Struct SequenceLearner 

pub struct SequenceLearner;

Sequence learner for discovering behavioral patterns

Implementations

impl SequenceLearner

pub async fn discover_sequences_from_traces( provider: &dyn TraceQueryProvider, min_frequency: f64, min_requests_per_trace: Option<usize>, ) -> Result<Vec<BehavioralSequence>>

Discover sequences from trace correlation

Uses trace_id/span_id to find correlated request sequences and identifies common patterns.

This implementation:

1. Queries database for requests grouped by trace_id
2. Orders by timestamp within each trace
3. Identifies common subsequences (e.g., login → list → detail)
4. Calculates transition probabilities between steps
5. Returns reusable behavioral sequences

pub fn learn_sequence_pattern( sequences: &[Vec<(String, String, Option<u64>)>], min_frequency: f64, ) -> Result<Vec<BehavioralSequence>>

Learn sequence pattern from a set of request sequences

Analyzes multiple sequences to extract common patterns and calculate confidence scores.

Each sequence is represented as (endpoint, method, delay_ms) tuples.

pub fn match_sequence<'a>( sequences: &'a [BehavioralSequence], endpoint: &str, method: &str, conditions: Option<&HashMap<String, String>>, ) -> Option<(&'a BehavioralSequence, usize)>

Check if an incoming request matches a learned sequence

Returns the matching sequence and current step index if found.

pub fn find_next_step( sequence: &BehavioralSequence, current_step_idx: usize, endpoint: &str, method: &str, ) -> Option<usize>

Find the next step in a sequence given the current step

Returns the next step index if found, or None if sequence is complete.