pattern

Documentation

Overview

Package pattern provides pattern detection and recognition functionality.

Package pattern provides pattern detection and recognition functionality.

Package pattern provides pattern detection and recognition functionality.

Index

• Constants
• func DynamicQuality(frequency int, confidence float64, projectCount int, maxFrequency int, ...) float64
• func QualityScore(frequency int, confidence float64, projectCount int, maxFrequency int) float64
• func RunDecay(ctx context.Context, patternStore *gorm.PatternStore) (int, error)
• func TemporalDecay(daysSinceLastSeen float64) float64
• type DetectionResult
• type Detector
  • func NewDetector(patternStore *gorm.PatternStore, observationStore *gorm.ObservationStore, ...) *Detector
  • func (d *Detector) AnalyzeObservation(ctx context.Context, obs *models.Observation) (*DetectionResult, error)
  • func (d *Detector) AnalyzeRecentObservations(ctx context.Context) error
  • func (d *Detector) BatchRecalculateConfidence(ctx context.Context) (int, error)
  • func (d *Detector) CandidateCount() int
  • func (d *Detector) GetPatternInsight(ctx context.Context, patternID int64) (string, error)
  • func (d *Detector) SetSyncFunc(fn PatternSyncFunc)
  • func (d *Detector) Start()
  • func (d *Detector) Stop()
• type DetectorConfig
  • func DefaultConfig() DetectorConfig
• type PatternSyncFunc

Constants

const (
	// DeprecateThreshold marks a pattern for soft-deprecation when Q_dynamic drops below this value.
	DeprecateThreshold = 0.10
	// DeleteThreshold marks a pattern as eligible for hard deletion when Q_dynamic reaches exactly 0.
	DeleteThreshold = 0.0
)

Lifecycle thresholds for pattern quality management.

Functions

func DynamicQuality

func DynamicQuality(frequency int, confidence float64, projectCount int, maxFrequency int, daysSinceLastSeen float64) float64

DynamicQuality computes the final quality score with temporal decay applied. Q_dynamic = QualityScore(...) * TemporalDecay(daysSinceLastSeen)

func QualityScore

func QualityScore(frequency int, confidence float64, projectCount int, maxFrequency int) float64

QualityScore computes a pattern's quality baseline score. Formula: Q_base = 0.3*log_freq + 0.5*sigmoid_conf + 0.2*diversity

Components:

• log_freq: log2(frequency+1) / log2(maxFrequency+1), normalized 0-1.
• sigmoid_conf: 1 / (1 + exp(-10*(confidence-0.5))), centered at 0.5.
• diversity: min(projectCount, 5) / 5.0, normalized 0-1.

func RunDecay

func RunDecay(ctx context.Context, patternStore *gorm.PatternStore) (int, error)

RunDecay processes all active patterns, computing dynamic quality and deprecating low-quality ones. Returns the count of deprecated patterns.

func TemporalDecay

func TemporalDecay(daysSinceLastSeen float64) float64

TemporalDecay computes the decay multiplier based on days since last seen. Three phases:

• Grace period (0-7 days): returns 1.0, no decay.
• Gaussian decay (7-60 days): smooth decay from 1.0 to ~exp(-3) ≈ 0.05.
• Linear terminal (60-90 days): linear ramp from Gaussian end value to 0.

Types

type DetectionResult

type DetectionResult struct {
	MatchedPattern *models.Pattern
	MatchScore     float64
	IsNewPattern   bool
}

DetectionResult contains the result of pattern detection.

type Detector

type Detector struct {
	// contains filtered or unexported fields
}

Detector detects and tracks recurring patterns across observations.

func NewDetector

func NewDetector(patternStore *gorm.PatternStore, observationStore *gorm.ObservationStore, config DetectorConfig) *Detector

NewDetector creates a new pattern detector.

func (*Detector) AnalyzeObservation

func (d *Detector) AnalyzeObservation(ctx context.Context, obs *models.Observation) (*DetectionResult, error)

AnalyzeObservation processes a new observation for pattern detection. This is called synchronously when a new observation is stored.

func (*Detector) AnalyzeRecentObservations

func (d *Detector) AnalyzeRecentObservations(ctx context.Context) error

AnalyzeRecentObservations analyzes recent observations for pattern detection. This is used for background batch analysis.

func (*Detector) BatchRecalculateConfidence

func (d *Detector) BatchRecalculateConfidence(ctx context.Context) (int, error)

BatchRecalculateConfidence recomputes confidence for all active patterns using the model formula (frequency + cross-project bonus) and persists updated values. Returns the number of patterns recalculated.

func (*Detector) CandidateCount

func (d *Detector) CandidateCount() int

CandidateCount returns the current number of pattern candidates.

func (*Detector) GetPatternInsight

func (d *Detector) GetPatternInsight(ctx context.Context, patternID int64) (string, error)

GetPatternInsight returns a formatted insight string for a pattern. GetPatternInsight returns a formatted insight string for a pattern.

func (*Detector) SetSyncFunc

func (d *Detector) SetSyncFunc(fn PatternSyncFunc)

SetSyncFunc sets the callback for syncing patterns to vector store.

func (*Detector) Start

func (d *Detector) Start()

Start begins background pattern analysis.

func (*Detector) Stop

func (d *Detector) Stop()

Stop stops background pattern analysis.

type DetectorConfig

type DetectorConfig struct {
	// MinMatchScore is the minimum similarity score to consider a match (0.0-1.0).
	MinMatchScore float64
	// MinFrequencyForPattern is the minimum occurrences before creating a pattern.
	MinFrequencyForPattern int
	// AnalysisInterval is how often to run background pattern analysis.
	AnalysisInterval time.Duration
	// MaxPatternsToTrack is the maximum number of active patterns.
	MaxPatternsToTrack int
	// MaxCandidates is the maximum number of candidates to track (LRU eviction).
	MaxCandidates int
}

DetectorConfig contains configuration for the pattern detector.

func DefaultConfig

func DefaultConfig() DetectorConfig

DefaultConfig returns the default detector configuration.

type PatternSyncFunc

type PatternSyncFunc func(pattern *models.Pattern)

PatternSyncFunc is a callback for syncing patterns to vector store.