Implement comprehensive 10-dimension code quality analysis

The review system now analyzes code across multiple dimensions:
- Unnecessary complexity
- Poor abstractions
- Unintended code deletion
- Hallucinated components
- Style inconsistencies
- Security vulnerabilities
- Performance issues
- Code duplication
- Incomplete error handling
- Test coverage gaps

Each dimension provides structured analysis with severity levels,
specific locations, detailed explanations, and actionable
recommendations. The UI has been updated to present this detailed
analysis in a clear, organized format with appropriate color coding
for different severity levels.

Updated CLI description to reflect these enhanced capabilities.

Extract quality dimensions into dedicated enum for better maintainability

This refactoring improves the code review system by:
- Creating a QualityDimension enum to centralize dimension definitions
- Adding methods to generate descriptions and display names
- Updating the review system to use this enum throughout
- Adding support for a new "Best Practices" dimension
- Simplifying dimension iteration in the UI
- Improving review formatting with consistent structure

The changes make the code more maintainable by eliminating duplicate
dimension definitions and centralizing quality analysis logic.

Add comprehensive documentation for 11 quality dimensions in code review

This update enhances the code review feature documentation by:
- Adding all 11 quality dimensions to the README feature list
- Providing detailed descriptions for each dimension in both files
- Including information about severity levels and report structure
- Adding a dedicated section for the Best Practices dimension
- Reorganizing the code review docs with the new section numbering

The changes ensure users understand the full capabilities of
the multi-dimensional code quality analysis system.

Implement specialized waiting messages for the code review feature

This change enhances the user experience during code reviews by:
- Adding 30 playful, themed waiting messages specific to code quality analysis
- Creating a dedicated get_review_waiting_message() function
- Simplifying the spinner message format in the review command
- Using cosmic and analytical imagery that aligns with the multi-dimensional
  code quality analysis system

The new messages provide users with entertaining feedback while they wait
for the comprehensive code review to complete.

Move shared types into specialized modules for better organization

This refactoring improves code organization by:
- Creating a new types.rs module for commit message types
- Moving review-related types from context.rs to review.rs
- Updating import paths across the codebase
- Adding re-exports in lib.rs for backward compatibility
- Maintaining consistent functionality while improving structure

The changes reduce duplication and make the code more maintainable by
clearly separating concerns between different components.

Improve the visual appearance and readability of code reviews

This update enhances the code review UI with:
- Modernized headers with decorative Unicode characters
- Improved section organization with consistent visual styling
- Better formatting for issue details with clear severity indicators
- Smart location formatting to handle various file path formats
- Text wrapping for long explanations to improve readability
- Dimension-specific emojis to visually distinguish categories
- Consistent bullet point styling across all list sections

Added tests for the new location formatting functionality to ensure
proper handling of various path formats.

Optimize regex pattern compilation across file analyzers

This performance improvement replaces all on-demand regex compilations
with Lazy static initialization using once_cell, which:

- Eliminates redundant regex compilation costs during analysis
- Improves error handling with descriptive messages for each pattern
- Maintains consistent naming conventions with _RE suffix
- Adds helpful documentation comments for each regex pattern

The change affects all file analyzer modules and the change_analyzer,
reducing runtime overhead while keeping the same functionality.

Remove String conversion implementations for response types

This refactoring improves type safety and validation by:
- Adding JsonSchema trait to response types for better validation
- Removing custom From<String> implementations that were error-prone
- Enhancing JSON parsing with provider-specific handling
- Adding fallback extraction methods for malformed responses
- Creating specialized parsing functions for different provider outputs

The changes affect core LLM interaction code and response types while
maintaining the same functionality with more robust error handling.