Add a wrapper around AST that exposes fields

[asimpletune]

The process for exposing the data that's parsed is a little strange and convoluted. There's an additional "rewriting" step, where the original rfc5322 grammar is rewritten with the help of a metagrammar, and variable names A-Z are assigned to each atom, before being saved. In the same file that coordinates the rewriting, there's the opportunity to rewrite some of those automatically assigned field names to ones that are more easily read.

After this, the resulting rewritten grammar can then be used to generate a new parser, while the original rfc5322 grammar remains untouched.

The reason for doing this is the grammar is too big to easily add computed fields without at the same time making it illegible and basically turning the grammar into a program. It's very important for the original grammar to easily be understandable, but also for there to be an easy way for changes made to it to automatically cause changes downstream.

As a final thought, there are at least big remaining issues left in the project:

1. There needs to be a lot more tests written, especially around things like obsolete fields
2. The project needs to reach a baseline acceptable level of tooling, like linters, a tsconfig file, etc...
3. Nothing special is being done for the body parsing. Some other projects I've seen are able to split the body into several parts, each corresponding to a different content-type. This is more inline with how email works in the real world, although it's probably out of scope for 5322 spec.