new mocks for community tests, list_pop_communities fix

See merge request !466

Merge follow data from given block range on python side, not sql

See merge request !396

added mock data and description for detailed follow_op tests

added debug parameter for bridge.get_relationship_between_accounts to make it possible to detect existence of hive_follow records even if they contain default data
added logging for invalid cases of follow_op

temp disable bench fail

See merge request !471

Change xdist lib to parallel

See merge request !402

Deadlock fix

See merge request !468

mentions unique constraint upgrade fix

See merge request !467

Fixed upgrade definition of hive_mentions_ux1 constraint to match original one specified in schema.py

[ABW]: changed the way follow list resetting works (closer to previous solution but with separate SQL functions) since there should not be too many such operations
replaced constants with enum values for clarity
eliminated unnecessary subqueries (replaced with outer join) and internal order clause, simplified query loop
checks during validation reordered to exit earlier in case of fatal error
[Note] whole follow code most likely needs more test scenarios - to be added

We need a method to also touch all data for given follower present
in the follow_items_to_flush in case when we have following scenario:
 1. data for mass action exists in DB
 2. data for mass action exists in follow_items_to_flush
 3. mass action is received when 1 and 2 or 2 are true

Issue 116

See merge request !441

mocks changed so each scenario uses separate post (last number in post permlink shows which scenario) - necessary so tests don't collide and/or overshadow potential problems

Added threads into post synchronization

See merge request !446

relinked to current tests_api master

See merge request !465

[ABW]: mostly tabs replaced with spaces; no actual code changes; relinked to current tests_api master