Module feedback 

Implicit Feedback Collection

Automatically captures user interaction signals to improve search quality. Unlike explicit feedback (“this was helpful”), implicit feedback is derived from natural usage patterns.

Signal Types

Signal	What It Means	Learning Impact
Query	User searched for something	Records query patterns
View	User looked at a result	Weak positive signal
Select	User chose this result	Strong positive signal
Ignore	Result shown but not viewed	Weak negative signal
Dwell	Time spent on result	Strength of interest
FollowUp	Query after viewing result	Indicates gap or relation
CoAccess	Items accessed together	Hidden relationship

Example

use converge_knowledge::learning::FeedbackCollector;

let collector = FeedbackCollector::new();

// Start a session
let session = collector.start_session();

// Track a search
let query_id = collector.record_query(&session, "rust async patterns", &results);

// User views result #2
collector.record_view(&session, query_id, results[1].entry_id);

// User selects (clicks, copies, etc.) result #2
collector.record_select(&session, query_id, results[1].entry_id);

// Later, process accumulated feedback
let signals = collector.drain_signals();
learning_engine.apply_implicit_feedback(signals);

Structs

FeedbackCollector

Collector for implicit feedback signals.

FeedbackConfig

Configuration for the feedback collector.

FeedbackProcessor

Process raw signals into learning updates.

FeedbackSignal

A captured feedback signal with context.

ProcessedFeedback

Processed feedback ready for the learning engine.

QueryId

Query identifier for tracking result interactions.

SessionId

Session identifier for grouping related interactions.

Enums

SignalType

Type of implicit signal captured.