Decide what a core-periphery or factional ideal means for a multilevel network

[jhollway]

net_by_core() and net_by_factions() fit a network to an ideal built by a
manynet::create_*() function, and correlate the two:

obs   <- manynet::as_matrix(.data)
ideal <- manynet::as_matrix(manynet::create_core(.data, mark = mark))

A create_*() function builds one layer. A multilevel network holds two, and
reports itself as two-mode, so the ideal comes back with the shape of a
two-mode network while the observed matrix is square over every node. On
fict_marvel that is a 53x141 ideal against a 194x194 observation, which
failed with "incompatible dimensions" (#26).

Both functions now stop with a message naming the shape and suggesting
to_mode1() or to_uniplex(). What they should do instead is open:

• Is there a core-periphery ideal for a network whose levels are of different
  kinds? A core that spans both levels is not the same claim as a core in each.
• The same question for a factional ideal.
• If there is one, the network has to be assembled from a one-mode and a
  two-mode layer, so the ideal probably has to be assembled the same way rather
  than produced by a single create_*() call.

Related to #26.