# command_handler.py
import pandas as pd
from linebot.v3.messaging import TextMessage, ImageMessage

# 匯入服務函式
from cwa_service import fetch_cwa_alarm_list, fetch_significant_earthquakes
from usgs_service import fetch_global_last24h_text, fetch_taiwan_df_this_year
from plotting_service import create_and_save_map
from ai_service import generate_ai_text
from weather_service import fetch_forecast_by_location
from config import CURRENT_YEAR, HF_SPACE_URL

def get_help_message() -> TextMessage:
    """回傳包含所有可用指令的說明訊息。"""
    text = (
        "📖 指令列表\n\n"
        "【地震資訊】\n"
        "• /global - 查詢全球近24小時顯著地震\n"
        "• /taiwan - 顯示今年台灣顯著地震列表\n"
        "• /map - 繪製台灣顯著地震地圖\n"
        "• /alert - 獲取 CWA 最新地震預警\n"
        "• /significant - CWA 最新顯著有感地震\n\n"
        "【AI 與工具】\n"
        "• /ai <問題> - 與 AI 助理對話\n"
        "• /weather <地點> - 查詢天氣預報\n\n"
        "【基本指令】\n"
        "• /info - 關於此機器人\n"
        "• /help - 顯示此說明"
    )
    return TextMessage(text=text)

def get_info_message() -> TextMessage:
    """回傳機器人資訊。"""
    text = (
        "🤖 關於我\n\n"
        "我是一個多功能助理機器人，提供地震查詢、天氣資訊與 AI 對話功能。\n\n"
        "• 版本: 2.0\n"
        "• 資料來源: CWA, USGS, Hugging Face\n"
        "• 開發者: dayichen"
    )
    return TextMessage(text=text)

def get_taiwan_earthquake_list() -> TextMessage:
    """回傳近期的台灣地震文字列表。"""
    result = fetch_taiwan_df_this_year()
    if isinstance(result, pd.DataFrame):
        count = len(result)
        lines = [f"🇹🇼 今年 ({CURRENT_YEAR} 年) 台灣區域顯著地震 (M≥5.0)，共 {count} 筆:", "-" * 20]
        for _, row in result.head(15).iterrows():
            t = row["time_utc"].strftime("%Y-%m-%d %H:%M")
            lines.append(f"震級: {row['magnitude']:.1f} | 日期時間: {t} (UTC)\n地點: {row['place']}")
        if count > 15:
            lines.append(f"... (還有 {count - 15} 筆，請用 /map 查看全部)")
        reply_text = "\n\n".join(lines)
    else:
        reply_text = result
    return TextMessage(text=reply_text)

def get_taiwan_earthquake_map(base_url: str) -> list:
    """產生並回傳台灣地震地圖的訊息。"""
    result = fetch_taiwan_df_this_year()
    if isinstance(result, pd.DataFrame):
        filename = create_and_save_map(result)
        image_url = f"{(HF_SPACE_URL or base_url)}/static/{filename}"
        return [
            TextMessage(text="🗺️ 已為您繪製今年台灣區域 M≥5.0 地震分佈圖（UTC）。"),
            ImageMessage(original_content_url=image_url, preview_image_url=image_url),
        ]
    else:
        return [TextMessage(text=result)]

def process_message(user_message_raw: str, request_base_url: str) -> list:
    """處理使用者的文字訊息並回傳一個包含回覆訊息的列表。"""
    # [修正] 加上這一行來定義 user_message 變數
    user_message = (user_message_raw or "").strip()
    
    # 將舊指令對應到新指令，提供向下相容
    if user_message.lower() in ["地震", "quake"]:
        user_message = "/global"
    elif user_message.lower() in ["台灣地震", "臺灣地震"]:
        user_message = "/taiwan"
    elif user_message.lower() in ["台灣地震畫圖", "臺灣地震畫圖"]:
        user_message = "/map"
    elif user_message.lower() == "地震預警":
        user_message = "/alert"

    if user_message.startswith('/'):
        parts = user_message.split(' ', 1)
        command = parts[0].lower()
        arg = parts[1] if len(parts) > 1 else ""

        if command == '/help':
            return [get_help_message()]
        elif command == '/info':
            return [get_info_message()]
        elif command == '/global':
            return [TextMessage(text=fetch_global_last24h_text())]
        elif command == '/taiwan':
            return [get_taiwan_earthquake_list()]
        elif command == '/map':
            return get_taiwan_earthquake_map(request_base_url)
        elif command == '/alert':
            return [TextMessage(text=fetch_cwa_alarm_list(limit=5))]
        elif command == '/significant':
            return [TextMessage(text=fetch_significant_earthquakes(limit=5))]
        elif command == '/weather':
            if not arg:
                return [TextMessage(text="請輸入地點，例如：/weather 台北")]
            return [TextMessage(text=fetch_forecast_by_location(arg))]
        elif command == '/ai':
            if not arg:
                return [TextMessage(text="請輸入問題，例如：/ai 台灣最高的山是哪座？")]
            return [TextMessage(text=generate_ai_text(arg))]

    # 若非指令，則預設交給 AI 處理 (作為備援)
    return [TextMessage(text=generate_ai_text(user_message))]