from typing import Any, Dict from server.tasks.base_task import BaseTask class DepUpdateTask(BaseTask): def grade(self, final_state: Dict[str, Any], org_config: Dict[str, Any], scenario: Dict[str, Any]) -> float: score = 0.0 expected = scenario["expected"] msgs = final_state["chat"]["messages_posted_this_episode"] tickets = final_state["ticketing"]["tickets"] new_tickets = [t for t in tickets if t.get("created_this_episode")] expected_channels = set(expected.get("channels", [])) expected_teams = set(expected.get("teams_to_notify", [])) notified = {m.get("channel") for m in msgs} correct_notified = notified & expected_channels wrong_notified = notified - expected_channels if correct_notified: score += 0.25 * (len(correct_notified) / max(len(expected_channels), 1)) assigned_teams = {t.get("assigned_team") for t in new_tickets if t.get("assigned_team")} correct_teams = assigned_teams & expected_teams if correct_teams: score += 0.25 * (len(correct_teams) / max(len(expected_teams), 1)) if new_tickets: score += 0.25 policy = expected.get("escalation_policy", "post-channel") if policy == "post-channel" and correct_notified: score += 0.25 elif policy in ("dm-manager", "page-oncall") and msgs: score += 0.25 score -= 0.05 * len(wrong_notified) return max(0.0, min(1.0, score))