app.py

import streamlit as st
import pandas as pd
import numpy as np
import requests
import time
from collections import defaultdict
import json
import os
import re
from datetime import datetime, timedelta, time as dt_time
import io
import warnings
import matplotlib.pyplot as plt
import matplotlib.font_manager as font_manager
from matplotlib.lines import Line2D
from matplotlib.backends.backend_pdf import PdfPages
import matplotlib.gridspec as gridspec
import base64
import math
from pypinyin import lazy_pinyin, Style
from itertools import combinations
# --- 新增：加载环境变量 ---
from dotenv import load_dotenv
load_dotenv()  # 加载本地 .env 文件


# --- 全局配置和常量 ---
TOKEN_FILE = 'token_data.json'
# --- 环境变量获取 (替代硬编码) ---
# 使用 os.getenv 获取，如果获取不到默认为空字符串或特定默认值
GAODE_API_KEY = os.getenv("GAODE_API_KEY", "")
ADCODE = os.getenv("ADCODE")
CINEMA_ID = os.getenv("CINEMA_ID")

# --- 打印功能相关常量 ---
BUSINESS_START = "09:30"
BUSINESS_END = "01:30"
BORDER_COLOR = 'grey'
DATE_COLOR = '#A9A9A9'
A5_WIDTH_IN = 5.83
A5_HEIGHT_IN = 8.27
NUM_COLS = 3

# --- 打印字体清单 ---
ALL_FONTS = {
    "思源黑体-常规 (推荐 LED 屏)": "SimHei.ttf",
    "思源黑体-重体 (推荐散场表)": "SourceHanSansOLD-Heavy-2.otf",
    "思源黑体-粗体": "SourceHanSansOLD-Bold-2.otf",
    "思源宋体-常规": "SourceHanSansCN-Normal.otf",
    "苹方-中黑": "PingFangSC-Medium.otf",
    "苹方-半粗": "PingFangSC-Semibold.otf",
    "苹方-极细": "PingFangSC-Ultralight.otf",
    "阿里巴巴普惠体-常规": "Alibaba-PuHuiTi.ttf",
    "阿里巴巴普惠体-粗体": "AlibabaPuHuiTi-Bold.otf",
    "阿里巴巴普惠体-重体": "AlibabaPuHuiTi-Heavy.otf",
}
# 检查可用字体
AVAILABLE_FONTS = {name: fname for name, fname in ALL_FONTS.items() if os.path.exists(fname)}

# --- 忽略特定警告 ---
# 忽略 openpyxl 的样式警告
warnings.filterwarnings("ignore", category=UserWarning, module="openpyxl")
# 忽略 pandas 日期解析的警告 (针对无法推断格式的情况)
warnings.filterwarnings("ignore", message="Could not infer format")

# --- 页面基础设置 ---
st.set_page_config(layout="wide", page_title="影城工作便捷工具")


# --- 1. API 数据获取模块 ---

# --- 1.1 Token 管理 ---
def save_token(token_data):
    """将Token数据保存到JSON文件"""
    try:
        with open(TOKEN_FILE, 'w', encoding='utf-8') as f:
            json.dump(token_data, f, ensure_ascii=False, indent=4)
        return True
    except Exception as e:
        st.error(f"保存Token失败: {e}")
        return False


def load_token():
    """从JSON文件加载Token数据"""
    if os.path.exists(TOKEN_FILE):
        try:
            with open(TOKEN_FILE, 'r', encoding='utf-8') as f:
                return json.load(f)
        except (json.JSONDecodeError, FileNotFoundError):
            return None
    return None


def login_and_get_token():
    """执行登录操作并获取新的Token"""
    st.write("Token无效或已过期，正在尝试重新登录...")


    # 获取环境变量
    username = os.getenv("CINEMA_USERNAME")
    password = os.getenv("CINEMA_PASSWORD")
    res_code = os.getenv("CINEMA_RES_CODE")
    device_id = os.getenv("CINEMA_DEVICE_ID")

    # 简单检查，防止未配置环境变量导致后续请求莫名报错
    if not all([username, password, res_code]):
        st.error("登录失败：未配置用户名、密码或影院编码环境变量。")
        return None

    session = requests.Session()
    session.headers.update({
        'Host': 'app.bi.piao51.cn',
        'Accept': 'application/json, text/javascript, */*; q=0.01',
        'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148',
    })

    login_url = 'https://app.bi.piao51.cn/cinema-app/credential/login.action'
    login_headers = {
        'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
        'Origin': 'https://app.bi.piao51.cn',
    }
    # 使用变量
    login_data = {
        'username': username,
        'password': password,
        'type': '1',
        'resCode': res_code,
        'deviceid': device_id,
        'dtype': 'ios',
    }

    try:
        response_login = session.post(login_url, headers=login_headers, data=login_data, allow_redirects=False,
                                      timeout=15)
        if not (300 <= response_login.status_code < 400 and 'token' in session.cookies):
            st.error(f"登录步骤 1 失败，未能获取 Session Token。状态码: {response_login.status_code}")
            return None

        user_info_url = 'https://app.bi.piao51.cn/cinema-app/security/logined.action'
        response_user_info = session.get(user_info_url, timeout=10)
        response_user_info.raise_for_status()

        user_info = response_user_info.json()
        if user_info.get("success") and user_info.get("data", {}).get("token"):
            token_data = user_info['data']
            if save_token(token_data): st.toast("登录成功，已获取并保存新 Token！", icon="🔑")
            return token_data
        else:
            st.error(f"登录步骤 2 失败，未能从 JSON 中提取 Token。响应: {user_info.get('msg')}")
            return None

    except requests.exceptions.RequestException as e:
        st.error(f"登录请求过程中发生网络错误: {e}")
        return None


# --- 1.2 API 数据抓取 (排片相关) ---
def fetch_hall_info(token):
    url = 'https://cawapi.yinghezhong.com/showInfo/getShowHallInfo'
    params = {'token': token, '_': int(time.time() * 1000)}
    headers = {'Origin': 'https://caw.yinghezhong.com', 'User-Agent': 'Mozilla/5.0'}
    response = requests.get(url, params=params, headers=headers, timeout=10)
    response.raise_for_status()
    data = response.json()
    if data.get('code') == 1 and data.get('data'):
        return {item['hallId']: item['seatNum'] for item in data['data']}
    else:
        raise Exception(f"获取影厅信息失败: {data.get('msg', '未知错误')}")


def fetch_schedule_data(token, show_date):
    url = 'https://cawapi.yinghezhong.com/showInfo/getHallShowInfo'
    params = {'showDate': show_date, 'token': token, '_': int(time.time() * 1000)}
    headers = {'Origin': 'https://caw.yinghezhong.com', 'User-Agent': 'Mozilla/5.0'}
    response = requests.get(url, params=params, headers=headers, timeout=15)
    response.raise_for_status()
    data = response.json()
    if data.get('code') == 1:
        return data.get('data', [])
    elif data.get('code') == 500:
        raise ValueError("Token 可能已失效")
    else:
        raise Exception(f"获取排片数据失败: {data.get('msg', '未知错误')}")


def get_api_data_with_token_management(show_date):
    token_data = load_token()
    token = token_data.get('token') if token_data else None
    if not token:
        token_data = login_and_get_token()
        if not token_data: return None, None
        token = token_data.get('token')

    try:
        schedule_list = fetch_schedule_data(token, show_date)
        hall_seat_map = fetch_hall_info(token)
        return schedule_list, hall_seat_map
    except ValueError:
        st.toast("Token 已失效，正在尝试重新登录并重试...", icon="🔄")
        token_data = login_and_get_token()
        if not token_data: return None, None
        token = token_data.get('token')
        try:
            schedule_list = fetch_schedule_data(token, show_date)
            hall_seat_map = fetch_hall_info(token)
            return schedule_list, hall_seat_map
        except Exception as e:
            st.error(f"重试获取数据失败: {e}");
            return None, None
    except Exception as e:
        st.error(f"获取 API 数据时发生错误: {e}");
        return None, None


# --- 1.3 新增：电影名称API (获取标准电影名) ---
@st.cache_data(show_spinner=False, ttl=600)
def fetch_canonical_movie_names(token, date_str):
    """
    获取指定日期的官方电影名称列表（唯一名称）。
    用于后续对原始排片数据中的电影名进行标准化清洗。
    """
    url = 'https://app.bi.piao51.cn/cinema-app/mycinema/movieSellGross.action'
    params = {
        'token': token,
        'startDate': date_str,
        'endDate': date_str,
        'dateType': 'day',
        'cinemaId': CINEMA_ID
    }
    headers = {
        'Host': 'app.bi.piao51.cn', 'X-Requested-With': 'XMLHttpRequest', 'jwt': '0',
        'Accept': 'application/json, text/javascript, */*; q=0.01',
        'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148',
    }

    try:
        response = requests.get(url, params=params, headers=headers, timeout=10)
        response.raise_for_status()
        data = response.json()
        if data.get('code') == 'A00000' and data.get('results'):
            # 提取 results 列表中的 movieName，排除 "总计" 等非电影项
            names = [item['movieName'] for item in data['results'] if
                     item.get('movieName') and item['movieName'] != '总计']
            return names
    except Exception as e:
        # 这里的错误不打断主流程，返回空列表，后续会回退到基础清洗逻辑
        print(f"获取标准电影名称失败: {e}")
    return []


def process_api_data(schedule_list, hall_seat_map, token=None, show_date=None):
    if not schedule_list:
        st.warning("未获取到任何排片数据。");
        return pd.DataFrame()
    df = pd.DataFrame(schedule_list)
    df['座位数'] = df['hallId'].map(hall_seat_map).fillna(0).astype(int)
    df.rename(columns={'movieName': '影片名称', 'showStartTime': '放映时间', 'soldBoxOffice': '总收入',
                       'soldTicketNum': '总人次'}, inplace=True)

    # 获取标准电影名列表并进行清洗
    canonical_names = []
    if token and show_date:
        canonical_names = fetch_canonical_movie_names(token, show_date)

    df['影片名称'] = df['影片名称'].apply(lambda x: clean_movie_title(x, canonical_names))

    required_cols = ['影片名称', '放映时间', '座位数', '总收入', '总人次']
    df = df[required_cols]
    df.dropna(subset=['影片名称', '放映时间'], inplace=True)
    for col in ['座位数', '总收入', '总人次']:
        df[col] = pd.to_numeric(df[col], errors='coerce').fillna(0)
    df['放映时间'] = pd.to_datetime(df['放映时间'], format='%H:%M', errors='coerce').dt.time
    df.dropna(subset=['放映时间'], inplace=True)
    return df


# --- 1.4 API 数据抓取 (销售相关) ---
def fetch_sales_data_from_api(token, selected_date):
    """从API获取指定日期的销售数据"""
    url = 'https://app.bi.piao51.cn/cinema-app/mycinema/retailTop.action'
    date_str = selected_date.strftime('%Y-%m-%d')
    headers = {
        'Host': 'app.bi.piao51.cn', 'X-Requested-With': 'XMLHttpRequest', 'jwt': '0',
        'Accept': 'application/json, text/javascript, */*; q=0.01',
        'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_7 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148',
    }
    params = {
        'dateType': 'day', 'startDate': date_str, 'endDate': date_str, 'noEvent': '1',
        'token': token, 'qTime': date_str, 'cinemaId': CINEMA_ID,
    }
    try:
        response = requests.get(url, params=params, headers=headers, timeout=15)
        response.raise_for_status()
        data = response.json()
        if data.get('code') == 'A00000':
            return data.get('results', [])
        elif "login" in response.text:
            raise ValueError("Token可能已失效")
        else:
            st.error(f"获取 API 数据失败: {data.get('msg', '未知错误')}")
            return None
    except requests.exceptions.RequestException as e:
        st.error(f"API 请求网络错误: {e}")
        return None


def get_sales_data_with_token_management(selected_date):
    """带Token管理的API销售数据获取流程"""
    token_data = load_token()
    token = token_data.get('token') if token_data else None
    if not token:
        token_data = login_and_get_token()
        if not token_data: return None
        token = token_data.get('token')

    try:
        api_results = fetch_sales_data_from_api(token, selected_date)
        return api_results
    except ValueError:
        st.toast("Token 已失效，正在尝试重新登录并重试...", icon="🔄")
        token_data = login_and_get_token()
        if not token_data: return None
        token = token_data.get('token')
        try:
            api_results = fetch_sales_data_from_api(token, selected_date)
            return api_results
        except Exception as e:
            st.error(f"重试获取数据失败: {e}")
            return None
    except Exception as e:
        st.error(f"获取 API 数据时发生错误: {e}")
        return None


# --- 2. 核心数据分析模块 ---
def clean_movie_title(raw_title, canonical_names=None):
    """
    电影名称标准化清洗函数
    """
    if not isinstance(raw_title, str):
        return raw_title

    base_name = None

    # 1. 尝试匹配标准名称
    if canonical_names:
        # 按长度倒序排序，确保最长匹配优先（解决"你好"vs"你好明天"的问题）
        sorted_names = sorted(canonical_names, key=len, reverse=True)
        for name in sorted_names:
            if name in raw_title:
                base_name = name
                break

