Struct NeuralResonanceModule 

pub struct NeuralResonanceModule { /* private fields */ }

Neural Resonance Module

Uses Baked Mask Encoders instead of an external LLM. Each mask’s signature_vector is baked into a tiny MLP (~66KB). Total memory: O(num_masks * 66KB) — fits any VPS.

Implementations

impl NeuralResonanceModule

pub fn new(config: NeuralConfig) -> Result<Self, String>

Create new neural module

pub fn load_model(&mut self) -> Result<(), String>

Load model (marks as ready — no external model files needed)

pub fn register_mask(&mut self, mask: &MaskProfile) -> Result<(), String>

Register mask — bakes its signature into a dedicated MLP encoder

pub fn record_traffic( &self, session_id: [u8; 16], packet_size: u16, iat_ms: f64, entropy: f64, )

Record traffic sample for session

pub fn check_resonance( &self, session_id: [u8; 16], mask_id: &str, ) -> Result<ResonanceResult, String>

Perform resonance check (Patent 1: Signal Reconstruction Resonance)

Encodes live traffic into a 64-dim feature vector, passes it through the mask’s baked encoder, and computes reconstruction MSE.

pub fn record_telemetry(&mut self, mask_id: &str, packet_loss: f64, rtt_ms: f64)

Record telemetry for anomaly detection

pub fn is_mask_anomalous(&self, mask_id: &str) -> bool

Check if mask is anomalous (possible DPI blocking)

pub fn get_or_create_stats(&self, session_id: [u8; 16]) -> TrafficStats

Get or create session stats

pub fn cleanup_stats(&self, session_id: [u8; 16])

Cleanup old session stats

pub fn total_memory_bytes(&self) -> usize

Total memory usage for all baked encoders

pub fn encoder_count(&self) -> usize

Number of registered mask encoders