sleep algorithm underestimates sleep by four minutes per sleep bout

[ggilestro]

pysolo sleep algorithm scans activity using the following formula

single_flies5min[fly] = [( single_flies[fly][i-b1:i+a1].sum() <= inactivity ) for i in range (d*c)]

where in normal conditions b1+a1=5
However this will result in not counting the very first three minutes of inactivity and the very last minute of inactivity in every bout >= 5 minutes, that is an underestimation of "sleep" of 4 minutes per bout.

[ggilestro]

Keeping in mind the threshold for sleep detection is absolutely arbitrary, this is not a big issue. However, for future use, I am changing the algorithm to a version that will count all minutes of inactivity.