    # 2. 回退逻辑：如果没传列表或没匹配到，使用空格分割
    if not base_name:
        base_name = raw_title.split(' ', 1)[0]

    # 3. 后缀追加逻辑
    raw_upper = raw_title.upper()
    suffix = ""

    if "HDR LED" in raw_upper:
        suffix = "（HDR LED）"
    elif "CINITY" in raw_upper:
        suffix = "（CINITY）"
    elif "杜比" in raw_upper or "DOLBY" in raw_upper:
        suffix = "（杜比视界）"
    elif "IMAX" in raw_upper:
        if "3D" in raw_upper:
            suffix = "（数字IMAX3D）"
        else:
            suffix = "（数字IMAX）"
    elif "巨幕" in raw_upper:
        if "立体" in raw_upper:
            suffix = "（中国巨幕立体）"
        else:
            suffix = "（中国巨幕）"
    elif "3D" in raw_upper:
        suffix = "（数字3D）"

    # **修复**: 只有当 base_name 自身不包含该后缀时才添加
    if suffix and suffix not in base_name:
        return f"{base_name}{suffix}"

    return base_name


def style_efficiency(row):
    green, red = 'background-color: #E6F5E6;', 'background-color: #FFE5E5;'
    seat_eff, session_eff = row.get('座次效率', 0), row.get('场次效率', 0)
    if seat_eff > 1.5 or session_eff > 1.5: return [green] * len(row)
    if seat_eff < 0.5 or session_eff < 0.5: return [red] * len(row)
    return [''] * len(row)


def style_summary_efficiency(row):
    green, red = 'background-color: #E6F5E6;', 'background-color: #FFE5E5;'
    if (row.get('全部座次效率', 0) > 1.5 or row.get('全部场次效率', 0) > 1.5 or
            row.get('黄金时段座次效率', 0) > 1.5 or row.get('黄金时段场次效率', 0) > 1.5):
        return [green] * len(row)
    if (row.get('全部座次效率', 0) < 0.5 or row.get('全部场次效率', 0) < 0.5 or
            row.get('黄金时段座次效率', 0) < 0.5 or row.get('黄金时段场次效率', 0) < 0.5):
        return [red] * len(row)
    return [''] * len(row)


def process_and_analyze_data(df):
    if df.empty: return pd.DataFrame()

    # 确保有清洗后的列名
    if '影片名称_清理后' not in df.columns and '影片名称' in df.columns:
        df['影片名称_清理后'] = df['影片名称']

    analysis_df = df.groupby('影片名称_清理后').agg(座位数=('座位数', 'sum'), 场次=('影片名称_清理后', 'size'),
                                                    票房=('总收入', 'sum'), 人次=('总人次', 'sum')).reset_index()
    analysis_df.rename(columns={'影片名称_清理后': '影片'}, inplace=True)
    analysis_df = analysis_df.sort_values(by='票房', ascending=False).reset_index(drop=True)
    total_seats, total_sessions, total_revenue = analysis_df['座位数'].sum(), analysis_df['场次'].sum(), analysis_df[
        '票房'].sum()
    with np.errstate(divide='ignore', invalid='ignore'):
        analysis_df['均价'] = np.divide(analysis_df['票房'], analysis_df['人次']).fillna(0)
        analysis_df['座次比'] = np.divide(analysis_df['座位数'], total_seats).fillna(0)
        analysis_df['场次比'] = np.divide(analysis_df['场次'], total_sessions).fillna(0)
        analysis_df['票房比'] = np.divide(analysis_df['票房'], total_revenue).fillna(0)
        analysis_df['座次效率'] = np.divide(analysis_df['票房比'], analysis_df['座次比']).fillna(0)
        analysis_df['场次效率'] = np.divide(analysis_df['票房比'], analysis_df['场次比']).fillna(0)
    final_cols = ['影片', '座位数', '场次', '票房', '人次', '均价', '座次比', '场次比', '票房比', '座次效率',
                  '场次效率']
    return analysis_df[final_cols]


# --- 2.1 销售数据处理与分析模块 ---
def transform_api_data_to_df(api_results):
    """将API返回的JSON列表转换为与Excel格式一致的DataFrame"""
    if not api_results: return pd.DataFrame()
    records = []
    for item in api_results:
        is_package = item.get('goodsAllName') and str(item.get('goodsAllName')).strip() != ""
        record = {
            '售卖位置': '线上渠道',
            '一级分类': '套餐' if is_package else '单品',
            '售卖键名称': item.get('goodsName'),
            '数量': item.get('goodsSoldNums', 0),
            '实收总金额': item.get('goodsSoldIncomes', 0)
        }
        records.append(record)
    return pd.DataFrame(records)


def process_sales_data(df):
    """核心分析函数，处理DataFrame并返回最终结果"""
    if df.empty:
        st.warning("没有可供分析的数据。")
        return None, ""
    required_columns = ['售卖位置', '一级分类', '售卖键名称', '数量', '实收总金额']
    if not all(col in df.columns for col in required_columns):
        missing_cols = [col for col in required_columns if col not in df.columns]
        st.error(f"数据缺少必要的列: {', '.join(missing_cols)}。")
        return None, ""
    for col in ['数量', '实收总金额']:
        df[col] = pd.to_numeric(df[col], errors='coerce').fillna(0)
    df.dropna(subset=['售卖键名称', '一级分类'], inplace=True)
    df['售卖位置'] = df['售卖位置'].fillna('未知渠道')

    df_package = df[df['一级分类'] == '套餐'].copy()
    df_non_package = df[df['一级分类'] != '套餐'].copy()

    if not df_package.empty:
        package_summary = df_package.groupby(['售卖位置', '售卖键名称']).agg(
            {'数量': 'sum', '实收总金额': 'sum'}).reset_index()
        package_summary = package_summary[package_summary['数量'] != 0]
        top_10_package = package_summary.sort_values(by='实收总金额', ascending=False).head(10)[
            ['售卖键名称', '数量', '实收总金额']]
    else:
        top_10_package = pd.DataFrame(columns=['售卖键名称', '数量', '实收总金额'])

    if not df_non_package.empty:
        non_package_summary = df_non_package.groupby(['售卖位置', '售卖键名称']).agg(
            {'数量': 'sum', '实收总金额': 'sum'}).reset_index()
        non_package_summary = non_package_summary[non_package_summary['数量'] != 0]
        top_5_non_package = non_package_summary.sort_values(by='实收总金额', ascending=False).head(5)[
            ['售卖键名称', '数量', '实收总金额']]
    else:
        top_5_non_package = pd.DataFrame(columns=['售卖键名称', '数量', '实收总金额'])

    summary_df = pd.concat([top_10_package, top_5_non_package], ignore_index=True)
    final_display_df = pd.DataFrame(index=range(15), columns=['售卖键名称', '数量', '实收总金额'])
    final_display_df['售卖键名称'] = ''
    final_display_df['数量'] = np.nan
    final_display_df['实收总金额'] = np.nan

    if not summary_df.empty:
        final_display_df.iloc[:len(summary_df)] = summary_df.to_numpy()

    copy_df = final_display_df.copy()
    # **修改**: 先转换为 object 类型再填充，避免 FutureWarning
    copy_df = copy_df.astype(object)
    copy_df.fillna('', inplace=True)
    copy_text = copy_df.to_csv(sep='\t', index=False, header=False)
    return final_display_df, copy_text


# --- 2.2 影片映出日累计报表数据处理模块 ---
def process_and_filter_data_for_report(schedule_list, hall_seat_map, selected_date_str, token=None):
    """核心数据处理函数 for 影片映出日累计报表"""
    if not schedule_list:
        st.warning("未获取到任何排片数据。")
        return pd.DataFrame()

    df = pd.DataFrame(schedule_list)
    # 过滤掉观影人数为0的场次
    df['soldTicketNum'] = pd.to_numeric(df['soldTicketNum'], errors='coerce').fillna(0)
    df = df[df['soldTicketNum'] > 0].copy()

    if df.empty:
        st.info("所有场次的观影人数均为 0，没有可显示的数据。")
        return pd.DataFrame()

    # 获取标准电影名并清洗
    canonical_names = []
    if token and selected_date_str:
        canonical_names = fetch_canonical_movie_names(token, selected_date_str)

    df['影片'] = df['movieName'].apply(lambda x: clean_movie_title(x, canonical_names))
    df['座位数'] = df['hallId'].map(hall_seat_map).fillna(0).astype(int)

    # 计算上座率
    with np.errstate(divide='ignore', invalid='ignore'):
        df['上座率%'] = np.divide(df['soldTicketNum'], df['座位数']) * 100
        df['上座率%'] = df['上座率%'].fillna(0)

    df.rename(columns={'showStartTime': '放映时间', 'hallName': '影厅', 'soldTicketNum': '人数合计'}, inplace=True)
    df['放映日期'] = selected_date_str
    final_cols = ['影片', '放映日期', '放映时间', '影厅', '人数合计', '座位数', '上座率%']
    result_df = df[final_cols]
    result_df = result_df.sort_values(by='放映时间').reset_index(drop=True)

    return result_df


# --- 2.3 打印功能数据处理与布局模块 ---

def get_font_properties(font_path, size=14):
    """通用字体加载函数"""
    if font_path and os.path.exists(font_path):
        return font_manager.FontProperties(fname=font_path, size=size)
    else:
        st.warning(f"警告：未找到字体文件 '{font_path}'，显示可能不正确。将使用默认字体。")
        return font_manager.FontProperties(family='sans-serif', size=size)


def get_pinyin_abbr(text):
    """获取中文文本前两个字的拼音首字母"""
    if not text: return ""
    chars = [c for c in text if '\u4e00' <= c <= '\u9fff'][:2]
    if not chars: return ""
    pinyin_list = lazy_pinyin(chars, style=Style.FIRST_LETTER)
    return ''.join(pinyin_list).upper()


def format_seq(n):
    """将数字或字符转换为带圈序号 (①, ②, ③...)，非数字则直接返回"""
    try:
        n = int(n)
    except (ValueError, TypeError):
        return str(n)
    if n <= 0: return str(n)
    circled_chars = "①②③④⑤⑥⑦⑧⑨⑩⑪⑫⑬⑭⑮⑯⑰⑱⑲⑳㉑㉒㉓㉔㉕㉖㉗㉘㉙㉚㉛㉜㉝㉞㉟㊱㊲㊳㊴㊵㊶㊷㊸㊹㊺㊻㊼㊽㊾㊿"
    if 1 <= n <= 50: return circled_chars[n - 1]
    return f'({n})'


def process_schedule_df(df, base_date, split_time_str, time_adjustment_minutes=0):
    """
    处理排片DataFrame，生成LED屏数据和散场时间数据
    """
    if df is None or df.empty:
        return None, None, None

    # 'LED屏排片表' 数据处理
    led_df = df.copy()
    try:
        # 优先提取 'X号'，失败则取第一个字符 + '号'
        extracted = led_df['Hall'].astype(str).str.extract(r'(\d+号)')
        fallback = led_df['Hall'].astype(str).str[0] + '号'
        led_df['Hall'] = extracted[0].fillna(fallback)

        led_df['StartTime_dt'] = pd.to_datetime(led_df['StartTime'], format='%H:%M', errors='coerce').apply(
            lambda t: t.replace(year=base_date.year, month=base_date.month, day=base_date.day) if pd.notnull(t) else t)
        led_df['EndTime_dt'] = pd.to_datetime(led_df['EndTime'], format='%H:%M', errors='coerce').apply(
            lambda t: t.replace(year=base_date.year, month=base_date.month, day=base_date.day) if pd.notnull(t) else t)
        led_df.loc[led_df['EndTime_dt'] < led_df['StartTime_dt'], 'EndTime_dt'] += timedelta(days=1)
        led_df = led_df.sort_values(['Hall', 'StartTime_dt'])
        merged_rows = []
        for _, group in led_df.groupby('Hall'):
            current = None
            for _, row in group.sort_values('StartTime_dt').iterrows():
                if current is None:
                    current = row.copy()
                elif row['Movie'] == current['Movie']:
                    current['EndTime_dt'] = row['EndTime_dt']
                else:
                    merged_rows.append(current)
                    current = row.copy()
            if current is not None: merged_rows.append(current)
        merged_df = pd.DataFrame(merged_rows)
        merged_df['StartTime_dt'] -= timedelta(minutes=10)
        merged_df['EndTime_dt'] -= timedelta(minutes=5)
        merged_df['Seq'] = merged_df.groupby('Hall').cumcount() + 1
        merged_df['StartTime_str'] = merged_df['StartTime_dt'].dt.strftime('%H:%M')
        merged_df['EndTime_str'] = merged_df['EndTime_dt'].dt.strftime('%H:%M')
        led_schedule_df = merged_df[['Hall', 'Seq', 'Movie', 'StartTime_str', 'EndTime_str']]
    except Exception as e:
        st.error(f"处理 'LED 屏打印表' 数据时出错: {e}")
        led_schedule_df = None

