...
The same as above, with FuzzyClustering. It only exists in the
CVS branch CalReconFuzzyClustering.
Output Treatment
TO BE DETAILED MORE.
As said in the general description, the log and ref outputs are processed before being compared. Here are the steps of the process :
- we remove all the lines except the ones starting with "test_CalRecon".
- we extract and sort the energies displayed by the profile tool, the
last layer tool, and the display of the clusters. - we compute a mean value for each kind of energy, and a hashed value
from the positions and directions (remember, especially for this last
hash value : the idea is not to have meaningfull values, but something
we could check for change, so a hashed value is fine enough).