chore: import upstream snapshot with attribution

functions/edge/conditions.py

@@ -0,0 +1,48 @@
+"""Edge condition helpers used by workflow YAML definitions."""
+
+import re
+
+def contains_keyword(data: str) -> bool:
+    """Check if data contains the keyword 'trigger'."""
+    return "trigger" in data.lower()
+
+def length_greater_than_5(data: str) -> bool:
+    """Check if data length is greater than 5."""
+    return len(data) > 5
+
+def always_false(data: str) -> bool:
+    """Always return false for testing."""
+    return False
+
+def not_contains_keyword(data: str) -> bool:
+    """Check if data contains the keyword 'trigger'."""
+    return "trigger" not in data.lower()
+
+def code_pass(data: str) -> bool:
+    return "==CODE EXECUTION FAILED==" not in data
+
+def code_fail(data: str) -> bool:
+    return "==CODE EXECUTION FAILED==" in data
+
+def _extract_verdict(data: str) -> str | None:
+    """Parse `Verdict: CONTINUE|STOP` markers from agent outputs."""
+    match = re.search(r"verdict\s*:\s*(\w+)", data, re.IGNORECASE)
+    if not match:
+        return None
+    return match.group(1).upper()
+
+
+def need_reflection_loop(data: str) -> bool:
+    """Return True when the Reasoner asks for another Reflexion iteration."""
+    verdict = _extract_verdict(data)
+    if verdict is None:
+        return True  # Default to continuing until STOP is explicit
+    return verdict not in {"STOP", "DONE"}
+
+
+def should_stop_loop(data: str) -> bool:
+    """Return True when the Reasoner signals the Reflexion loop can stop."""
+    verdict = _extract_verdict(data)
+    if verdict is None:
+        return False
+    return verdict in {"STOP", "DONE"}