evolution.toon.yaml

# code2llm/evolution | 3602 func | 164f | 2026-05-06
# generated in 0.42s

NEXT[1] (ranked by impact):
  [1] !! SPLIT           code2llm/exporters/index_generator/renderer.py
      WHY: 637L, 1 classes, max CC=1
      EFFORT: ~4h  IMPACT: 637


RISKS[1]:
  ⚠ Splitting code2llm/exporters/index_generator/renderer.py may break 2 import paths

METRICS-TARGET:
  CC̄:          2.3 → ≤1.6
  max-CC:      14 → ≤7
  god-modules: 1 → 0
  high-CC(≥15): 0 → ≤0
  hub-types:   0 → ≤0

PATTERNS (language parser shared logic):
  _extract_declarations() in base.py — unified extraction for:
    - TypeScript: interfaces, types, classes, functions, arrow funcs
    - PHP: namespaces, traits, classes, functions, includes
    - Ruby: modules, classes, methods, requires
    - C++: classes, structs, functions, #includes
    - C#: classes, interfaces, methods, usings
    - Java: classes, interfaces, methods, imports
    - Go: packages, functions, structs
    - Rust: modules, functions, traits, use statements

  Shared regex patterns per language:
    - import: language-specific import/require/using patterns
    - class: class/struct/trait declarations with inheritance
    - function: function/method signatures with visibility
    - brace_tracking: for C-family languages ({ })
    - end_keyword_tracking: for Ruby (module/class/def...end)

  Benefits:
    - Consistent extraction logic across all languages
    - Reduced code duplication (~70% reduction in parser LOC)
    - Easier maintenance: fix once, apply everywhere
    - Standardized FunctionInfo/ClassInfo models

HISTORY:
  prev CC̄=2.3 → now CC̄=2.3