Improve seismogram plotting scripts.

[fabiolsilva]

Modify the seismogram plotting scripts to better handle cases where the time series contain a large leading and/or trailing area of inactivity. Plotting the entire time series for a 400s long file, will generate a plot with very little detail that is not very useful. One possible solution would be estimating the start of the activity using the station distance from the rupture to guess when to start plotting. Another idea is to use Arias intensity to detect the start/end of the activity and then add a buffer on both sides.