Passing ivec complete to fedorov

[mdlerch]

Is this necessary when we are already passing Xc (the candidate set?)

I was trying to future proof for when there exists an option to specify necessary design points.

I don't see how it would work with the complete vector (naively I was thinking that the vector would just be a subset of 1:n). Now I suspect what we will need is a vector that indexes the non-necessary indexes in current.

So, if it's an 8 point design and we want to specify that points 1 and 2 are required, then current would be a vector of length 8 with the indexes of the current points and swappable (or something) will be a vector of [2,3,4,5,6,7,8](that is, 0 and 1 are not swappable).

https://github.com/mdlerch/OptimalDesign/blob/1214e8bf0b2350dc6818db8e057505420a66cfd3/src/fedorov.cpp#L13-14

[mdlerch]

I updated the complete argument based on adding the repeated option. Now the argument is called candidateidx is a list of legal indices to use as next proposition. We should add a new argument if/when we get to the point of necessary design points.