feat: add signature extraction mode for static code analysis

[RMLaroche]

Summary

This PR adds a new --signature-mode feature that extracts class and method signatures from code files without requiring AI or API keys. This enables fast, offline static code analysis.

Key Features

• Static Code Analysis: Extracts class definitions, method signatures, and function signatures
• Multi-Language Support: Python (AST), Java, JavaScript, TypeScript (Tree-sitter when available)
• Offline Operation: No API key required, works completely offline
• Fast Processing: Much faster than AI summarization for signature extraction
• Mode Compatibility: Works with both --update (default) and --scan-all modes

Technical Implementation

• StaticAnalyzer: New module for signature extraction with fallback strategy
• Python AST Fallback: Uses built-in ast module when Tree-sitter unavailable
• Tree-sitter Integration: Multi-language parsing when tree-sitter-languages available
• Graceful Degradation: Falls back to Python AST if Tree-sitter fails
• Comprehensive Testing: 11 test cases covering all functionality

Usage Examples

# Extract signatures from current directory
codectx --signature-mode

# Extract signatures from all files (not just changed)
codectx --signature-mode --scan-all

# Extract signatures with custom output file
codectx --signature-mode --output-file signatures.md

Output Format

Generates structured markdown with:

• Class definitions with inheritance information
• Method signatures with type annotations and decorators
• Function signatures with parameters and return types
• Visibility modifiers (for Java)

Test Coverage

• ✅ 11 comprehensive test cases
• ✅ Python AST signature extraction
• ✅ Error handling and fallback scenarios
• ✅ Output formatting validation
• ✅ Multi-language file detection
• ✅ Integration with existing CLI framework

🤖 Generated with Claude Code