    # '散场时间快捷打印' 数据处理
    times_df = df.copy()
    try:
        # 优先提取数字，失败则取第一个字符
        num_part = times_df['Hall'].str.extract(r'(\d+)')[0]
        char_part = times_df['Hall'].astype(str).str[0]
        times_df['Hall'] = num_part.fillna(char_part)
        times_df.dropna(subset=['Hall', 'StartTime', 'EndTime'], inplace=True)
        times_df['StartTime_dt'] = pd.to_datetime(times_df['StartTime'], format='%H:%M', errors='coerce').apply(
            lambda t: datetime.combine(base_date, t.time()) if pd.notnull(t) else pd.NaT)
        times_df['EndTime_dt'] = pd.to_datetime(times_df['EndTime'], format='%H:%M', errors='coerce').apply(
            lambda t: datetime.combine(base_date, t.time()) if pd.notnull(t) else pd.NaT)
        times_df.loc[times_df['EndTime_dt'] < times_df['StartTime_dt'], 'EndTime_dt'] += timedelta(days=1)

        # 应用时间提前量
        if time_adjustment_minutes > 0:
            times_df['EndTime_dt'] -= timedelta(minutes=time_adjustment_minutes)

        business_start_dt = datetime.combine(base_date, datetime.strptime(BUSINESS_START, "%H:%M").time())
        business_end_dt = datetime.combine(base_date, datetime.strptime(BUSINESS_END, "%H:%M").time())
        if business_end_dt < business_start_dt: business_end_dt += timedelta(days=1)
        times_df = times_df[(times_df['EndTime_dt'] >= business_start_dt) & (times_df['EndTime_dt'] <= business_end_dt)]
        times_df = times_df.sort_values('EndTime_dt')
        split_dt = datetime.combine(base_date, split_time_str)
        part1 = times_df[times_df['EndTime_dt'] <= split_dt].copy()
        part2 = times_df[times_df['EndTime_dt'] > split_dt].copy()

        # 使用 %H:%M 保证两位小时
        part1['EndTime'] = part1['EndTime_dt'].dt.strftime('%H:%M')
        part2['EndTime'] = part2['EndTime_dt'].dt.strftime('%H:%M')

        times_part1_df = part1[['Hall', 'EndTime']]
        times_part2_df = part2[['Hall', 'EndTime']]
    except Exception as e:
        st.error(f"处理 '散场时间表' 数据时出错: {e}")
        times_part1_df, times_part2_df = None, None

    return led_schedule_df, times_part1_df, times_part2_df


def process_file_upload(file, split_time_str, time_adjustment_minutes=0):
    """
    Handles file upload, reads excel and calls the core processing function.
    """
    try:
        date_df = pd.read_excel(file, header=None, skiprows=7, nrows=1, usecols=[3])
        date_str = pd.to_datetime(date_df.iloc[0, 0]).strftime('%Y-%m-%d')
        base_date = pd.to_datetime(date_str).date()
    except Exception:
        date_str = datetime.today().strftime('%Y-%m-%d')
        base_date = datetime.today().date()

    try:
        df = pd.read_excel(file, header=9, usecols=[1, 2, 4, 5])
        df.columns = ['Hall', 'StartTime', 'EndTime', 'Movie']
        df['Hall'] = df['Hall'].ffill()
        df.dropna(subset=['StartTime', 'EndTime', 'Movie'], inplace=True)
    except Exception as e:
        st.error(f"读取数据时出错: {e}。请检查文件格式是否为'放映时间核对表'。")
        return None, None, None, date_str

    led_data, times_p1, times_p2 = process_schedule_df(df, base_date, split_time_str, time_adjustment_minutes)
    return led_data, times_p1, times_p2, date_str


def create_print_layout_led(data, date_str, font_path, generate_png=False):
    """生成LED屏排片表的PDF/PNG"""
    if data is None or data.empty: return None
    A4_width_in, A4_height_in = 8.27, 11.69
    dpi = 300
    total_content_rows = len(data)
    layout_rows = max(total_content_rows, 25)
    totalA = layout_rows + 2
    row_height = A4_height_in / totalA
    data = data.reset_index(drop=True)
    data['hall_str'] = '$' + data['Hall'].str.replace('号', '') + '^{\\#}$'
    data['seq_str'] = data['Seq'].apply(format_seq)
    data['pinyin_abbr'] = data['Movie'].apply(get_pinyin_abbr)
    data['time_str'] = data['StartTime_str'] + ' - ' + data['EndTime_str']
    temp_fig = plt.figure(figsize=(A4_width_in, A4_height_in), dpi=dpi)
    renderer = temp_fig.canvas.get_renderer()
    base_font_size_pt = (row_height * 0.9) * 72
    seq_font_size_pt = (row_height * 0.5) * 72

    def get_col_width_in(series, font_size_pt, is_math=False):
        if series.empty: return 0
        font_prop = get_font_properties(font_path, font_size_pt)
        longest_str_idx = series.astype(str).str.len().idxmax()
        max_content = str(series.loc[longest_str_idx])
        text_width_px, _, _ = renderer.get_text_width_height_descent(max_content, font_prop, ismath=is_math)
        return (text_width_px / dpi) * 1.1

    margin_col_width = row_height
    hall_col_width = get_col_width_in(data['hall_str'], base_font_size_pt, is_math=True)
    seq_col_width = get_col_width_in(data['seq_str'], seq_font_size_pt)
    pinyin_col_width = get_col_width_in(data['pinyin_abbr'], base_font_size_pt)
    time_col_width = get_col_width_in(data['time_str'], base_font_size_pt)
    movie_col_width = A4_width_in - (
            margin_col_width * 2 + hall_col_width + seq_col_width + pinyin_col_width + time_col_width)
    plt.close(temp_fig)
    col_widths = {'hall': hall_col_width, 'seq': seq_col_width, 'movie': movie_col_width, 'pinyin': pinyin_col_width,
                  'time': time_col_width}
    col_x_starts = {}
    current_x = margin_col_width
    for col_name in ['hall', 'seq', 'movie', 'pinyin', 'time']:
        col_x_starts[col_name] = current_x
        current_x += col_widths[col_name]

    def draw_figure(fig, ax):
        renderer = fig.canvas.get_renderer()
        for col_name in ['hall', 'seq', 'movie', 'pinyin']:
            x_line = col_x_starts[col_name] + col_widths[col_name]
            line_top_y, line_bottom_y = A4_height_in - row_height, row_height
            ax.add_line(
                Line2D([x_line, x_line], [line_bottom_y, line_top_y], color='gray', linestyle=':', linewidth=0.5))
        last_hall_drawn = None
        for i, row in data.iterrows():
            y_bottom = A4_height_in - (i + 2) * row_height
            y_center = y_bottom + row_height / 2
            if row['Hall'] != last_hall_drawn:
                ax.text(col_x_starts['hall'] + col_widths['hall'] / 2, y_center, row['hall_str'],
                        fontproperties=get_font_properties(font_path, base_font_size_pt), ha='center', va='center')
                last_hall_drawn = row['Hall']
            ax.text(col_x_starts['seq'] + col_widths['seq'] / 2, y_center, row['seq_str'],
                    fontproperties=get_font_properties(font_path, seq_font_size_pt), ha='center', va='center')
            ax.text(col_x_starts['pinyin'] + col_widths['pinyin'] / 2, y_center, row['pinyin_abbr'],
                    fontproperties=get_font_properties(font_path, base_font_size_pt), ha='center', va='center')
            ax.text(col_x_starts['time'] + col_widths['time'] / 2, y_center, row['time_str'],
                    fontproperties=get_font_properties(font_path, base_font_size_pt), ha='center', va='center')
            movie_font_size = base_font_size_pt
            movie_font_prop = get_font_properties(font_path, movie_font_size)
            text_w_px, _, _ = renderer.get_text_width_height_descent(row['Movie'], movie_font_prop, ismath=False)
            text_w_in = text_w_px / dpi
            max_width_in = col_widths['movie'] * 0.9
            if text_w_in > max_width_in:
                movie_font_size *= (max_width_in / text_w_in)
                movie_font_prop = get_font_properties(font_path, movie_font_size)
            ax.text(col_x_starts['movie'] + 0.05, y_center, row['Movie'], fontproperties=movie_font_prop, ha='left',
                    va='center')
            is_last_in_hall = (i == len(data) - 1) or (row['Hall'] != data.loc[i + 1, 'Hall'])
            line_start_x, line_end_x = margin_col_width, A4_width_in - margin_col_width
            if is_last_in_hall:
                ax.add_line(Line2D([line_start_x, line_end_x], [y_bottom, y_bottom], color='black', linestyle='-',
                                   linewidth=0.8))
            else:
                ax.add_line(Line2D([line_start_x, line_end_x], [y_bottom, y_bottom], color='gray', linestyle=':',
                                   linewidth=0.5))

    outputs = {}
    fig = plt.figure(figsize=(A4_width_in, A4_height_in), dpi=300)
    ax = fig.add_axes([0, 0, 1, 1])
    ax.set_axis_off();
    ax.set_xlim(0, A4_width_in);
    ax.set_ylim(0, A4_height_in)
    ax.text(margin_col_width, A4_height_in - row_height, date_str, fontproperties=get_font_properties(font_path, 10),
            color=DATE_COLOR, ha='left', va='bottom', transform=ax.transData)
    draw_figure(fig, ax)
    pdf_buf = io.BytesIO()
    fig.savefig(pdf_buf, format='pdf', dpi=dpi, bbox_inches='tight', pad_inches=0)
    pdf_buf.seek(0)
    outputs['pdf'] = f"data:application/pdf;base64,{base64.b64encode(pdf_buf.getvalue()).decode()}"
    if generate_png:
        png_buf = io.BytesIO()
        fig.savefig(png_buf, format='png', dpi=dpi, bbox_inches='tight', pad_inches=0)
        png_buf.seek(0)
        outputs['png'] = f"data:image/png;base64,{base64.b64encode(png_buf.getvalue()).decode()}"
    plt.close(fig)
    return outputs


def create_print_layout_times(data, title, date_str, font_path, size_multiplier=1.1, hall_format='Default',
                              generate_png=False):
    """生成散场时间表的PDF/PNG"""
    if data is None or data.empty: return None

    def generate_figure():
        total_items = len(data)
        num_rows = math.ceil(total_items / NUM_COLS) if total_items > 0 else 1
        data_area_height_in, cell_width_in = A5_HEIGHT_IN, A5_WIDTH_IN / NUM_COLS
        cell_height_in = data_area_height_in / num_rows
        target_width_pt, target_height_pt = (cell_width_in * 0.9) * 72, (cell_height_in * 0.9) * 72
        font_size_based_on_width = target_width_pt / (8 * 0.6)
        base_fontsize = min(font_size_based_on_width, target_height_pt) * size_multiplier
        fig = plt.figure(figsize=(A5_WIDTH_IN, A5_HEIGHT_IN), dpi=300)
        fig.subplots_adjust(left=0, right=1, top=1, bottom=0)
        gs = gridspec.GridSpec(num_rows, NUM_COLS, hspace=0, wspace=0, figure=fig)
        data_values = data.values.tolist()
        while len(data_values) % NUM_COLS != 0: data_values.append(['', ''])
        rows_per_col_layout = math.ceil(len(data_values) / NUM_COLS)
        sorted_data = [['', '']] * len(data_values)
        for i, item in enumerate(data_values):
            if item[0] and item[1]:
                row_in_col, col_idx = i % rows_per_col_layout, i // rows_per_col_layout
                new_index = row_in_col * NUM_COLS + col_idx
                if new_index < len(sorted_data): sorted_data[new_index] = item
        is_first_cell_with_data = True
        for idx, (hall, end_time) in enumerate(sorted_data):
            if hall and end_time:
                row_grid, col_grid = idx // NUM_COLS, idx % NUM_COLS
                ax = fig.add_subplot(gs[row_grid, col_grid])
                for spine in ax.spines.values():
                    spine.set_visible(True);
                    spine.set_linestyle((0, (1, 2)))
                    spine.set_color(BORDER_COLOR);
                    spine.set_linewidth(0.75)
                if is_first_cell_with_data:
                    ax.text(0.05, 0.95, f"{date_str} {title}",
                            fontproperties=get_font_properties(font_path, base_fontsize * 0.5), color=DATE_COLOR,
                            ha='left', va='top', transform=ax.transAxes)
                    is_first_cell_with_data = False
                if hall_format == 'Superscript':
                    display_text = f'${str(hall)}^{{\\#}}$ {end_time}'
                elif hall_format == 'Circled':
                    display_text = f'{format_seq(hall)} {end_time}'
                else:  # Default
                    display_text = f"{str(hall)} {end_time}"
                ax.text(0.5, 0.5, display_text, fontproperties=get_font_properties(font_path, base_fontsize),
                        ha='center', va='center', transform=ax.transAxes)
                ax.set_xticks([]);
                ax.set_yticks([]);
                ax.set_facecolor('none')
        return fig

    fig_for_output = generate_figure()
    outputs = {}
    pdf_buffer = io.BytesIO()
    with PdfPages(pdf_buffer) as pdf:
        pdf.savefig(fig_for_output)
    pdf_buffer.seek(0)
    outputs['pdf'] = f"data:application/pdf;base64,{base64.b64encode(pdf_buffer.getvalue()).decode()}"
    if generate_png:
        png_buffer = io.BytesIO()
        fig_for_output.savefig(png_buffer, format='png')
        png_buffer.seek(0)
        outputs['png'] = f'data:image/png;base64,{base64.b64encode(png_buffer.getvalue()).decode()}'
    plt.close(fig_for_output)
    return outputs


def display_pdf(base64_pdf):
    """在Streamlit中嵌入显示PDF"""
    return f'<iframe src="{base64_pdf}" width="100%" height="800" type="application/pdf"></iframe>'


