The `FindSubtreeRoots`

component filters an unordered set of row locations down to a
lexicographically ordered set of row locations where each location corresponds to the root
of a subtree.

Where multiple solutions are possible this algorithm will find the set of subtrees with the minimal number of subtrees by grouping connected row locations together into a single subtree.

The algorithm used to perform this conversion makes a simplifying assumption that the roots of all subtrees in the result share a common parent. If this assumption is untrue then the input set is unsuitable and the algorithm will return an empty optional.

The above example fails due to a *shallow* root, a root which has a depth *smaller* than the first
root we encounter.

This example fails due to a *deep* root, a root which has a depth *larger* than the first
root we encounter.

This example fails due to the fact that the parents of the subtree roots are not the same.

These conditions are checked using the functions defined in `Subtree.h`

.

This algorithm is used by Job Tree View in the `selectedSubtreeRoots`

method which is
required to implement non-trivial copy and paste.