* Usage of pytest-benchmarks is dropped since it call tests
  sequentially which very long time to complete, instead original
  tavern tests are used
* Time measuring is done via durations switch to pytest and tavern
* During benchmarking validate_response.py is disabled using nevly
  introduced env variable
* Each run creates xml junit file with time data
* After all runs data from files are combined and report file is
  generated
* Tests above threshold are marked with red

* Reverted hardcoded block number and changed to use CI var HIVEMIND_MAX_BLOCK.
* Please add a possibility to set HIVEMIND_MAX_BLOCK.
* operation_extractor:
** name of parameter `hivemind_address` changed to `hived_url` as we
   scan blocks from `hived`
** added description to the script

* New script for running benchmarks
* Benchmarks are now runned several times to cheat caching
* Data from each run is combined together

Do not allow benchmark tests to fail, failed test is a failed job. Changed html file names to avoid confusion

* added script to process xml reports
* script will parse xml reports and create html reports with api
  execution times
* if execution time is greater than treshold it will be marked on red
  and task will fail

[ABW]: list_comments with by_root, by_parent, by_last_update and by_author_last_update have their main parameters required (nonempty)
list_comments_by_cashout_time requires valid start post when not completely skipped
[Fix] list_comments_by_cashout_time and list_comments_by_permlink would in some cases select special null post in inner query giving one post below limit
[Fix] list_comments_by_last_update incorrectly used found parent author id (matching it with post id)
tests rebound to proper version (above and earlier changes required a lot of changes in database api tests)

[ABW]: tests rebound to newest master; also removed old CI group in bridge_api since all test are now separate

[ABW]: tests rebound to version that has no more condenser_api tests in old scheme (with common .yaml)

changed behaviour of some condenser_api calls: when start_author or start_permlink is present, they need to point to correct post or error is generated
fixed crash when start_author or start_permlink is passed to some condenser_api calls

added new test groups to CI with at least single test to make it easier to move tests accourding to new scheme (only follow/tags have no negative tests so no such groups were added)

[ABW]: added separate tests to CI (all tests will eventually be run through folders instead of common .yaml file)

- removed tests_require from setup.py it is depreciated in favor of tox
- setup test env in tox
- removed gitpython dependency