# --- 3. TMS 及天气查询模块 ---
@st.cache_data(show_spinner=False, ttl=600)
def fetch_and_process_server_movies(priority_movie_titles=None):
    if priority_movie_titles is None: priority_movie_titles = []

    # 获取环境变量
    app_secret = os.getenv("TMS_APP_SECRET")
    ticket = os.getenv("TMS_TICKET")
    theater_id_str = os.getenv("TMS_THEATER_ID")
    x_session_id = os.getenv("TMS_X_SESSION_ID")

    # 转换 ID 为整数
    try:
        theater_id = int(theater_id_str) if theater_id_str else 0
    except ValueError:
        st.error("环境变量 TMS_THEATER_ID 格式错误，应为数字。")
        return {}, []

    token_headers = {
        'Host': 'oa.hengdianfilm.com:7080', 'Content-Type': 'application/json',
        'Origin': 'http://115.239.253.233:7080', 'Connection': 'keep-alive',
        'Accept': 'application/json, text/javascript, */*; q=0.01',
        'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_5_0 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) CriOS/138.0.7204.156 Mobile/15E148 Safari/604.1',
        'Accept-Language': 'zh-CN,zh-Hans;q=0.9',
    }

    # 使用变量
    token_json_data = {'appId': 'hd', 'appSecret': app_secret, 'timeStamp': int(time.time() * 1000)}
    # 动态构建 URL
    token_url = f'http://oa.hengdianfilm.com:7080/cinema-api/admin/generateToken?token=hd&murl=?token=hd&murl=ticket={ticket}'

    response = requests.post(token_url, headers=token_headers, json=token_json_data, timeout=10)

    response.raise_for_status()
    token_data = response.json()
    if token_data.get('error_code') != '0000': raise Exception(f"获取Token失败: {token_data.get('error_desc')}")
    auth_token = token_data['param']
    all_movies, page_index = [], 1
    while True:
        list_headers = {
            'Accept': 'application/json, text/javascript, */*; q=0.01',
            'Content-Type': 'application/json; charset=UTF-8',
            'Origin': 'http://115.239.253.233:7080', 'Proxy-Connection': 'keep-alive',
            'Token': auth_token,
            'User-Agent': 'Mozilla/5.0 ...',
            'X-SESSIONID': x_session_id,  # 使用变量
        }
        list_params = {'token': 'hd', 'murl': 'ContentMovie'}
        # 使用变量
        list_json_data = {'THEATER_ID': theater_id, 'SOURCE': 'SERVER', 'ASSERT_TYPE': 2, 'PAGE_CAPACITY': 20,
                          'PAGE_INDEX': page_index}

        list_url = 'http://oa.hengdianfilm.com:7080/cinema-api/cinema/server/dcp/list'
        response = requests.post(list_url, params=list_params, headers=list_headers, json=list_json_data, verify=False)
        response.raise_for_status()
        movie_data = response.json()
        if movie_data.get("RSPCD") != "000000": raise Exception(f"获取影片列表失败: {movie_data.get('RSPMSG')}")
        body = movie_data.get("BODY", {});
        movies_on_page = body.get("LIST", [])
        if not movies_on_page: break
        all_movies.extend(movies_on_page)
        if len(all_movies) >= body.get("COUNT", 0): break
        page_index += 1;
        time.sleep(1)
    movie_details = {m.get('CONTENT_NAME'): {'assert_name': m.get('ASSERT_NAME'),
                                             'halls': sorted([h.get('HALL_NAME') for h in m.get('HALL_INFO', [])]),
                                             'play_time': m.get('PLAY_TIME')} for m in all_movies if
                     m.get('CONTENT_NAME')}
    by_hall = defaultdict(list)
    for content_name, details in movie_details.items():
        for hall_name in details['halls']:
            by_hall[hall_name].append({'content_name': content_name, 'details': details})
    for hall_name in by_hall:
        by_hall[hall_name].sort(
            key=lambda item: (item['details']['assert_name'] is None or item['details']['assert_name'] == '',
                              item['details']['assert_name'] or item['content_name']))
    view2_list = [{'assert_name': d['assert_name'], 'content_name': c, 'halls': d['halls'], 'play_time': d['play_time']}
                  for c, d in movie_details.items() if d.get('assert_name')]
    priority_list = [item for item in view2_list if
                     any(p_title in item['assert_name'] for p_title in priority_movie_titles)]
    other_list_items = [item for item in view2_list if item not in priority_list]
    priority_list.sort(key=lambda x: x['assert_name']);
    other_list_items.sort(key=lambda x: x['assert_name'])
    final_sorted_list = priority_list + other_list_items
    return dict(sorted(by_hall.items())), final_sorted_list


@st.cache_data(show_spinner=False, ttl=600)
def get_weather_forecast(target_date):
    """获取指定日期的天气信息并格式化为标题字符串"""
    if not target_date:
        return "当日放映影片"
    url = "https://restapi.amap.com/v3/weather/weatherInfo"
    params = {'key': GAODE_API_KEY, 'city': ADCODE, 'extensions': 'all', 'output': 'JSON'}
    try:
        response = requests.get(url, params=params, timeout=5)
        response.raise_for_status()
        data = response.json()
        if data.get('status') == '1' and data.get('forecasts'):
            target_date_str = target_date.strftime('%Y-%m-%d')
            for day_cast in data['forecasts'][0].get('casts', []):
                if day_cast.get('date') == target_date_str:
                    weekday_map = {'1': '一', '2': '二', '3': '三', '4': '四', '5': '五', '6': '六', '7': '日'}
                    week = weekday_map.get(day_cast.get('week'), '')
                    weather = day_cast.get('dayweather', '未知')
                    max_temp = day_cast.get('daytemp', 'N/A')
                    min_temp = day_cast.get('nighttemp', 'N/A')
                    return f"今日放映影片（{target_date_str}，星期{week}，{weather}，{max_temp}℃ / {min_temp}℃）"
    except Exception as e:
        print(f"天气 API 请求失败: {e}")
    return f"今日放映影片（{target_date.strftime('%Y-%m-%d')}，天气未知）"


def get_circled_number(hall_name):
    mapping = {'1': '①', '2': '②', '3': '③', '4': '④', '5': '⑤', '6': '⑥', '7': '⑦', '8': '⑧', '9': '⑨'}
    num_str = ''.join(filter(str.isdigit, hall_name));
    return mapping.get(num_str, '')


def format_play_time(time_str):
    if not time_str or not isinstance(time_str, str): return None
    try:
        parts = time_str.split(':');
        hours = int(parts[0]);
        minutes = int(parts[1])
        return hours * 60 + minutes
    except (ValueError, IndexError):
        return None


def add_tms_locations_to_analysis(analysis_df, tms_movie_list):
    locations = []
    for _, row in analysis_df.iterrows():
        movie_title = row['影片']
        found_versions = []
        for tms_movie in tms_movie_list:
            if tms_movie['assert_name'].startswith(movie_title):
                version_name = tms_movie['assert_name'].replace(movie_title, '').strip()
                circled_halls = " ".join(sorted([get_circled_number(h) for h in tms_movie['halls']]))
                found_versions.append(f"{version_name}：{circled_halls}" if version_name else circled_halls)
        locations.append('｜'.join(found_versions))
    analysis_df['影片所在影厅位置'] = locations
    return analysis_df


# --- 4. 新增的经营数据API函数 ---
def fetch_income_data(token, date_str):
    """获取指定日期的票房、卖品收入和卖品占比"""
    url = 'https://app.bi.piao51.cn/cinema-app/mycinema/incomeProportion.action'
    params = {'cinemaId': CINEMA_ID, 'token': token, 'qTimeStart': date_str, 'qTimeEnd': date_str}
    headers = {'Host': 'app.bi.piao51.cn', 'X-Requested-With': 'XMLHttpRequest', 'User-Agent': 'Mozilla/5.0'}
    try:
        response = requests.get(url, params=params, headers=headers, timeout=10)
        response.raise_for_status()
        data = response.json()
        if data.get('code') == 'A00000' and data.get('results'):
            day_data = data['results'][0]
            return {
                "ticket_income": float(day_data.get('ticketIncome') or 0),
                "goods_income": float(day_data.get('goodsIncome') or 0),
                "sold_incomes_zb": float(day_data.get('soldIncomesZb') or 0)
            }
    except Exception as e:
        st.warning(f"获取经营收入数据失败: {e}")
    return None


def fetch_membership_data(token, date_str):
    """获取指定日期的文旅卡开卡数"""
    url = 'https://app.bi.piao51.cn/cinema-app/mycinema/membership.action'
    params = {'token': token, 'cinemaId': CINEMA_ID, 'startDate': date_str, 'endDate': date_str}
    headers = {'Host': 'app.bi.piao51.cn', 'X-Requested-With': 'XMLHttpRequest', 'User-Agent': 'Mozilla/5.0'}
    try:
        response = requests.get(url, params=params, headers=headers, timeout=10)
        response.raise_for_status()
        data = response.json()
        if data.get('code') == 'A00000' and data.get('results'):
            for card_type in data['results']:
                if card_type.get('cardLevelName') == '文旅消费卡':
                    return int(card_type.get('sendCardNums', 0))
            return 0  # 没找到文旅卡
    except Exception as e:
        st.warning(f"获取会员开卡数据失败: {e}")
    return 0


# --- 6. 新增：场次合理性检查日志函数 ---
def generate_schedule_check_logs(schedule_list, date_str):
    """
    生成场次合理性检查日志
    :param schedule_list: 排片数据列表 (list of dicts)
    :param date_str: 排片日期字符串 (YYYY-MM-DD)
    :return: 格式化后的日志文本 (str)
    """
    if not schedule_list:
        return "无排片数据，无法进行合理性检查。"

    # 转换为 DataFrame 方便处理
    df_original = pd.DataFrame(schedule_list)
    # 重命名列以符合逻辑处理习惯
    df_original.rename(
        columns={'hallName': 'Hall', 'movieName': 'filmName', 'showStartTime': 'StartTime', 'showEndTime': 'EndTime'},
        inplace=True)

    # 预处理：转换时间列，简化影厅名
    df_original['startTime'] = pd.to_datetime(df_original['StartTime'], format='%H:%M', errors='coerce').apply(
        lambda t: datetime.combine(datetime.strptime(date_str, '%Y-%m-%d').date(), t.time()) if pd.notnull(
            t) else pd.NaT)
    df_original['endTime'] = pd.to_datetime(df_original['EndTime'], format='%H:%M', errors='coerce').apply(
        lambda t: datetime.combine(datetime.strptime(date_str, '%Y-%m-%d').date(), t.time()) if pd.notnull(
            t) else pd.NaT)
    # 处理跨天时间
    df_original.loc[df_original['endTime'] < df_original['startTime'], 'endTime'] += timedelta(days=1)

    # 简单影厅名处理 (仅提取数字或主要标识)
    def simplify_hall(name):
        import re
        match = re.search(r'(\d+号?)', str(name))
        return match.group(1) if match else str(name)[:2]

    df_original['simpleHallName'] = df_original['Hall'].apply(simplify_hall)

    df_check = df_original.sort_values(by='startTime').reset_index(drop=True)
    final_log_parts = []

