bin_apply does not downsample

[qgeissmann]

e.g. we want to sample at 1/10 s

dt <- toy_dam_data()
regular_data <- behavr::bin_apply_all(dt,
									x = t,
									y = activity,
									x_bin_length = 10)

unique(diff(regular_data[,t]))

we get 60
this explains

rethomics/zeitgebr#7