RMarkdown: Resolve Figure references in Bookdown syntax

[alex-ketch]

We're already using Bookdown syntax for encoding Figure captions, we should also handle encoding/decoding of figures (and other nodes) given in Bookdown syntax e.g. \@ref(fig:label).

The full specification can be found at https://bookdown.org/yihui/bookdown/cross-references.html

Encoding to RMD is a little more complex than in the MD as we need to look up the reference node type and alter the encoding accordingly:

• Equations: \@ref(eq:label)
• Theorems: \@ref(prefix:label) where prefix is one of these environments
• Figure: \@ref(label)
• Tables: \@ref(label)(although there is potential complexity for tables, please see the spec)

As a first step we can translate MDAST LinkReferences into Bookdown \@ref(label) format. Haven't quite figured out how Bookdown treats reference text as in the equivalent of [some text][label].