Create S-expression parser to reverse token extraction

[jwalsh]

Implement a parser that can read S-expression AST files and rebuild the original text. This is the reverse operation of the extract-tokens.scm script and would allow for:

1. Round-trip verification of the AST format
2. Reconstruction of text from modified ASTs
3. AST-based text generation or transformation pipelines

The parser should:

• Handle all node types in the schema
• Properly process metadata attributes
• Maintain whitespace and formatting based on the AST structure
• Support both full document and partial AST fragment reconstruction

This would complement the existing token extraction functionality and provide a complete bidirectional toolchain.