Implement Phase 4 reasoning trigger gate with cooldown#137
Open
dilshadalikhan2004 wants to merge 2 commits into
Open
Implement Phase 4 reasoning trigger gate with cooldown#137dilshadalikhan2004 wants to merge 2 commits into
dilshadalikhan2004 wants to merge 2 commits into
Conversation
|
Caution Review failedFailed to post review comments 📝 WalkthroughWalkthroughThis PR refactors the memory trigger reasoning-gate API from a TrackLifecycleEvent-plus-action-strings pattern to a TrackSequence-based pattern, updates action-classifier configuration defaults, and revises tests to use TrackSequence/ActionHint with fixtures and helpers. ChangesMemory Trigger API and Test Refactoring
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related issues
Possibly related PRs
Suggested labels
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
services/memory/trigger.py (1)
11-11: 💤 Low value
_reasoning_cooldownsgrows unbounded and is unsynchronized.Entries are added on every successful trigger (Line 66) and only removed via
reset_cooldown. If reasoning fails orreset_cooldownis never called for some tracks, the dict accumulates one entry per track_id for the process lifetime. Additionally, the check-then-set on Lines 60-66 is not atomic, so concurrent invocations for the sametrack_idcould both trigger. If the pipeline ever processes tracks across threads, consider a lock and/or periodic eviction of stale entries.Also applies to: 57-66
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@services/memory/trigger.py` at line 11, _reasoning_cooldowns currently grows unbounded and is updated without synchronization; make access atomic and evict stale entries by introducing a lock (e.g., _reasoning_cooldowns_lock) and a TTL constant (e.g., REASONING_COOLDOWN_TTL) and wrap all reads/writes (checks, setting on trigger, and reset_cooldown) to _reasoning_cooldowns in that lock so the check-then-set around the trigger is atomic; additionally, on each access (or periodically) prune entries older than TTL so failed or forgotten track_ids don't accumulate, and ensure failures also clear their entries via reset_cooldown.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@services/memory/trigger.py`:
- Around line 43-45: The current gating in should_trigger_reasoning only ensures
seq.zones_visited is non-empty so it fires for any visited zone; either update
the function docstring to state it intentionally triggers on any visited zone,
or change the gating to require an intersection with the restricted-zone set
(e.g., filter seq.zones_visited against the restricted zones provided by Memory
or a RESTRICTED_ZONES constant) before returning True; locate
should_trigger_reasoning in services/memory/trigger.py and use the same
restricted-zone identifier used in services/memory/memory.py (or the
Memory.restricted_zones attribute) to perform the filtered check.
---
Nitpick comments:
In `@services/memory/trigger.py`:
- Line 11: _reasoning_cooldowns currently grows unbounded and is updated without
synchronization; make access atomic and evict stale entries by introducing a
lock (e.g., _reasoning_cooldowns_lock) and a TTL constant (e.g.,
REASONING_COOLDOWN_TTL) and wrap all reads/writes (checks, setting on trigger,
and reset_cooldown) to _reasoning_cooldowns in that lock so the check-then-set
around the trigger is atomic; additionally, on each access (or periodically)
prune entries older than TTL so failed or forgotten track_ids don't accumulate,
and ensure failures also clear their entries via reset_cooldown.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 706987e1-7a7a-41b2-896e-ae8f59ea75a6
📒 Files selected for processing (3)
libs/config/settings.pyservices/memory/trigger.pytests/test_trigger.py
Comment on lines
+43
to
45
| # Zone check | ||
| if not seq.zones_visited: | ||
| return False |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Inspect how zones_visited / zones_key is populated to see whether it is restricted-only.
rg -nP -C4 '\bzones_visited\b' --type=py
rg -nP -C4 '_zones_key|zones_present|restricted' --type=pyRepository: Devnil434/Eagle
Length of output: 50371
Fix restricted-zone gating in should_trigger_reasoning
services/memory/trigger.pyonly checks thatseq.zones_visitedis non-empty (if not seq.zones_visited: return False), so it will fire for any visited zone.services/memory/memory.pypopulateszones_visitedby collectingevent.zonevalues with no restricted-only filtering;tests/test_memory.py::test_zones_visited_populatedshows"safe_corridor"and"restricted_door"both ending up inseq.zones_visited.
Update the docstring/intent to match “any visited zone” or filterseq.zones_visitedagainst the restricted-zone set before allowing the trigger to proceed.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@services/memory/trigger.py` around lines 43 - 45, The current gating in
should_trigger_reasoning only ensures seq.zones_visited is non-empty so it fires
for any visited zone; either update the function docstring to state it
intentionally triggers on any visited zone, or change the gating to require an
intersection with the restricted-zone set (e.g., filter seq.zones_visited
against the restricted zones provided by Memory or a RESTRICTED_ZONES constant)
before returning True; locate should_trigger_reasoning in
services/memory/trigger.py and use the same restricted-zone identifier used in
services/memory/memory.py (or the Memory.restricted_zones attribute) to perform
the filtered check.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implemented the Phase 4 reasoning trigger gate to prevent unnecessary VLM/LLM invocations by evaluating track sequences before triggering reasoning.
Changes
Added
should_trigger_reasoning(seq: TrackSequence) -> boolAdded in-memory cooldown tracking
Added
reset_cooldown(track_id)Added suspicious action detection for:
LINGERINGNEAR_KEYPADREPEATED_APPROACHAdded dwell threshold validation using centralized settings
Updated trigger tests to use
TrackSequenceTests
Added/updated tests covering:
All trigger tests pass successfully.
Note
The existing trigger tests were based on the older
TrackLifecycleEventinterface. They were updated to align with theTrackSequence-based API specified in Issue #42.Summary by CodeRabbit
New Features
Refactor