While I was sitting in the hospital, bored out of my mind, I noticed that the different BoxMap constructors have different names for the same type of thing. Specifically, the keyword arguments are not uniform:
BoxMap(:montecarlo, f, domain, no_of_points)
BoxMap(:grid, f, domain, no_of_points)
BoxMap(:interval, f, domain, no_subintervals)
Now the keywords all have the prefix n_ as in
BoxMap(:montecarlo, f, domain, n_points)
BoxMap(:grid, f, domain, n_points)
BoxMap(:interval, f, domain, n_subintervals)
For very expansive systems, the default BoxMap(:adaptive, ...) can cause errors due to the svd algorithm not converging or the singular values being too large to represent as Int64. There is now a more detailed error message that helps explain this error so it doesn't just look like a bug.
Finally, I added a more general function for computing the box dimension of any sequence of Boxsets.
While I was sitting in the hospital, bored out of my mind, I noticed that the different BoxMap constructors have different names for the same type of thing. Specifically, the keyword arguments are not uniform:
BoxMap(:montecarlo, f, domain, no_of_points)
BoxMap(:grid, f, domain, no_of_points)
BoxMap(:interval, f, domain, no_subintervals)
Now the keywords all have the prefixn_
as inBoxMap(:montecarlo, f, domain, n_points)
BoxMap(:grid, f, domain, n_points)
BoxMap(:interval, f, domain, n_subintervals)
For very expansive systems, the default
BoxMap(:adaptive, ...)
can cause errors due to the svd algorithm not converging or the singular values being too large to represent asInt64
. There is now a more detailed error message that helps explain this error so it doesn't just look like a bug.Finally, I added a more general function for computing the box dimension of any sequence of
Boxset
s.