It resulted from changing
the exception type from the SQLAlchemy side
in the latest 1.4.45 version.

several pipelines were writing to the same base at the same time and
sometimes database were not deleted after tests so it could not be
created in future tests

Done required refactoring - collecting was moved to a separate .sh file

logger setup moved to logger.py

Mateusz Żebrak authored Feb 08, 2022 and Mateusz Żebrak committed Feb 25, 2022

Refactor manual.insertion.py that the pylint passes.

Mateusz Żebrak authored Feb 08, 2022 and Mateusz Żebrak committed Feb 25, 2022

If the --zscore-limit argument is not provided - the limit value will be
fetched from database. Each testcase has a separately determined value,
which currently defaults to 1.

Additionaly, constants from test_common.py are moved to constants.py.

Mateusz Żebrak authored Feb 08, 2022 and Mateusz Żebrak committed Feb 25, 2022

Downloading images without bearer token fixed

Mateusz Żebrak authored Feb 08, 2022 and Mateusz Żebrak committed Feb 25, 2022

HAF replay produces same data as HIVED replay, so we need to distinguish
them. We re doing this by different description stored in `description`
column in the `benchmark_description` table.

To calculate z_score and show correct bars on the grafana chart - we
need to pass `--desc-like` and `--exec-env-desc-like` arguments.