fix(perception): structural Resource-ID bypass gate + dead code purge in ActionMemory
P0-1: Toggle intents (follow/like/save) resolved via Resource-ID fast-path now skip VLM verification entirely. Eliminates the #1 session failure: VLM hallucinating Follow→Like, causing zero follows per session. P0-2: Purge 52 lines of unreachable dead code after unconditional return at line 290. Non-toggle structural delta verification now properly reachable. Remove 4 DEBUG-prefixed logger.info pollution statements. TDD: 7 new tests covering bypass gate, negative guard, non-toggle delta, and debug log pollution detection.
This commit is contained in:
@@ -196,6 +196,23 @@ class ActionMemory:
|
||||
state_toggles = ["like", "save", "follow", "heart"]
|
||||
is_toggle = any(t in intent_lower for t in state_toggles)
|
||||
|
||||
# ── P0-1: Structural Resource-ID Bypass Gate ──
|
||||
# If the clicked node was resolved via a structural Resource-ID that
|
||||
# directly matches the toggle intent, VLM verification is SKIPPED.
|
||||
# This eliminates the #1 session failure: VLM hallucinating Follow→Like.
|
||||
if is_toggle and self._last_click_context:
|
||||
clicked_rid = (self._last_click_context.get("node_dict", {}).get("resource_id", "") or "").lower()
|
||||
if clicked_rid:
|
||||
for intent_keyword, required_markers in TOGGLE_INTENT_MARKERS.items():
|
||||
if intent_keyword in intent_lower:
|
||||
if any(marker in clicked_rid for marker in required_markers):
|
||||
logger.info(
|
||||
f"⚡ [ActionMemory] Structural Resource-ID bypass: '{intent}' matched "
|
||||
f"'{clicked_rid}'. Skipping VLM verification — O(1) trust."
|
||||
)
|
||||
return True
|
||||
break # Only check the first matching intent keyword
|
||||
|
||||
# ── VLM Verification Fallback ──
|
||||
|
||||
# If we are highly confident (e.g. pulled from Qdrant memory), bypass heavy VLM
|
||||
@@ -270,10 +287,8 @@ class ActionMemory:
|
||||
)
|
||||
return False
|
||||
|
||||
# Fallback to structural delta
|
||||
logger.info(f"DEBUG: len(pre_click_xml)={len(pre_click_xml)} len(post_click_xml)={len(post_click_xml)}")
|
||||
# ── Structural Delta Verification ──
|
||||
diff = abs(len(pre_click_xml) - len(post_click_xml))
|
||||
logger.info(f"DEBUG: diff={diff}")
|
||||
|
||||
if is_toggle:
|
||||
if diff > 1000:
|
||||
@@ -288,58 +303,18 @@ class ActionMemory:
|
||||
f"⚠️ [ActionMemory] Zero structural shift (diff={diff}) for state-toggle '{intent}'. Verification FAIL."
|
||||
)
|
||||
return False
|
||||
# If the intent is an abstract goal (like "find customers"), diff > 50 is NOT enough.
|
||||
# We must force visual VLM confirmation because clicking the wrong thing (like "Create highlight")
|
||||
# also produces a large diff but achieves the wrong goal.
|
||||
if diff > 50:
|
||||
# Is it a standard structural transition?
|
||||
from GramAddict.core.screen_topology import ScreenTopology
|
||||
|
||||
# We don't have screen type here, so we just check if it's in the HD Map keys
|
||||
logger.info(f"DEBUG: intent is '{intent}'")
|
||||
logger.info(
|
||||
f"DEBUG: TRANSITIONS keys are: {[list(t.keys()) for t in ScreenTopology.TRANSITIONS.values()]}"
|
||||
)
|
||||
is_standard = any(intent in transitions for transitions in ScreenTopology.TRANSITIONS.values())
|
||||
logger.info(f"DEBUG: is_standard={is_standard}")
|
||||
|
||||
if is_standard:
|
||||
logger.debug(
|
||||
f"🧠 [ActionMemory] Structural change detected for known navigation '{intent}'. Verification PASS."
|
||||
)
|
||||
return True
|
||||
else:
|
||||
logger.info(
|
||||
f"👁️ [ActionMemory] Abstract intent '{intent}' caused UI change. Forcing VLM visual verification..."
|
||||
)
|
||||
# For abstract intents, we must visually verify if it actually helped!
|
||||
# If device is available, we use VLM. If not, we fail safe.
|
||||
if device:
|
||||
from GramAddict.core.perception.semantic_evaluator import SemanticEvaluator
|
||||
|
||||
evaluator = SemanticEvaluator()
|
||||
prompt = f"The user just attempted to perform the action: '{intent}'. Does the current screen match the expected outcome? Answer ONLY with the word YES or NO."
|
||||
try:
|
||||
response = evaluator._query_vlm(prompt, device.get_screenshot_b64())
|
||||
decision = _parse_yes_no(response) if response else None
|
||||
if decision is True:
|
||||
return True
|
||||
else:
|
||||
logger.warning(
|
||||
f"⚠️ [ActionMemory] VLM rejected success for abstract intent '{intent}'. Response: '{response}'"
|
||||
)
|
||||
return False
|
||||
except Exception as e:
|
||||
logger.error(f"VLM visual verification failed: {e}")
|
||||
|
||||
logger.warning(f"⚠️ [ActionMemory] Cannot visually verify abstract intent '{intent}'. Failing safe.")
|
||||
return False
|
||||
|
||||
# If diff <= 50 for non-toggle
|
||||
logger.warning(
|
||||
f"⚠️ [ActionMemory] Insufficient structural change (diff={diff}) for non-toggle '{intent}'. Verification FAIL."
|
||||
# ── Non-Toggle Structural Delta ──
|
||||
if diff > 50:
|
||||
logger.debug(
|
||||
f"🧠 [ActionMemory] Structural change detected ({diff} chars) for '{intent}'. Verification PASS."
|
||||
)
|
||||
return False
|
||||
return True
|
||||
|
||||
logger.warning(
|
||||
f"⚠️ [ActionMemory] Insufficient structural change (diff={diff}) for non-toggle '{intent}'. Verification FAIL."
|
||||
)
|
||||
return False
|
||||
|
||||
|
||||
def _intent_matches_node(intent: str, semantic_string: str) -> bool:
|
||||
|
||||
Reference in New Issue
Block a user