Andy Dixon
e9c1b25c80
Phase 4 (ADR-0001). internal/attention scores thought candidates as a normalised weighted sum over ten named features with a per-term breakdown, ranked by (score, id) so ordering is total and replayable. internal/workspace holds the capacity-limited workspace (refresh by subject, decayed competition, explained admit/reject outcomes) and the decaying working memory. internal/brain runs one runtime per active mind: debounced wake-ups from the event bus (own output never re-triggers), a settle window so concurrently committing events are not skipped, bounded concurrent Observe calls with per-process isolation and timeouts, deterministic scoring, a loop detector that suppresses recurring thoughts and raises a metacognition warning, executive dispatch by thought kind, and one transaction per cycle whose events are published only after commit. Every candidate, selection, attention shift and broadcast is an event with provenance back to the triggering observation. Operators can run a cycle, pause, resume, reset transient state and change attention weights live; the manager reconciles runtimes with mind status. cognition/attend is the first process: deterministic candidates from external events using an urgency lexicon, novelty against context and trust-based operator priority.
58 lines
1.6 KiB
Go
58 lines
1.6 KiB
Go
package brain
|
|
|
|
// loopDetector spots candidates that keep being admitted without the world
|
|
// changing (T20). It keeps a ring of recent admitted fingerprints; when one
|
|
// recurs threshold times within the window it is suppressed and a
|
|
// metacognition warning is raised once per streak.
|
|
type loopDetector struct {
|
|
window int
|
|
threshold int
|
|
ring []string
|
|
warned map[string]bool
|
|
}
|
|
|
|
func newLoopDetector(window, threshold int) *loopDetector {
|
|
return &loopDetector{window: window, threshold: threshold, warned: map[string]bool{}}
|
|
}
|
|
|
|
func fingerprint(c Candidate) string { return c.Process + "|" + c.Kind + "|" + c.Subject }
|
|
|
|
// count returns how many times fp appears in the window.
|
|
func (l *loopDetector) count(fp string) int {
|
|
n := 0
|
|
for _, x := range l.ring {
|
|
if x == fp {
|
|
n++
|
|
}
|
|
}
|
|
return n
|
|
}
|
|
|
|
// looping reports whether admitting fp again would exceed the threshold.
|
|
func (l *loopDetector) looping(fp string) bool { return l.count(fp) >= l.threshold }
|
|
|
|
// record notes an admission.
|
|
func (l *loopDetector) record(fp string) {
|
|
l.ring = append(l.ring, fp)
|
|
if len(l.ring) > l.window {
|
|
l.ring = l.ring[len(l.ring)-l.window:]
|
|
}
|
|
if l.count(fp) < l.threshold {
|
|
delete(l.warned, fp)
|
|
}
|
|
}
|
|
|
|
// shouldWarn returns true the first time a fingerprint is suppressed in a streak.
|
|
func (l *loopDetector) shouldWarn(fp string) bool {
|
|
if l.warned[fp] {
|
|
return false
|
|
}
|
|
l.warned[fp] = true
|
|
return true
|
|
}
|
|
|
|
// snapshot exports the ring for persistence.
|
|
func (l *loopDetector) snapshot() []string { return append([]string{}, l.ring...) }
|
|
|
|
func (l *loopDetector) restore(ring []string) { l.ring = append([]string{}, ring...) }
|