parse: Reserve keywords for reference names

parse: Parse a reference node

src: Rename syntax to token

parse: Remove SyntaxType ...

Once again I find myself in the pattern of having a useless 'type' value
that just complicates my life. In this case the correct appproach is to
output the AST directly.

parse: Parse text and bools to AST types

parse: Check types when clearing notes, add extra tests

syntax: Refactor draw_syntax_random and add draw_syntax_texttoken ...

I realized draw_syntax_classified is the same as draw_syntax_random, so
just merge them.

Also add draw_syntax_texttoken explicitly meant to weed out cases where
token type is not set.

tests: Clean up syntax and parse code

tests: Clarify function documentation

tests: Optimize and move draw_syntax_not_token to tests_syntax ...

Drawing random tokens and throwing them away was killing test
performance.

tests: Refactor mid-test assumes in to custom draws ...

This should help with performance as Hypothesis as we no longer need to
do complex unrelated draws first.

tests: Random insert_random to insert_random_within

tests: Refactor clear_notes tests

parse: Handle parent_context in template_parse_invalid

tests: Fix some typos for text tests

tests: Split up parser tests for empty cases

tests: Use one test per invalid case for text and notes

tests: Add template_parse_invalid

tests: Move insert_random to parse.templates

tests: Add a template for parsing valid syntax

tests: Refactor getter and setter tests in to a template

parse: Pass a context to the parse function

parse: Have parse_text handle contexts properly

parse: Have parse_bool handle contexts properly

tests: Clarify test name in test_clear_notes

parse: Use an error context for clear_notes

src: Pass a parent context to skip_note

parse: Have skip_note create a PARSE_NOT ParseContext

parse: Allow attaching a ParseContext to a ParseErrorException

parse: Add parse contexts for later error reporting