First event at 0 triggers incorrect warning

[SNAnalyst]

If the event list has a first event that occurs at exactly time = 0, remify gives the following warning:

Warning: at least two events (or more) occurred at the same time point. The interevent time of such events will be evenly spaced.

This is incorrect. If time = 0 is not admissable (if so, why not?), then there should be a message that time = 0 is not admissable. Otherwise, it should not trigger this warning.

[jupepis]

The general assumption is that the first event is always observed after the starting time point, so t[1] > t[0] with t[0] being the starting time point (usually set at 0). Therefore, in the case where the first event occurs at t[1]=0, it means that the waiting time for the first event is 0, and it would be seen as an artifact of the data because actually some time has elapsed after the start of the study.

However, I was also wondering if we always want to keep the first event in the likelihood or use it just for the calculation of the statistics and then in the optimization part to consider the sequence starting from the second event.

I can now set a warning and anyway let the function process the first event time being 0.

[jupepis]

Solved in the next commit:

• in general, we need a time[0] when we started to record/annotate events, and at this time point, time[0], nothing happened,
• otherwise, if time[1] is 0, its waiting time will artificially be 0, and I do not think it is a realistic assumption.

At the moment, if time[1] is 0, its waiting time will be set to 0.

I would still think about some other alternative as to which waiting time to assume for the event observed at t[1]=0.