Text

The Text node is the catch-all: any input fragment that doesn’t match a typed production lands here. The grammar is total — it never fails on stray input, it just emits a Text node.

Fields

Field	Type	Description
content	String	The captured text content

Why a catch-all?

Two reasons:

1. Forward compatibility. Inputs that use constructs not yet implemented (e.g. aromatic rings before v0.3) parse successfully, with the unknown fragment as a Text node. The renderer echoes the source verbatim — no data loss.
2. Punctuation and whitespace. Reaction separators (+), padding spaces, and other syntactic glue between typed nodes are captured as Text so round-trip is exact.

Grammar production

text_run := (not structural-char)*

where structural-char excludes:
  ` (math escape)
  + (reaction separator)
  arrow tokens (<=>, <->, ->, <-)
  group brackets ( ) [ ] { }

The exclusion keeps structural boundaries crisp.

Formatter behaviour

Every formatter’s visit_text emits the content verbatim (with appropriate escaping for the output format).

Formatter	Output
MathML	<mtext>content</mtext>
Text	content
HTML	escaped content
LaTeX	content
SVG	content

Round-trip

Text nodes preserve their content exactly. The Text formatter is the canonicaliser; any input that survives parsing as Text will round-trip character-for-character.

When Text appears

A typical input contains Text nodes for:

• The + separators in reactions (2H_2 + O_2 has Text(" ") between molecules — though the grammar usually trims this).
• Multi-digit stoichiometric coefficients that don’t form a valid element (e.g. H_2O3 parses as Molecule[H_2, O], Text(“3”)).
• Punctuation the user includes (H_2O., with a literal period).

Children contract

Text is a leaf node — children returns [].