Improve inter-eventblock-distances in sequences

[ReneSkukies]

Request to improve control over inter-eventblock-distances.

In #104 event sequences are added to UnfoldSim. These are implemented expressed as for example {SR}-blocks, such that in this case an S event is always followed by an R event. You can add an underscore to the formula {SR_}, to indicate that there should be no overlap between one SR-block (i.e. pair) and the next.

It would be handy to have more control over the time between those blocks. For example, one could implement a keyword that takes in a distribution from which to pull the distances between the last event of one block and the first of the next. But I am also not sure how this is currently implemented and how viable that is

[behinger]

Iirc currently I take the size of the longest component/basis function and multiply it times 2 or 3. We could either allow to specify a factor, or specify a distance in samples. What would you prefer?

[ReneSkukies]

I can see merit in both:

A factor might be more intuitiv to the user. Takes away having to think about and calculate the samples from latencies (and thus a possible bug)

Samples is a better way to control the exact overlap/ distance ('cause I don't think there is an easy/ direct way to get the length of the components/ basis functions/ ERPs). But maybe I think samples is a better way, now that I think about it...

How would you handle implementing a random jitter? Cause from the above it seems both options would be a static distance

[behinger]

• samples it is then. By default we do 2*maxlength
• Easy way is: maxlength([c1,c2,c3]) - but not sure it is exported
• random jitter: you just provide a function that fullfills the following interface:

(rng::AbstractRNG,n)::Vector{Int} with size n.

e.g. rand