feat: named tables and array type

[dvhthomas]

Summary

Add named markdown tables and a first-class Array type to CalcMark, enabling tabular data to serve as a computation source.

Canonical use case: Consulting SOW where a table of roles, rates, and headcount computes line totals and grand totals — no data duplication between tables and calc blocks.

Key capabilities

• Directive-based table naming: <!-- table: rates (role, rate, hc) -->
• Dot-access syntax for column arrays: rates.rate, rates.hc
• Element-wise arithmetic: rates.rate * rates.hc
• Aggregate functions on arrays: sum(), avg(), min(), max(), count()
• Array interpolation in table cells: {{line_costs}} per row

Documents

• Requirements: docs/brainstorms/2026-04-06-named-tables-and-arrays-requirements.md
• Plan: docs/plans/2026-04-06-003-feat-named-tables-and-arrays-plan.md

Plan structure (10 units, 6 phases)

1. Type Foundation: Array + Table types
2. Syntax: AST node, lexer DOT generalization, parser MemberAccess + new function tokens
3. Document Pipeline: Table directive parsing + extraction from TextBlocks
4. Semantic + Interpreter: MemberAccess eval, element-wise ops, aggregate functions
5. Display + Interpolation: Array formatting, array-in-table-cell interpolation, JSON output
6. Integration: Classifier, feature registry, golden tests

[dvhthomas]

Plan ready

Requirements and implementation plan committed:

• Requirements: docs/brainstorms/2026-04-06-named-tables-and-arrays-requirements.md
• Plan: docs/plans/2026-04-06-003-feat-named-tables-and-arrays-plan.md

10 implementation units across 6 phases. Key P0 from review addressed (parser arg count relaxation for array-accepting functions).