Show the actual configuration settings used in the collect mode on html report
mtakahar opened this issue · 1 comments
What's printed under the "Configuration" section of html report currently comes from the config specified in the "report" mode of the runner.py script instead of the settings actually used in the "collect" mode.
This is okay as long as the same settings are specified in both modes and the collection runs that are compared against each other, however, this is inconvenient and can be misleading when comparing two different settings, e.g. Batched Nested Loop disabled vs. enabled, colocated vs. uncolocated, etc....
Also when manually running the scripts during experiments, the collect vs. report mode config can easily go out of the sync.
Would be great if the configuration settings are saved into the .json file produced by the "collect" mode, and the "report" mode shows them with the differences highlighted.
Framework will now report configuration if it is available in json files