    # --- Rule 1: 同影片场次间隔过近 ---
    logs_r1 = []
    for film_name in df_check['filmName'].unique():
        film_schedules = df_check[df_check['filmName'] == film_name].sort_values(by='startTime').reset_index()
        if len(film_schedules) > 1:
            for i in range(len(film_schedules) - 1):
                s1, s2 = film_schedules.iloc[i], film_schedules.iloc[i + 1]
                interval = (s2['startTime'] - s1['startTime']).total_seconds() / 60
                if interval < 30:
                    log_entry = f"《{s1['filmName']}》{s1['simpleHallName']}【{s1['startTime'].strftime('%H:%M')}】和 {s2['simpleHallName']}【{s2['startTime'].strftime('%H:%M')}】开场时间距离 {int(interval)} 分钟"
                    logs_r1.append(log_entry)

    final_log_parts.append("规则一：同影片场次间隔过近（少于 30 分钟）")
    if logs_r1:
        for i, log in enumerate(logs_r1, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 2: 30 分钟内影片开场超过 4 场 ---
    logs_r2 = []
    i = 0
    processed_indices_r2 = set()
    while i < len(df_check):
        if i in processed_indices_r2:
            i += 1
            continue
        window_start_time = df_check.iloc[i]['startTime']
        window_end_time_30min = window_start_time + timedelta(minutes=30)
        window_df = df_check[
            (df_check['startTime'] >= window_start_time) & (df_check['startTime'] < window_end_time_30min)]

        if len(window_df) > 4:
            start_t_str = window_df.iloc[0]['startTime'].strftime('%H:%M')
            end_t_str = window_df.iloc[-1]['startTime'].strftime('%H:%M')
            log_message_lines = [f"【{start_t_str} - {end_t_str}】开场时间比较集中："]
            for _, row in window_df.iterrows():
                log_message_lines.append(
                    f"  {row['simpleHallName']}《{row['filmName']}》> {row['startTime'].strftime('%H:%M')}")
                processed_indices_r2.add(row.name)
            logs_r2.append("\n".join(log_message_lines))
        i += 1

    final_log_parts.append("\n规则二：30 分钟内影片开场超过 4 场")
    if logs_r2:
        for i, log in enumerate(logs_r2, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 3: 场次开场间隔超过 30 分钟 ---
    logs_r3 = []
    if len(df_check) > 1:
        for i in range(len(df_check) - 1):
            s1_start, s2_start = df_check.iloc[i]['startTime'], df_check.iloc[i + 1]['startTime']
            gap = (s2_start - s1_start).total_seconds() / 60
            if gap > 30:
                log_entry = f"【{s1_start.strftime('%H:%M')} ~ {s2_start.strftime('%H:%M')}】缺少影片开场，间隔 {int(gap)} 分钟"
                logs_r3.append(log_entry)

    final_log_parts.append("\n规则三：场次开场间隔超过 30 分钟")
    if logs_r3:
        for i, log in enumerate(logs_r3, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 4: 最早/最晚场次时间检查 ---
    logs_r4 = []
    if not df_check.empty:
        first_sched = df_check.iloc[0]
        last_sched = df_check.iloc[-1]
        if first_sched['startTime'].time() > dt_time(10, 0):
            logs_r4.append(
                f"最早一场 {first_sched['simpleHallName']}《{first_sched['filmName']}》{first_sched['startTime'].strftime('%H:%M')} 晚于 10:00")
        if last_sched['startTime'].time() < dt_time(22, 30):
            logs_r4.append(
                f"最晚一场 {last_sched['simpleHallName']}《{last_sched['filmName']}》{last_sched['startTime'].strftime('%H:%M')} 早于 22:30")

    final_log_parts.append("\n规则四：最早一场晚于 10:00，最晚一场早于 22:30")
    if logs_r4:
        for i, log in enumerate(logs_r4, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 5: 影厅空闲时间超过 1 小时 (10:00-23:00) ---
    logs_r5 = []
    today_date = datetime.strptime(date_str, '%Y-%m-%d').date()
    window_start_time_limit_r5 = datetime.combine(today_date, dt_time(10, 0))
    window_end_time_limit_r5 = datetime.combine(today_date, dt_time(23, 0))

    unique_halls_r5 = df_original['simpleHallName'].unique()
    for hall_name in unique_halls_r5:
        hall_df = df_original[df_original['simpleHallName'] == hall_name].sort_values(by='startTime')
        if len(hall_df) > 1:
            for i in range(len(hall_df) - 1):
                prev_sched_end = hall_df.iloc[i]['endTime']
                curr_sched_start = hall_df.iloc[i + 1]['startTime']
                if prev_sched_end < window_end_time_limit_r5 and curr_sched_start > window_start_time_limit_r5:
                    idle_duration_minutes = (curr_sched_start - prev_sched_end).total_seconds() / 60
                    if idle_duration_minutes > 60:
                        log_entry = f"{hall_name} 【{prev_sched_end.strftime('%H:%M')} ~ {curr_sched_start.strftime('%H:%M')}】无影片在播，时长 {int(idle_duration_minutes)} 分钟"
                        logs_r5.append(log_entry)

    final_log_parts.append("\n规则五：影厅空闲时间超过 1 小时（10:00-23:00）")
    if logs_r5:
        for i, log in enumerate(logs_r5, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 6: 影厅场次转换时间检查 ---
    logs_r6 = []
    for hall_name in df_original['simpleHallName'].unique():
        hall_df = df_original[df_original['simpleHallName'] == hall_name].sort_values(by='startTime')
        if len(hall_df) > 1:
            for i in range(len(hall_df) - 1):
                prev_sched = hall_df.iloc[i]
                next_sched = hall_df.iloc[i + 1]
                conversion_time = (next_sched['startTime'] - prev_sched['endTime']).total_seconds() / 60
                if conversion_time < 10:
                    logs_r6.append(
                        f"{hall_name} {prev_sched['endTime'].strftime('%H:%M')} 《{prev_sched['filmName']}》结束后影厅空闲时间仅为 {int(conversion_time)} 分钟")

    final_log_parts.append("\n规则六：影厅场次转换时间检查")
    if logs_r6:
        for i, log in enumerate(logs_r6, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 7: 动态散场和入场高峰预警 ---
    logs_r7 = []
    final_log_parts.append("\n规则七：动态散场和入场高峰预警")

    if not df_check.empty:
        start_time = df_check.iloc[0]['startTime'].replace(second=0, microsecond=0)
        end_time = df_check.iloc[-1]['endTime']
        current_time = start_time
        reported_windows = set()

        while current_time < end_time:
            window_end = current_time + timedelta(minutes=10)
            starts_in_window = df_check[(df_check['startTime'] >= current_time) & (df_check['startTime'] < window_end)]
            ends_in_window = df_check[(df_check['endTime'] > current_time) & (df_check['endTime'] <= window_end)]

            if len(starts_in_window) + len(ends_in_window) > 5:
                window_tuple = (current_time.strftime('%H:%M'), window_end.strftime('%H:%M'))
                if window_tuple not in reported_windows:
                    exit_halls = "、".join(ends_in_window['simpleHallName'])
                    entry_halls = "、".join(starts_in_window['simpleHallName'])
                    log_msg = f"【{current_time.strftime('%H:%M')} ~ {window_end.strftime('%H:%M')}】"
                    if not ends_in_window.empty:
                        log_msg += f"，{exit_halls}集中散场"
                    if not starts_in_window.empty:
                        if not ends_in_window.empty:
                            log_msg += "，同时"
                        else:
                            log_msg += "，"
                        log_msg += f"{entry_halls}即将入场"
                    log_msg += "，预计人流瞬时压力过大。"
                    logs_r7.append(log_msg)
                    reported_windows.add(window_tuple)
            current_time += timedelta(minutes=5)

        # Part 2: Simultaneous start
        start_groups = df_check.groupby('startTime').filter(lambda x: len(x) > 3)
        for time_val, group in start_groups.groupby('startTime'):
            halls = "、".join(group['simpleHallName'])
            logs_r7.append(f"{time_val.strftime('%H:%M')}，{halls}电影同时开场，注意预计人流瞬时压力过大。")

        # Part 3: Simultaneous end
        end_groups = df_check.groupby('endTime').filter(lambda x: len(x) > 3)
        for time_val, group in end_groups.groupby('endTime'):
            halls = "、".join(group['simpleHallName'])
            logs_r7.append(f"{time_val.strftime('%H:%M')}，{halls}电影同时散场，注意预计人流瞬时压力过大。")

    if logs_r7:
        for i, log in enumerate(logs_r7, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 8: “幽灵厅”预警 ---
    logs_r8 = []
    final_log_parts.append("\n规则八：影厅结束运营过早预警")
    for hall_name in df_original['simpleHallName'].unique():
        last_sched = df_original[df_original['simpleHallName'] == hall_name].nlargest(1, 'endTime').iloc[0]
        # 简单判断：如果最后一场结束时间早于 22:30 且就是当天（不是跨天到凌晨）
        if last_sched['endTime'].date() == today_date and last_sched['endTime'].time() < dt_time(22, 30):
            logs_r8.append(f"{hall_name} 最后一场于【{last_sched['endTime'].strftime('%H:%M')}】结束，过早停运。")
    if logs_r8:
        for i, log in enumerate(logs_r8, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    # --- Rule 9: 真正意义的黄金时段热门影片排片密度检查 ---
    logs_r9 = []
    final_log_parts.append("\n规则九：黄金时段热门影片排片密度检查")
    if not df_check.empty:
        weekday = today_date.weekday()
        golden_hours_r9 = [
            [(dt_time(14, 0), dt_time(16, 0)), (dt_time(19, 0), dt_time(22, 0))],
            [(dt_time(14, 0), dt_time(15, 30)), (dt_time(19, 0), dt_time(22, 20))],
            [(dt_time(14, 30), dt_time(16, 0)), (dt_time(19, 0), dt_time(21, 40))],
            [(dt_time(14, 0), dt_time(16, 0)), (dt_time(19, 0), dt_time(22, 0))],
            [(dt_time(14, 0), dt_time(15, 0)), (dt_time(19, 0), dt_time(22, 0))],
            [(dt_time(14, 0), dt_time(16, 0)), (dt_time(19, 0), dt_time(22, 0))],
            [(dt_time(14, 0), dt_time(17, 0)), (dt_time(19, 0), dt_time(21, 30))]
        ][weekday]

        film_counts = df_check['filmName'].value_counts()
        if not film_counts.empty:
            max_count = film_counts.iloc[0]
            # 定义热门影片：排片量接近第一名（95%以上）的影片
            hot_films = film_counts[film_counts >= max_count * 0.95].index.tolist()

            golden_hour_schedules = df_check[
                df_check['startTime'].apply(lambda dt: any(start <= dt.time() < end for start, end in golden_hours_r9))]

            for film in hot_films:
                hot_film_total_in_golden = len(golden_hour_schedules[golden_hour_schedules['filmName'] == film])
                golden_total = len(golden_hour_schedules)
                if golden_total > 0:
                    ratio = hot_film_total_in_golden / golden_total
                    if ratio < 0.3:  # 热门影片黄金时段占比低于30%预警
                        period_str = " 和 ".join(
                            [f"{s.strftime('%H:%M')}-{e.strftime('%H:%M')}" for s, e in golden_hours_r9])
                        logs_r9.append(f"《{film}》在核心黄金时段 {period_str} 排片占比仅为{ratio:.1%}，低于 30%。")

    if logs_r9:
        for i, log in enumerate(logs_r9, 1):
            final_log_parts.append(f"{i}. {log}")
    else:
        final_log_parts.append("（无）")

    return "\n".join(final_log_parts)


def check_tms_file_availability(schedule_list, tms_data, date_str):
    """
    对比排片表和TMS数据，检查影厅是否缺失对应的影片文件
    优化：仅匹配核心片名（去除版本后缀），优化影厅名显示，合并日志输出
    """
    if not schedule_list:
        return "未获取到排片数据，无法检查。"

    if not tms_data:
        return "未获取到 TMS 数据，无法检查。"

    # --- 内部辅助函数 ---
    def get_core_movie_name(raw_name):
        """
        获取核心片名用于匹配：
        1. 先执行标准的 clean_movie_title (统一命名)
        2. 再去除所有括号及括号内的内容 (去除版本/制式信息)
        例如：'疯狂动物城2（数字3D）' -> '疯狂动物城2'
        """
        # 1. 基础清洗 (利用现有的逻辑处理中英文/特殊后缀)
        # 注意：这里我们不传入 canonical_names，只做规则清洗
        name = clean_movie_title(raw_name)
        # 2. 正则去除中文全角括号及内容 （...）
        name = re.sub(r'（.*?）', '', name)
        # 3. 正则去除英文半角括号及内容 (...)
        name = re.sub(r'\(.*?\)', '', name)
        return name.strip()

    def clean_hall_display_name(raw_name):
        """去除影厅名两端多余的 【】 [] 符号"""
        return str(raw_name).strip('【】[] ')

    def get_hall_key_num(name):
        """提取影厅数字ID用于数据匹配 (如 '1号厅' -> '1')"""
        nums = re.findall(r'\d+', str(name))
        return nums[0] if nums else str(name)

