Glyphs as sublayers of term cell layers

[ghost]

Currently we draw an entire term cell as an image that is the contents of a single layer. Interaction (highlighting, selection) is at the layer level, hence we’re always highlighting or selecting an entire term cell.

Now that we have modelled hReps & glyphs for issues #75 and #72, we can actually split a term cell layer into sublayers, each sublayer mapping to a glyph, which would allow us to interact with individual (or groups of) glyphs. For example, in a term cell with two terms, we’d be able to select a single term.