Add functions to help detect and fix cycles in neurons

[jefferis]

Several of the sample neurons have a loop due to a tracing error

> table(sapply(Cell07PNs, function(x) igraph::is.dag(as.ngraph(x))))

FALSE  TRUE 
    9    31 

i.e. 9/40 neurons have cycles. This is not an uncommon error in the wild. It would be good to think of a way to find these and fix them.