context
Our regression test (test.sh) uses a mock setup that's very simple and unrealistic. This test is executed upon pull request.
problem
We want to check that our changes don't break existing features in real case scenarios, with different realistic use cases.
expected result
Devs can run a script that will simulate wildfires using the Kitral fuel model on 2 landscapes where fires have occurred. The results will be compared to target results obtained with a stable version of the algorithm. If results differ, the test fails.
If the test fails as expected, then the dev will update the target results accordingly and later changes will have to comply with these new results.
context
Our regression test (test.sh) uses a mock setup that's very simple and unrealistic. This test is executed upon pull request.
problem
We want to check that our changes don't break existing features in real case scenarios, with different realistic use cases.
expected result
Devs can run a script that will simulate wildfires using the Kitral fuel model on 2 landscapes where fires have occurred. The results will be compared to target results obtained with a stable version of the algorithm. If results differ, the test fails.
If the test fails as expected, then the dev will update the target results accordingly and later changes will have to comply with these new results.