Tests for 'symmetry' regression i.e. does a change to code make any difference whether the model input domain terrain is flipped from left to right. So in this case, the river flows East-West. The flipped test case simply flips the data so that it now runs West-East. A typical breaking change to the model should break both tests, if only one is broken, something more subtle may be wrong...
For reference: these are the DEMs (outlet on the West - original, outlet on the East - flipped)
Units are 'grid cells' at 50m resolution
Possibly need to set OMP_NUM_THREADS in the travis yml and run_tests script to replicate results exactly between local machine and travis server, but otherwise OK to merge now
Tests for 'symmetry' regression i.e. does a change to code make any difference whether the model input domain terrain is flipped from left to right. So in this case, the river flows East-West. The
flipped
test case simply flips the data so that it now runs West-East. A typical breaking change to the model should break both tests, if only one is broken, something more subtle may be wrong...For reference: these are the DEMs (outlet on the West - original, outlet on the East - flipped) Units are 'grid cells' at 50m resolution