    # ------------------

    # 1. 预处理 TMS 数据
    # 结构: {'1': ['Zootopia2', '疯狂动物城2', ...], ...}
    tms_map = defaultdict(set)  # 使用 set 提高查找效率

    for hall_name, movies in tms_data.items():
        hall_key = get_hall_key_num(hall_name)
        for movie in movies:
            # 收集 Assert Name (显示名)
            if movie.get('details', {}).get('assert_name'):
                # 同样对 TMS 里的名字取核心名，提高匹配率
                core_tms_name = get_core_movie_name(str(movie['details']['assert_name']))
                tms_map[hall_key].add(core_tms_name.upper())
                # 保留原始 Assert Name 用于兜底匹配
                tms_map[hall_key].add(str(movie['details']['assert_name']).upper())

            # 收集 Content Name (文件名/UUID)
            if movie.get('content_name'):
                tms_map[hall_key].add(str(movie['content_name']).upper())

    # 2. 遍历排片数据进行检查
    missing_logs = []
    checked_combinations = set()

    for item in schedule_list:
        hall_raw = item.get('hallName') or item.get('Hall')
        movie_raw = item.get('movieName') or item.get('Movie')

        if not hall_raw or not movie_raw:
            continue

        # 准备数据
        hall_num = get_hall_key_num(hall_raw)
        hall_display = clean_hall_display_name(hall_raw)  # 清洗后的影厅名

        # 获取排片的核心片名 (去掉版本后缀)
        target_movie_core = get_core_movie_name(movie_raw).upper()

        # 组合键去重 (同一厅同一部片只报一次)
        combo_key = (hall_num, target_movie_core)
        if combo_key in checked_combinations:
            continue
        checked_combinations.add(combo_key)

        # 检查逻辑
        if hall_num not in tms_map:
            # 找不到影厅数据暂不报错，可能是未映射或设备离线，避免刷屏
            continue

            # 核心匹配：检查 TMS 集合中是否包含核心片名
        # 方式A：精确匹配核心名 (推荐，最准)
        # 方式B：模糊包含 (target in tms_file)

        has_file = False
        tms_files = tms_map[hall_num]

        # 策略：只要 TMS 中有一个文件名 包含 我们的核心排片名，就视为有片
        # 例如：排片 core='疯狂动物城2'，TMS='疯狂动物城2_IMAX' -> 匹配成功
        for tms_file in tms_files:
            if target_movie_core in tms_file:
                has_file = True
                break

        if not has_file:
            # 记录日志，使用清洗后的影厅名和排片原名
            missing_logs.append(f"【{hall_display}】排映《{movie_raw}》，但服务器未检测到包含“{target_movie_core}”的文件。")

    # 3. 格式化输出
    if not missing_logs:
        return None  # 返回 None 表示一切正常

    # 生成带编号的字符串
    formatted_output = []
    for idx, log in enumerate(missing_logs, 1):
        formatted_output.append(f"{idx}. ❌ 缺片警告：{log}")

    return "\n".join(formatted_output)

# --- 5. UI 渲染与交互逻辑 ---

def display_analysis_results(df_raw, data_source_name, date_for_display, query_tms_enabled):
    if df_raw.empty:
        st.info(f"请先从 {data_source_name} 加载数据。");
        return

    if data_source_name == "文件":
        token_data = load_token()
        if not token_data:
            token_data = login_and_get_token()

        token = token_data.get('token') if token_data else None
        date_str = date_for_display.strftime('%Y-%m-%d') if date_for_display else None

        canonical_names = []
        if token and date_str:
            canonical_names = fetch_canonical_movie_names(token, date_str)

        df_raw['影片名称_清理后'] = df_raw['影片名称'].apply(lambda x: clean_movie_title(x, canonical_names))
    else:
        df_raw['影片名称_清理后'] = df_raw['影片名称']

    date_str = f"{date_for_display} " if date_for_display else ""
    total_revenue, total_attendance, total_sessions = df_raw['总收入'].sum(), df_raw['总人次'].sum(), len(df_raw)
    st.markdown(
        f"> {date_str}数据总览：总票房 **¥{total_revenue:,.2f}** | 总人次 **{total_attendance:,.0f}** | 总场次 **{total_sessions:,.0f}**")

    format_config = {'座位数': '{:,.0f}', '场次': '{:,.0f}', '人次': '{:,.0f}', '票房': '{:,.2f}', '均价': '{:.2f}',
                     '座次比': '{:.2%}', '场次比': '{:.2%}', '票房比': '{:.2%}', '座次效率': '{:.2f}',
                     '场次效率': '{:.2f}'}
    full_day_analysis, prime_time_analysis = process_and_analyze_data(df_raw.copy()), process_and_analyze_data(
        df_raw[df_raw['放映时间'].between(dt_time(14, 0), dt_time(21, 0))].copy())

    if query_tms_enabled:
        with st.spinner("正在关联查询 TMS 服务器..."):
            try:
                priority_titles = full_day_analysis['影片'].unique().tolist()
                _, tms_movie_list = fetch_and_process_server_movies(priority_titles)
                full_day_analysis = add_tms_locations_to_analysis(full_day_analysis, tms_movie_list)
                prime_time_analysis = add_tms_locations_to_analysis(prime_time_analysis, tms_movie_list)
                if '影片所在影厅位置' in full_day_analysis.columns:
                    cols = full_day_analysis.columns.tolist();
                    full_day_analysis = full_day_analysis[cols[:1] + ['影片所在影厅位置'] + cols[1:-1]]
                if '影片所在影厅位置' in prime_time_analysis.columns:
                    cols = prime_time_analysis.columns.tolist();
                    prime_time_analysis = prime_time_analysis[cols[:1] + ['影片所在影厅位置'] + cols[1:-1]]
                st.toast("TMS 影片位置关联成功！", icon="🔗")
            except Exception as e:
                st.error(f"关联TMS失败: {e}")

    st.markdown("#### 全天排片效率分析");
    st.dataframe(full_day_analysis.style.format(format_config).apply(style_efficiency, axis=1),
                 use_container_width=True, hide_index=True)
    st.markdown("#### 黄金时段排片效率分析 (14:00-21:00)");
    st.dataframe(prime_time_analysis.style.format(format_config).apply(style_efficiency, axis=1),
                 use_container_width=True, hide_index=True)
    if not full_day_analysis.empty:
        st.markdown("### 排片效率汇总")
        full_day_summary = full_day_analysis.rename(
            columns={'场次': '全部场次', '座次效率': '全部座次效率', '场次效率': '全部场次效率'})
        full_day_cols_to_keep = ['影片', '票房', '全部场次', '全部座次效率', '全部场次效率']
        if '影片所在影厅位置' in full_day_summary.columns: full_day_cols_to_keep.insert(1, '影片所在影厅位置')
        full_day_summary = full_day_summary[full_day_cols_to_keep]
        prime_time_summary = prime_time_analysis.rename(
            columns={'场次': '黄金时段场次', '座次效率': '黄金时段座次效率', '场次效率': '黄金时段场次效率'})[
            ['影片', '黄金时段场次', '黄金时段座次效率', '黄金时段场次效率']]
        summary_df = pd.merge(full_day_summary, prime_time_summary, on='影片', how='left').fillna(0)
        summary_df['黄金时段场次'] = summary_df['黄金时段场次'].astype(int)
        summary_format_config = {'票房': '{:,.2f}', '全部场次': '{:,.0f}', '黄金时段场次': '{:,.0f}',
                                 '全部座次效率': '{:.2f}', '全部场次效率': '{:.2f}', '黄金时段座次效率': '{:.2f}',
                                 '黄金时段场次效率': '{:.2f}'}
        st.dataframe(summary_df.style.format(summary_format_config).apply(style_summary_efficiency, axis=1),
                     use_container_width=True, hide_index=True)


def fetch_and_process_daily_sessions(date_str, quiet=False):
    """获取并处理指定日期的排片场次，返回(场次字典, 总场次数)"""
    if not quiet: st.write(f"正在查询 {date_str} 的排片数据...")

    token_data = load_token()
    token = token_data.get('token') if token_data else None
    if not token:
        token_data = login_and_get_token()
        token = token_data.get('token') if token_data else None

    schedule, _ = get_api_data_with_token_management(date_str)

    if not schedule:
        if not quiet: st.warning(f"未能获取到 {date_str} 的排片数据。")
        return None, None, None  # 修改返回值，增加 raw_schedule

    total_sessions = len(schedule)
    df = pd.DataFrame(schedule)

    canonical_names = []
    if token:
        canonical_names = fetch_canonical_movie_names(token, date_str)

    df['影片名称_清理后'] = df['movieName'].apply(lambda x: clean_movie_title(x, canonical_names))

    sessions_map = df.groupby('影片名称_清理后').size().to_dict()
    # 返回 (映射表, 总场次, 原始排片列表)
    return sessions_map, total_sessions, schedule


def generate_efficiency_report_df(analysis_df, next_day_sessions_map=None, next_day_total_sessions=None):
    if analysis_df.empty: return pd.DataFrame()
    report_df = analysis_df[
        ['影片', '座位数', '场次', '票房', '人次', '均价', '座次比', '场次比', '票房比', '座次效率', '场次效率']].copy()
    report_df['情况说明'] = '';
    report_df['次日调整方案'] = ''
    if next_day_sessions_map is not None:
        report_df['次日场数'] = report_df['影片'].map(next_day_sessions_map).fillna(0).astype(int)
    else:
        report_df['次日场数'] = report_df['场次']
    totals = report_df[['座位数', '场次', '票房', '人次']].sum()
    totals['影片'] = ''
    totals['次日场数'] = next_day_total_sessions if next_day_total_sessions is not None else report_df['次日场数'].sum()
    report_df = pd.concat([report_df, pd.DataFrame(totals).T], ignore_index=True)
    return report_df


def generate_excel_paste_data(df):
    if df.empty: return ""
    lines, total_row_num = [], len(df) + 1
    for i, row in df.iterrows():
        excel_row_num = i + 2
        line = [row['影片'], row['座位数'], row['场次'], row['票房'], row['人次']]
        if i == len(df) - 1:  # 合计行
            line.extend(['', '', '', '', '', '', '', '', row['次日场数']])
        else:  # 数据行
            line.extend([f"=IFERROR(F{excel_row_num}/G{excel_row_num},0)", f"=D{excel_row_num}/D${total_row_num}",
                         f"=E{excel_row_num}/E${total_row_num}", f"=F{excel_row_num}/F${total_row_num}",
                         f"=IFERROR(K{excel_row_num}/I{excel_row_num},0)",
                         f"=IFERROR(K{excel_row_num}/J{excel_row_num},0)", row['情况说明'], row['次日调整方案'],
                         row['次日场数']])
        lines.append("\t".join(map(str, line)))
    return "\n".join(lines)


def get_business_date(df_with_datetime):
    """根据包含完整日期时间列的DataFrame计算营业日"""
    crossover_time = dt_time(6, 0)
    df_with_datetime['business_date'] = df_with_datetime['datetime'].apply(
        lambda dt: (dt - timedelta(days=1)).date() if dt.time() < crossover_time else dt.date()
    )
    return df_with_datetime['business_date'].mode()[0]


# --- 主应用 ---
def main():
    st.title('影城工作便捷工具')

    # 初始化 session_state
    if 'file_df' not in st.session_state: st.session_state.file_df, st.session_state.api_df = pd.DataFrame(), pd.DataFrame()
    if 'api_date' not in st.session_state: st.session_state.api_date = datetime.now().date()
    if 'file_date' not in st.session_state: st.session_state.file_date = None
    if 'today_movie_count' not in st.session_state: st.session_state.today_movie_count = 0
    if 'previous_day_movie_count' not in st.session_state: st.session_state.previous_day_movie_count = 0
    if 'daily_report_df' not in st.session_state: st.session_state.daily_report_df = pd.DataFrame()
    if 'processed_print_data' not in st.session_state: st.session_state.processed_print_data = None
    if 'check_logs' not in st.session_state: st.session_state.check_logs = ""

    tab1, tab2, tab_sales, tab_report, tab_print, tab3 = st.tabs(
        ["🔍 排片效率分析", "📋 次日排片效率分析报表", "🍿 卖品品类分析表", "📑 影片映出日累计表", "🖨️ 场次与散场打印",
         "🎬 TMS 影片查询"])

    with tab1:
        # 顶部控制区
        col_a, col_b = st.columns(2)
        with col_a:
            import_from_file = st.checkbox("从`影片映出日累计报表.xlsx`导入数据")
        with col_b:
            query_tms_for_location = st.checkbox("查询 TMS 找影片所在影厅")

