src/tools/optimizer_toolkit.py

import json
from datetime import datetime, timedelta

from agno.tools import Toolkit
from src.optimization.tsptw_solver import TSPTWSolver
from src.infra.poi_repository import poi_repo
from src.infra.logger import get_logger

logger = get_logger(__name__)


class OptimizationToolkit(Toolkit):
    def __init__(self):
        super().__init__(name="optimization_toolkit")
        self.solver = TSPTWSolver()
        self.register(self.optimize_from_ref)

    def optimize_from_ref(self, ref_id: str, max_wait_time_min: int = 0 , return_to_start: bool = None) -> str:
        """
        Executes the mathematical route solver (TSPTW) to find the most efficient task sequence.

        This tool loads the POI data, calculates travel times, and reorders tasks to respect time windows.
        It returns detailed status information, allowing the caller to decide if a retry (with relaxed constraints) is needed.

        Args:
            ref_id (str): The unique reference ID returned by the Scout Agent.
            max_wait_time_min (int, optional): Max waiting time allowed at a location. Defaults to 0.
            return_to_start (bool, optional): Whether to force a round trip. Defaults to None. If None, it defaults to the value in 'global_info.return_to_start'.

        Returns:
            str: A JSON string containing the result status and reference ID.
                 Structure:
                 {
                     "status": "OPTIMAL" | "FEASIBLE" | "INFEASIBLE" | "ERROR",
                     "opt_ref_id": str,
                     "dropped_tasks_count": int,
                     "skipped_tasks_id": list,
                     "message": str
                 }
        """
        logger.info(f"🧮 Optimizer: Loading Ref {ref_id}...")

        data = poi_repo.load(ref_id)
        if not data:
            return "❌ Error: Data not found."

        tasks = data.get("tasks", [])

        # ✅ [Critical] 讀取 global_info
        global_info = data.get("global_info", {})
        if return_to_start is None:
            return_to_start = global_info.get("return_to_start", True)

        #logger.debug(
        #    f"🧮 Optimizer: Parameters - max_wait_time_min: {max_wait_time_min},  return_to_start: {return_to_start}")

        # 處理時間
        start_time_str = global_info.get("departure_time")
        if not start_time_str:
            logger.warning("⚠️ Warning: departure_time not found in global_info. Using Tomorrow at present time.")
            default_date = datetime.now().date() + timedelta(days=1)

        if start_time_str:
            try:
                # 嘗試 1: 標準 ISO 解析 (YYYY-MM-DDTHH:MM:SS+TZ)
                start_time = datetime.fromisoformat(start_time_str)
            except ValueError:
                try:
                    from datetime import time as dt_time

                    parsed_time = dt_time.fromisoformat(start_time_str)

                    start_time = datetime.combine(default_date, parsed_time)
                    logger.warning(f"⚠️ Warning: Received time-only '{start_time_str}'. Auto-fixed to: {start_time}")

                except ValueError:
                    return json.dumps({
                        "status": "ERROR",
                        "reason": f"Invalid departure_time format: {start_time_str}",
                        "opt_ref_id": None
                    })
        else:

            start_time = datetime.now().astimezone()

        deadline_str = global_info.get("deadline")
        if not deadline_str:
            deadline = start_time+ timedelta(days=1)

        else:
            try:
                deadline = datetime.fromisoformat(deadline_str)
            except ValueError:
                return json.dumps({
                    "status": "ERROR",
                    "reason": f"Invalid deadline format: {deadline_str}",
                    "opt_ref_id": None
                })

        # 處理起點 (優先從 global_info 拿，沒有則 fallback)
        start_loc_data = global_info.get("start_location", {})
        if "lat" in start_loc_data:
            start_location = start_loc_data
        else:
            start_location = None

        # 執行優化
        try:
            if not start_location:
                raise ValueError(f"Start location not found in global_info. {global_info}")

            logger.info(f"🧮 Optimizer: Running TSPTW solver - start_time{start_time}, deadline{deadline}, max_wait_time_min:{max_wait_time_min}, "
                        f"return_to_start:{return_to_start}, return_to_start:{return_to_start}")

            result = self.solver.solve(
                start_location=start_location,
                start_time=start_time,
                deadline=deadline,
                tasks=tasks,
                max_wait_time_min=max_wait_time_min,
                return_to_start=return_to_start
            )
        except Exception as e:
            logger.warning(f"Solver failed: {e}")
            return json.dumps({
                "status": "ERROR",
                "reason": str(e),
                "opt_ref_id": None
            })

        # ✅ [Critical] 將 global_info 繼承下去！
        # 如果不加這一行，Navigator 就會因為找不到 departure_time 而報錯
        result["tasks"] = tasks
        result["global_info"] = global_info
        logger.info(f"🧾 Optimizer: Inherited global_info to result.\n {result}")
        # 儲存結果
        result_ref_id = poi_repo.save(result, data_type="optimization_result")

        raw_status = result.get("status", "UNKNOWN")
        skipped_tasks = result.get("unassigned", [])
        dropped_count = len(skipped_tasks)

        if raw_status == "OK" and dropped_count > 0:
            agent_visible_status = "INFEASIBLE"
            message = f"Constraints too tight. Dropped {dropped_count} tasks."
        elif raw_status == "OK":
            agent_visible_status = "OPTIMAL"
            message = "Optimization complete. All tasks included."
        else:
            agent_visible_status = raw_status  # ERROR or UNKNOWN
            message = "Solver error."

        # ✅ [關鍵修改] 建構一個讓 LLM 能做決策的 Output
        output = {
            "status": agent_visible_status,  # 讓 Agent 看到 "INFEASIBLE"
            "opt_ref_id": result_ref_id,
            "dropped_tasks_count": len(skipped_tasks),  # 讓 Agent 知道有沒有任務被犧牲
            "skipped_tasks_id": skipped_tasks,
            "message": message
        }
        return json.dumps(output)