Module types 

Common Types for ML Model Integration

SPARC Phase 1: Specification

This module defines common types for ML model configuration and hybrid detection mode that combines heuristic and ML approaches.

Design Principles

• Flexibility: Support multiple model variants (FP32, FP16, INT8)
• Graceful Degradation: Fallback to heuristics if ML fails
• Performance: Cache-aware configuration
• Observability: Rich metadata for monitoring

Structs

CacheSettings

Cache settings for ML result caching

InferenceMetrics

Inference performance metrics

MLConfig

ML detection configuration for scanners

Enums

DetectionMethod

Detection method used for a scan result

HybridMode

Hybrid detection mode