        if import_from_file:
            st.header("从本地文件导入数据")
            st.write("上传 `影片映出日累计报表.xlsx`，程序将自动处理数据。")
            uploaded_file = st.file_uploader("上传 Excel 文件", type=['xlsx', 'xls'], label_visibility="collapsed")
            if uploaded_file is not None:
                with st.spinner("正在处理文件..."):
                    try:
                        df = pd.read_excel(uploaded_file, skiprows=3, header=None)
                        df.rename(columns={0: '影片名称', 1: '放映日期', 2: '放映时间', 5: '总人次', 6: '总收入',
                                           7: '座位数'}, inplace=True)
                        df_for_date_calc = df[['放映日期', '放映时间']].copy()
                        df_for_date_calc['datetime_str'] = df_for_date_calc['放映日期'].astype(str).str.split(' ').str[
                                                               0] + ' ' + df_for_date_calc['放映时间'].astype(str)
                        df_for_date_calc['datetime'] = pd.to_datetime(df_for_date_calc['datetime_str'], errors='coerce')
                        df_for_date_calc.dropna(subset=['datetime'], inplace=True)
                        business_date = get_business_date(df_for_date_calc)
                        st.session_state.file_date = business_date
                        st.toast(f"文件营业日识别为: {business_date}", icon="🗓️")
                        df = df[['影片名称', '放映时间', '座位数', '总收入', '总人次']]
                        df.dropna(subset=['影片名称', '放映时间'], inplace=True)
                        for col in ['座位数', '总收入', '总人次']: df[col] = pd.to_numeric(df[col],
                                                                                           errors='coerce').fillna(0)
                        df['放映时间'] = pd.to_datetime(df['放映时间'], format='%H:%M:%S', errors='coerce').dt.time
                        df.dropna(subset=['放映时间'], inplace=True)
                        st.session_state.file_df = df
                    except Exception as e:
                        st.error(f"处理文件或识别日期时出错: {e}");
                        st.session_state.file_df, st.session_state.file_date = pd.DataFrame(), None
            # 展示分析结果 (传递 query_tms_for_location 的状态)
            display_analysis_results(st.session_state.file_df, "文件", st.session_state.file_date,
                                     query_tms_for_location)

        else:
            st.header("使用 API 获取数据")
            st.session_state.api_date = st.date_input("选择要查询的排片日期", value=st.session_state.api_date,
                                                      key="api_date_picker")
            if st.button("获取排片数据", key="fetch_api_data", icon="🫵"):
                with st.spinner(f"正在获取 {st.session_state.api_date} 的排片数据..."):

                    token_data = load_token()
                    if not token_data:
                        token_data = login_and_get_token()
                    token = token_data.get('token') if token_data else None

                    schedule, halls = get_api_data_with_token_management(st.session_state.api_date.strftime('%Y-%m-%d'))

                    if schedule is not None and halls is not None:
                        # 传入 token 和 date 以进行标准名清洗
                        processed_df = process_api_data(schedule, halls, token,
                                                        st.session_state.api_date.strftime('%Y-%m-%d'))
                        st.session_state.api_df = processed_df
                        if not processed_df.empty: st.toast(f"成功获取并处理了 {len(processed_df)} 条排片数据！",
                                                            icon="✅")
                    else:
                        st.session_state.api_df = pd.DataFrame()
            # 展示分析结果 (传递 query_tms_for_location 的状态)
            display_analysis_results(st.session_state.api_df, "API", st.session_state.api_date, query_tms_for_location)

    with tab2:
        # 确定数据源 (逻辑保持不变，API > 文件)
        source_df_raw, source_date, source_date_str = pd.DataFrame(), None, ""
        if not st.session_state.api_df.empty:
            source_df_raw, source_date, source_date_str = st.session_state.api_df, st.session_state.api_date, f"{st.session_state.api_date}"
            st.toast("正在使用来自 **API 获取** 的最新数据。", icon="☁️")
        elif not st.session_state.file_df.empty:
            source_df_raw, source_date, source_date_str = st.session_state.file_df, st.session_state.file_date, f"{st.session_state.file_date} (文件)"
            st.toast("API数据为空，正在使用来自 **文件导入** 的数据。", icon="📃")
        else:
            st.warning('没有可用的数据。请先在 "🔍 排片效率分析" 标签页加载数据。')

        st.header(f"{source_date_str} 排片效率分析与调整建议")
        if not source_df_raw.empty:
            if st.button("生成分析报表", icon="🫵"):
                with st.spinner("正在生成分析报表 (含跨日数据查询)..."):
                    next_day_sessions_map, next_day_total_sessions = None, None
                    previous_day_sessions_map = None
                    token_data = load_token()

                    if source_date and token_data:
                        token = token_data.get('token')
                        # 获取次日数据，并拿到原始排片表用于检查
                        next_day = source_date + timedelta(days=1)
                        next_day_str = next_day.strftime('%Y-%m-%d')
                        next_day_sessions_map, next_day_total_sessions, next_day_raw_schedule = fetch_and_process_daily_sessions(
                            next_day_str, quiet=True)

                        # 生成合理性检查日志
                        if next_day_raw_schedule:
                            logs = generate_schedule_check_logs(next_day_raw_schedule, next_day_str)
                            st.session_state.check_logs = logs
                        else:
                            st.session_state.check_logs = "无法获取次日排片详情，跳过检查。"

                        # 获取前一日数据
                        previous_day = source_date - timedelta(days=1)
                        previous_day_sessions_map, _, _ = fetch_and_process_daily_sessions(
                            previous_day.strftime('%Y-%m-%d'), quiet=True)

                        # 获取经营摘要数据
                        date_str = source_date.strftime('%Y-%m-%d')
                        income_data = fetch_income_data(token, date_str)
                        wenlv_cards = fetch_membership_data(token, date_str)
                        attendance = int(source_df_raw['总人次'].sum())

                        st.session_state.daily_summary_data = {
                            "ticket_income": income_data.get('ticket_income', 0.0) if income_data else 0.0,
                            "attendance": attendance,
                            "goods_income": income_data.get('goods_income', 0.0) if income_data else 0.0,
                            "sold_incomes_zb": income_data.get('sold_incomes_zb', 0.0) if income_data else 0.0,
                            "wenlv_cards": wenlv_cards
                        }
                    else:
                        st.error("无法确定源数据日期或Token，无法获取跨日及经营数据。")

                    if '影片名称_清理后' not in source_df_raw.columns:
                        source_df_raw['影片名称_清理后'] = source_df_raw['影片名称']

                    analysis_df = process_and_analyze_data(source_df_raw.copy())
                    st.session_state.today_movie_count = len(analysis_df)
                    st.session_state.previous_day_movie_count = len(
                        previous_day_sessions_map) if previous_day_sessions_map else 0

                    report_df = generate_efficiency_report_df(analysis_df, next_day_sessions_map,
                                                              next_day_total_sessions)
                    st.session_state.report_df = report_df
                    st.session_state.excel_paste_data = generate_excel_paste_data(report_df)

            if 'report_df' in st.session_state and not st.session_state.report_df.empty:
                st.markdown("#### 排片效率分析表");
                display_df = st.session_state.report_df.copy()
                display_df.insert(0, '序号', range(2, len(display_df) + 2))
                report_format = {'座位数': '{:,.0f}', '场次': '{:,.0f}', '人次': '{:,.0f}', '票房': '{:,.2f}',
                                 '均价': '{:.2f}', '座次比': '{:.2%}', '场次比': '{:.2%}', '票房比': '{:.2%}',
                                 '座次效率': '{:.2f}', '场次效率': '{:.2f}', '次日场数': '{:,.0f}'}
                display_df['均价'] = pd.to_numeric(display_df['均价'], errors='coerce').replace([np.inf, -np.inf],
                                                                                                np.nan)
                st.dataframe(display_df.style.format(report_format, na_rep="#DIV/0!"), use_container_width=True,
                             hide_index=True)

                n_diff = st.session_state.today_movie_count - st.session_state.previous_day_movie_count
                if n_diff > 0:
                    excel_copy_title = f"复制到 Excel (需要增加 {n_diff} 行，从第一个电影名字开始粘贴)"
                elif n_diff < 0:
                    excel_copy_title = f"复制到 Excel (需要减少 {abs(n_diff)} 行，从第一个电影名字开始粘贴)"
                else:
                    excel_copy_title = "复制到 Excel (行数保持不变，从第一个电影名字开始粘贴)"
                st.markdown(f"##### {excel_copy_title}");
                st.code(st.session_state.excel_paste_data, language='text')


                # 复制列表
                report_df = st.session_state.report_df
                movie_titles = report_df.iloc[:-1]['影片'].tolist()
                weather_title = get_weather_forecast(source_date)
                st.markdown(f"##### {weather_title}")
                st.code(''.join([f'《{title}》' for title in movie_titles]), language='text')

                # 经营摘要
                if 'daily_summary_data' in st.session_state:
                    summary_data = st.session_state.daily_summary_data
                    summary_text = (
                        f"花都店，今日票房：{summary_data.get('ticket_income', 0.0):.2f}元，"
                        f"观影人次：{summary_data.get('attendance', 0)}，"
                        f"卖品收入：{summary_data.get('goods_income', 0.0):.2f}元，"
                        f"卖品占比：{summary_data.get('sold_incomes_zb', 0.0):.2f}%，"
                        f"文旅卡：{summary_data.get('wenlv_cards', 0)}张。"
                    )
                    st.markdown(f"##### 今日经营数据概览")
                    st.code(summary_text)
                    st.markdown(f"> 上述文旅卡数量不是实际数据，API 获取的数据是需要次日 10 点更新数据，请查询鼎新报表系统里查询`卡发行`报表。")
                    st.markdown(f"> 抽奖券计算方法：先在鼎新报表系统里查询`卡发行`当日开卡数量然后查询`卡充值`当日详细的充值金额，**充值金额整除 200 加上卡发行数量即为抽奖券数量**。")

                st.markdown("#### 🔍 场次合理性检查日志")
                if st.session_state.check_logs:
                    st.code(st.session_state.check_logs)
                    st.markdown("#### 📡 次日排片 TMS 文件核对")
                    st.info(
                        "此功能将查询 TMS 服务器，检查次日排程的影厅是否有对应的影片文件（不区分语言和制式版本，仅匹配片名）。")

                    if st.button("开始核对 TMS 文件", key="check_tms_files_btn", icon="🕵️‍♂️"):
                        # 确定次日日期
                        check_date = source_date + timedelta(days=1)
                        check_date_str = check_date.strftime('%Y-%m-%d')

                        with st.spinner(f"正在获取 {check_date_str} 的排片数据并连接 TMS 服务器..."):
                            # 1. 获取次日排片
                            schedule_data, _ = get_api_data_with_token_management(check_date_str)

                            if not schedule_data:
                                st.error(f"无法获取 {check_date_str} 的排片数据，请检查网络或 Token。")
                            else:
                                try:
                                    # 2. 获取 TMS 数据
                                    df_sched = pd.DataFrame(schedule_data)
                                    priority_titles = df_sched[
                                        'movieName'].unique().tolist() if 'movieName' in df_sched.columns else []

                                    tms_hall_data, _ = fetch_and_process_server_movies(priority_titles)

                                    # 3. 执行比对 (使用新函数)
                                    logs_text = check_tms_file_availability(schedule_data, tms_hall_data,
                                                                            check_date_str)

                                    if logs_text is None:
                                        st.success(
                                            f"✅ 核对完成：{check_date_str} 所有排映影片在对应影厅服务器中均存在关联文件。")
                                    else:
                                        st.warning("⚠️ 发现潜在缺片风险！请检查以下影厅服务器：")
                                        # 这里使用 st.code 展示多行带编号的文本
                                        st.code(logs_text, language="text")

