Support packing data across multiple ciphertexts in a tensor

[j2kun]

The layout system supports describing a layout that packs data among multiple ciphertexts (e.g., (x, y, z) -> (x, F(x, y, z)) would use the first dimension as a ciphertext selector while the slot within a ciphertext is chosen according to F.

The initial convert-to-ciphertext-semantics pass supports only packing into a single ciphertext for now.