v0.7.0 - Add Feature Flags for Selective Compilation

Release Note for RustyML v0.7.0

Added

• Annotate mathematical utility functions with #[inline] for potential performance improvements during compilation
• add feature flags for selective compilation
• Add load_diabetes_raw_data function to provide static raw diabetes dataset and headers
• Add load_titanic_raw_data function to provide static raw Titanic dataset and headers
• Add cost calculation and reporting for ML models
• Add load_boston_housing_raw_data function to provide static raw Boston housing dataset and headers
• add batch processing for fit in Sequential model
• Add Linear activation function to support identity transformations
• Add sequential_test module with comprehensive tests for Sequential neural network functionality
• Add #[inline] attribute to performance-critical functions for potential inlining optimization in metric.rs
• Add comprehensive tests across modules
• Derive Copy and Clone for Activation enum to enable value duplication without ownership transfer

Changed

• Replace trait-based getter implementation with macros for improved reusability and consistency
• Refactor imports
• modularize traits
• Replace HashSet with AHashSet in DBSCAN and LDA
• Refactor dataset loading functions to use OnceLock for thread-safe memoization and add owned copy variants
• Refactor activation handling: consolidate apply_activation and activation_derivative logic into specialized in-place methods
• Consolidate helper_functions into layer.rs for improved modularity and maintainability. Refactor layer imports to remove dependency on helper_functions
• Update README for features clarification