This part of the for loop doesn't actually do what we want. Both if-statements do the same roll so that a delay of e.g. 5 is the same as -5.
Example traces[i] == [0, 1, 2, 3, 4, 5]
if r = +2, we want output to be [4, 5, 0, 1, 2, 3]
Line 106: [2, 3, 4, 5, 0, 0]
Line 107: [2, 3, 4, 5, 0, 1]
if r = -2, we want output to be [2, 3, 4, 5, 0, 1]
Line 109: [2, 3, 4, 5, 0, 0]
Line 110: [2, 3, 4, 5, 0, 1]
I guess the direction that things get moved for positive values of r. But I don't think that we want both positive and negative delays should give the same result.
For comparison, if we use numpy's roll, one gets
In [1]: np.roll([i for i in range(6)], 2)
Out[1]: array([4, 5, 0, 1, 2, 3])
In [2]: np.roll([i for i in range(6)], -2)
Out[2]: array([2, 3, 4, 5, 0, 1])
https://github.com/nu-radio/NuRadioMC/blob/19e807e04a4aa3ddeda5e06c4008a94f15f3ec37/NuRadioReco/utilities/noise.py#L105-L110
This part of the for loop doesn't actually do what we want. Both if-statements do the same roll so that a delay of e.g.
5is the same as-5.Example
traces[i] == [0, 1, 2, 3, 4, 5]if
r = +2, we want output to be[4, 5, 0, 1, 2, 3][2, 3, 4, 5, 0, 0][2, 3, 4, 5, 0, 1]if
r = -2, we want output to be[2, 3, 4, 5, 0, 1][2, 3, 4, 5, 0, 0][2, 3, 4, 5, 0, 1]I guess the direction that things get moved for positive values of
r. But I don't think that we want both positive and negative delays should give the same result.For comparison, if we use numpy's
roll, one getsSince we want to use the switch at https://github.com/nu-radio/NuRadioMC/blob/19e807e04a4aa3ddeda5e06c4008a94f15f3ec37/NuRadioReco/utilities/noise.py#L533-L537 it is probably preferred that we use numpy def (also what is used in the actual PA trigger module for simulations). So we could fix this by removing the
elifloop and definer = -rolling[i]to get the same result.