 4 years ago and Jason Salyers committed 4 years ago

[JES] Update to the get discussion query. Do the filtering in the sub query to increase performance, drop an uneeded join for the same reason. Add a limit of 4000 to prevent a huge amount of data being returned (test post that showcased this problem was a post with over 450k children)

[JES] I somehow wound up working on an older version of the file and encountered an error this morning that made me add this line. It was correct the way it was, but the bad file I was working on threw me off.

[JES] Final round of updates for blacklisting. Return ID's insteaad of text from the DB query and have the python side do a dictionary lookup to get the name. Response times are good until following a large blacklist, and even then they're still reasonable (under 500ms)

[JES] Union all works faster than a union. Tried switching order of the join but that didn't seem to make a difference

[JES] Modifications to the follow op for the new blacklisting options. Now to figure out how to test it

[JES] Trying to restore old mutes functionality. Decided to create a new method instead of changing the existing

[JES] Testing some more updates. Hopefully have get_discussion in condenser using the new blacklists

[JES] Some refactoring to eliminate the number of times posts get looped over to finish filling out their fields

[JES] Reformat the query a bit. Figured out indexes and such. Now onwards to refactoring the processing side with the new query info