bugs in KPM module
So far I have encountered 3 bugs related to the KPM module
- When the number of moments is odd and
operator=None
, the algorithm takes a shortcut but fails to calculate the last (odd) moment - When increasing the number of moments or random vectors, unnecessary extra calculations are performed
- Inconsistent behaviour when using
bounds
- If bounds=None, the initialization will calculate them, but they are not saved. So when
increasing_accuracy
, the bounds will be calculated again. - Two instances with a)
bounds=None
, and b)bounds=(lmin, lmax)
, will not yield the same densities.
- If bounds=None, the initialization will calculate them, but they are not saved. So when