Allow distance decay functions truncated at 120 minutes

[abyrd]

When using non-step decay functions with higher cutoffs (e.g. logistic decay with cutoff over 60 minutes), an analysis may fail with the error "Distance decay function must reach zero at or before 120 minutes." This is one of several general assertions that are checked on any decay function, but this one can easily fail on reasonable inputs. We should tolerate functions that are still above zero at 120 minutes.

Nonetheless, as this may produce surprising results we will want to show a warning to the user (e.g. in single point results). But once you're aware of the situation, the result is a straightforward interaction with our usual 120 minute travel time cap.