Closed d-ver closed 4 years ago
We must. Not should. :)
Please update the issue with a checklist per-tool.
The idea is to make a python script that deletes the uncommon metrics between the confronted outputs, which are then compared with Json-diff. In this case, the Json files produced by this new script must not contain any empty metrics.
Would be better to just extract the single metric we want to test and compare it.
Thanks for opening this issue @m51galaxy!
As @lu-zero said, we could implement these tests through a Python
script that, instead of cleaning up a json
file, extracts the metrics we are interested in from the json files produced by the tools and compares all of them in order to verify if they are identic
It's much simpler than our initial approach :)
We must implement a TEST mode (argparse ...) to check if different tools return the same results for the metrics they have in common.
This mode should produce a separate Json output file for each of the used tools.