                                except Exception as e:
                                    st.error(f"核对过程中发生错误: {e}")

    with tab_sales:
        # 卖品品类分析表 UI
        col_1, col_2 = st.columns(2)
        with col_1:
            sales_from_file = st.checkbox("从`商品销售汇总报表-已退减.xlsx`导入数据", key="sales_file_cb")

        if sales_from_file:
            st.header("从本地文件导入数据")
            st.write("请上传 `商品销售汇总报表-已退减.xlsx` 文件。")
            uploaded_file = st.file_uploader("上传 Excel 文件", type=['xlsx', 'xls'], label_visibility="collapsed",
                                             key="sales_file_uploader")
            if uploaded_file is not None:
                with st.spinner("正在读取并分析文件..."):
                    try:
                        df = pd.read_excel(uploaded_file, skiprows=3)
                        final_summary, copy_text = process_sales_data(df)
                        if final_summary is not None:
                            st.markdown("#### 销售总览 (套餐 Top 10 + 单品 Top 5)")
                            st.dataframe(final_summary, use_container_width=True, hide_index=True)
                            st.markdown("##### 复制到 Excel")
                            st.code(copy_text, language='text')
                    except Exception as e:
                        st.error(f"处理文件时发生错误: {e}")
        else:
            st.header("从服务器API获取实时数据")
            sales_date = st.date_input("选择要查询的日期", value=datetime.now().date(), key="sales_date_picker")
            if st.button("获取卖品销售数据", key="fetch_sales_data", icon="🫵"):
                with st.spinner(f"正在获取 {sales_date} 的销售数据..."):
                    api_results = get_sales_data_with_token_management(sales_date)
                    if api_results is not None:
                        st.toast(f"成功获取到 {len(api_results)} 条销售记录！", icon="✅")
                        df = transform_api_data_to_df(api_results)
                        final_summary, copy_text = process_sales_data(df)
                        if final_summary is not None:
                            st.markdown("#### 销售总览 (套餐 Top 10 + 单品 Top 5)")
                            st.dataframe(final_summary, use_container_width=True, hide_index=True)
                            st.markdown("##### 复制到 Excel")
                            st.code(copy_text, language='text')
                    else:
                        st.warning("未能从 API 获取到数据，请检查登录或网络连接。")

    with tab_report:
        # 影片映出日累计表 UI
        st.header("影片映出日累计报表生成")
        report_date = st.date_input("选择要查询的排片日期", value=datetime.now().date(), key="daily_report_date")

        if st.button("获取并生成报表", key="fetch_daily_report", icon="🫵"):
            report_date_str = report_date.strftime('%Y-%m-%d')
            with st.spinner(f"正在获取 {report_date_str} 的排片数据..."):

                token_data = load_token()
                if not token_data: token_data = login_and_get_token()
                token = token_data.get('token') if token_data else None

                schedule, halls_map = get_api_data_with_token_management(report_date_str)
                if schedule is not None and halls_map is not None:
                    processed_df = process_and_filter_data_for_report(schedule, halls_map, report_date_str, token)
                    st.session_state.daily_report_df = processed_df
                    if not processed_df.empty:
                        st.toast(f"成功获取并处理了 {len(processed_df)} 条有效场次数据！", icon="✅")
                else:
                    st.session_state.daily_report_df = pd.DataFrame()

        if not st.session_state.daily_report_df.empty:
            st.markdown(f"#### {report_date.strftime('%Y-%m-%d')} 影片映出日累计报表")
            st.dataframe(
                st.session_state.daily_report_df.style.format({
                    '人数合计': '{:,.0f}', '座位数': '{:,.0f}', '上座率%': '{:.2f}%'
                }),
                use_container_width=True, hide_index=True
            )
            import io
            output_buffer = io.BytesIO()
            st.session_state.daily_report_df.to_excel(output_buffer, index=False, engine='openpyxl')
            excel_data = output_buffer.getvalue()
            st.download_button(
                label="📥 下载 XLSX 报表文件",
                data=excel_data,
                file_name=f"{report_date.strftime('%Y-%m-%d')}_影片映出日累计报表.xlsx",
                mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
            )
        else:
            st.info("请选择日期并点击“获取并生成报表”以生成数据。")

    with tab_print:
        # 场次与散场时间快捷打印 UI
        st.header("场次与散场时间快捷打印")

        with st.expander("⚙️ 显示与打印设置", expanded=False):
            col1, col2 = st.columns(2)
            with col1:
                st.subheader("⚙️ 修改 LED 屏排片表设置")
                led_font_name = st.selectbox("字体选择", options=list(AVAILABLE_FONTS.keys()), index=0, key="led_font")
                led_font_path = AVAILABLE_FONTS.get(led_font_name)
                generate_png_led = st.checkbox("生成 PNG 图片", value=False, key="png_led")
            with col2:
                st.subheader("⚙️ 散场时间表设置")
                times_font_name = st.selectbox("字体选择", options=list(AVAILABLE_FONTS.keys()), index=1,
                                               key="times_font")
                times_font_path = AVAILABLE_FONTS.get(times_font_name)
                font_size_multiplier = st.slider("字体大小调节", min_value=0.8, max_value=1.5, value=1.2, step=0.05,
                                                 help="调整字体在单元格内的相对大小")
                split_time = st.time_input("白班 / 晚班分割时间", value=dt_time(17, 0),
                                           help="散场时间在此时间点之前（含）的为白班")
                time_adjustment = st.slider("时间提前 (分钟)", min_value=0, max_value=10, value=0,
                                            help="将所有散场时间提前 N 分钟显示")
                hall_display_format = st.radio("影厅号格式", options=['Default', 'Superscript', 'Circled'],
                                               format_func=lambda x:
                                               {'Default': '默认 (2 18:28)', 'Superscript': '上标 (2# 18:28)',
                                                'Circled': '带圈 (② 18:28)'}[x], horizontal=True)
                generate_png_times = st.checkbox("生成 PNG 图片", value=False, key="png_times")

        with st.expander("💡 使用帮助", expanded=False):
            st.markdown("""
            #### 🖨️ 功能简介
            本工具用于将影院的排期数据快速转换为两种形式的打印页：
            1.  **修改 LED 屏幕排片表打印**：A4 竖版，详细列出影厅、场次、影片名、拼音缩写和时间范围，方便员工在修改 LED 屏幕时快速查阅和输入。
            2.  **散场时间打印**：A5 竖版，以大字体分栏显示各影厅的散场时间，方便员工在疏散人群和清洁影厅时查阅。

            #### ⬇️ 操作步骤
            1.  **选择数据源**：
                * **从文件导入**：导出 `放映时间核对表.xls` 后，点击 "Browse files" 按钮上传。
                * **从 API 获取**：选择日期，点击 "获取排片数据" 按钮，程序将自动登录并拉取最新数据。
            2.  **调整设置 (可选)**：点击上方的 "显示与打印设置" 打开设置面板，根据需要调整字体、大小、格式等。
            3.  **预览与打印**：
                * 数据加载成功后，下方会自动生成预览。
                * 默认显示 **PDF 预览**，这是最适合打印的格式。可以直接在预览界面点击 🖨️ 打印按钮。
            """)

        print_tab1, print_tab2 = st.tabs(["☁️ 从 API 获取", "📁 从文件导入"])

        with print_tab2:
            uploaded_file_print = st.file_uploader("请上传 `放映时间核对表.xls` 文件", type=["xls"],
                                                   key="print_file_uploader")
            if uploaded_file_print:
                with st.spinner("正在处理文件，请稍候..."):
                    led_data, times_part1, times_part2, date_str = process_file_upload(uploaded_file_print, split_time,
                                                                                       time_adjustment)
                    st.session_state.processed_print_data = {
                        "led_data": led_data,
                        "times_part1": times_part1,
                        "times_part2": times_part2,
                        "date_str": date_str
                    }
                    if date_str:
                        st.toast(f"文件处理完成！排期日期：**{date_str}**", icon="🎉")

        with print_tab1:
            # 修改 value 为当前日期 + 1天
            print_api_date = st.date_input("选择要查询的排片日期", value=datetime.now().date() + timedelta(days=1),
                                           key="print_api_date_picker")
            if st.button("获取排片数据", key="fetch_print_api_data", icon="🫵"):
                with st.spinner(f"正在获取 {print_api_date} 的排片数据..."):
                    date_str_api = print_api_date.strftime('%Y-%m-%d')
                    # 重用 app2.py 中现有的 API 获取逻辑，不需要重复写 fetch_schedule_data
                    schedule_list, _ = get_api_data_with_token_management(date_str_api)

                    if schedule_list is not None and len(schedule_list) > 0:
                        df_api = pd.DataFrame(schedule_list)
                        df_api.rename(columns={'hallName': 'Hall', 'showStartTime': 'StartTime',
                                               'showEndTime': 'EndTime', 'movieName': 'Movie'}, inplace=True)
                        df_api = df_api[['Hall', 'StartTime', 'EndTime', 'Movie']]

                        led_data, times_part1, times_part2 = process_schedule_df(df_api, print_api_date, split_time,
                                                                                 time_adjustment)
                        st.session_state.processed_print_data = {
                            "led_data": led_data,
                            "times_part1": times_part1,
                            "times_part2": times_part2,
                            "date_str": date_str_api
                        }
                        st.toast(f"成功获取 {len(schedule_list)} 条排片数据！", icon="✅")
                    elif schedule_list is not None:
                        st.warning("成功连接API，但当天没有排片数据。")
                        st.session_state.processed_print_data = None
                    else:
                        st.error("获取API数据失败。")
                        st.session_state.processed_print_data = None

        # --- 显示打印预览结果 ---
        if st.session_state.processed_print_data:
            data = st.session_state.processed_print_data
            led_data, times_part1, times_part2, date_str = data["led_data"], data["times_part1"], data["times_part2"], \
            data["date_str"]


            # 显示 LED 屏排片表
            st.header("🖥️ 修改 LED 屏幕排片表打印")
            if led_data is not None and not led_data.empty:
                led_output = create_print_layout_led(led_data, date_str, led_font_path, generate_png_led)
                if led_output:
                    tabs = ["PDF 预览"]
                    if 'png' in led_output: tabs.append("PNG 预览")
                    tab_views = st.tabs(tabs)
                    with tab_views[0]:
                        st.markdown(display_pdf(led_output['pdf']), unsafe_allow_html=True)
                    if 'png' in led_output:
                        with tab_views[1]: st.image(led_output['png'], use_container_width=True)
            else:
                st.error("未能成功生成 '修改 LED 屏排片表'。请检查数据源。")

            # 显示散场时间快捷打印
            st.header("🔚 散场时间打印")
            col1, col2 = st.columns(2)
            with col1:
                if times_part1 is not None and not times_part1.empty:
                    part1_output = create_print_layout_times(times_part1, "A", date_str, times_font_path,
                                                             font_size_multiplier, hall_display_format,
                                                             generate_png_times)
                    if part1_output:
                        tabs1 = [f"白班 (≤ {split_time.strftime('%H:%M')}) PDF 预览"]
                        if 'png' in part1_output: tabs1.append(f"白班 (≤ {split_time.strftime('%H:%M')}) PNG 预览")
                        tab_views1 = st.tabs(tabs1)
                        with tab_views1[0]:
                            st.markdown(display_pdf(part1_output['pdf']), unsafe_allow_html=True)
                        if 'png' in part1_output:
                            with tab_views1[1]: st.image(part1_output['png'])
                else:
                    st.info(f"白班 (≤ {split_time.strftime('%H:%M')}) 没有排期数据。")

            with col2:
                if times_part2 is not None and not times_part2.empty:
                    part2_output = create_print_layout_times(times_part2, "C", date_str, times_font_path,
                                                             font_size_multiplier, hall_display_format,
                                                             generate_png_times)
                    if part2_output:
                        tabs2 = [f"晚班 (> {split_time.strftime('%H:%M')}) PDF 预览"]
                        if 'png' in part2_output: tabs2.append(f"晚班 (> {split_time.strftime('%H:%M')}) PNG 预览")
                        tab_views2 = st.tabs(tabs2)
                        with tab_views2[0]:
                            st.markdown(display_pdf(part2_output['pdf']), unsafe_allow_html=True)
                        if 'png' in part2_output:
                            with tab_views2[1]: st.image(part2_output['png'])
                else:
                    st.info(f"晚班 (> {split_time.strftime('%H:%M')}) 没有排期数据。")
        else:
            st.info("👆 请先从文件或API加载数据以生成预览。")

    with tab3:
        st.header("TMS 服务器影片内容查询")
        if st.button('点击查询 TMS 服务器', key="query_tms", icon="🫵"):
            with st.spinner("正在从 TMS 服务器获取数据中..."):
                try:
                    priority_titles, df_for_tms = [], pd.DataFrame()
                    if not st.session_state.api_df.empty:
                        df_for_tms = st.session_state.api_df
                    elif not st.session_state.file_df.empty:
                        df_for_tms = st.session_state.file_df
                    if not df_for_tms.empty:
                        # 优先使用清洗后的名字
                        if '影片名称_清理后' in df_for_tms.columns:
                            priority_titles = df_for_tms['影片名称_清理后'].unique().tolist()
                        else:
                            priority_titles = df_for_tms['影片名称'].apply(
                                lambda x: clean_movie_title(x)).unique().tolist()

                    halls_data, movie_list_sorted = fetch_and_process_server_movies(priority_titles)
                    st.toast("TMS 服务器数据获取成功！", icon="🎉")
                    st.markdown("#### 按影片查看所在影厅")
                    view2_data = [{'影片名称': item['assert_name'],
                                   '所在影厅': " ".join(sorted([get_circled_number(h) for h in item['halls']])),
                                   '文件名': item['content_name'], '时长(分钟)': format_play_time(item['play_time'])}
                                  for item in movie_list_sorted]
                    st.dataframe(pd.DataFrame(view2_data), hide_index=True, use_container_width=True)
                    st.markdown("#### 按影厅查看影片内容")
                    hall_tabs = st.tabs(list(halls_data.keys()))
                    for tab, hall_name in zip(hall_tabs, halls_data.keys()):
                        with tab:
                            view1_data = [{'影片名称': item['details']['assert_name'], '所在影厅': " ".join(
                                sorted([get_circled_number(h) for h in item['details']['halls']])),
                                           '文件名': item['content_name'],
                                           '时长(分钟)': format_play_time(item['details']['play_time'])} for item in
                                          halls_data[hall_name]]
                            st.dataframe(pd.DataFrame(view1_data), hide_index=True, use_container_width=True)
                except Exception as e:
                    st.error(f"查询TMS服务器时出错: {e}")


if __name__ == "__main__":
    main()