feat: [US-041] DOCX parser - OMML math equations

[developer0hye]

Summary

• Parse OMML (Office Math Markup Language) elements from DOCX paragraphs (m:oMath, m:oMathPara)
• Convert OMML elements to Typst math notation: fractions, super/subscripts, radicals, delimiters, n-ary operators, accents, matrices, equation arrays
• Add MathEquation IR type with inline/display math support
• Integrate via raw ZIP XML scanning pattern (same as WrapContext/NoteContext) since docx-rs doesn't parse m: namespace

Key changes

• New module parser/omml.rs: OMML-to-Typst converter with omml_to_typst() and scan_math_equations() functions
• IR extension ir/elements.rs: MathEquation struct and Block::MathEquation variant
• Parser integration parser/docx.rs: MathContext for body-child-index correlation between raw XML scan and docx-rs document tree
• Codegen render/typst_gen.rs: Typst $...$ (inline) and $ ... $ (display) math rendering

Test plan

• 13 unit tests for OMML conversion (fraction, superscript, subscript, sub+superscript, square root, nth root, parentheses, complex equation, sum with limits, E=mc², quadratic formula, scan display/inline/multiple)
• 3 DOCX integration tests (display math fraction, inline math superscript, complex math)
• 3 codegen tests (display math, inline math, complex math)
• 427 library tests pass
• cargo clippy --workspace -- -D warnings passes
• cargo fmt --all --check passes

🤖 Generated with Claude Code