Cell-veto in 2D: problem in computing event rates

[GabrieleTartero]

Hi Philipp and Werner, I might have found a bug in the JF implementation, concerning the cell-veto algorithm in 2D. To determine the event rate for every pair of cells, JF uses an object called Estimator , which is endowed with the method derivative_bound . The latter relies on the coordinates of the lower and upper corners of the target cell to do its computation. However, as you can see from one of my screenshots, the code is written as the tuples lower_corner and upper_corner always had three elements (=coordinates), but they only have two when working in 2D. Indeed, when running my conf file in 2D, I get a list index out of range error (other screenshot). Is this a JF limitation, or is it rather me getting something wrong with the configuration file? Thanks, Gabriele

[PhilippHoellmer]

Hi Gabriele,

this is indeed specifically implemented for three dimensions. To generalize this class, one should rely on jellyfysh.setting.dimension (as, e.g., in the DipoleInnerPointEstimator class).