app.py

import pandas as pd
import streamlit as st
import matplotlib.pyplot as plt
import matplotlib.font_manager as font_manager
import io
import base64
import os
from datetime import datetime, timedelta
import math
from pypinyin import lazy_pinyin, Style
from matplotlib.backends.backend_pdf import PdfPages
import matplotlib.gridspec as gridspec
from matplotlib.patches import FancyBboxPatch

# --- Constants for "Quick Print" (放映场次核对表) ---
SPLIT_TIME = "17:30"
BUSINESS_START = "09:30"
BUSINESS_END = "01:30"
BORDER_COLOR = '#A9A9A9'
DATE_COLOR = '#A9A9A9'

# --- Helper functions for "LED Screen" (放映时间核对表) ---
def get_font(size=14):
    """Loads a specific font file, falling back to a default if not found."""
    font_path = "simHei.ttc"
    if not os.path.exists(font_path):
        font_path = "SimHei.ttf"  # Fallback font
    # Add a final fallback for systems without Chinese fonts
    try:
        return font_manager.FontProperties(fname=font_path, size=size)
    except RuntimeError:
        # If the font file is not found, use a default font that should exist.
        # This will likely not render Chinese characters correctly but prevents crashing.
        return font_manager.FontProperties(family='sans-serif', size=size)


def get_pinyin_abbr(text):
    """Gets the first letter of the Pinyin for the first two Chinese characters of a text."""
    if not text:
        return ""
    # Extract the first two Chinese characters
    chars = [c for c in text if '\u4e00' <= c <= '\u9fff']
    chars = chars[:2]
    # Get the first letter of the pinyin for each character
    pinyin_list = lazy_pinyin(chars, style=Style.FIRST_LETTER)
    return ''.join(pinyin_list).upper()

# --- Processing logic for "LED Screen" (放映时间核对表) ---
def process_schedule_led(file):
    """Processes the '放映时间核对表.xls' file."""
    try:
        # Attempt to read the date from a specific cell
        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:
        # Fallback to the current date if reading fails
        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)
        df['Hall'] = df['Hall'].astype(str).str.extract(r'(\d+号)')

        # Convert times to datetime objects, handling overnight screenings
        df['StartTime_dt'] = pd.to_datetime(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
        )
        df['EndTime_dt'] = pd.to_datetime(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
        )
        df.loc[df['EndTime_dt'] < df['StartTime_dt'], 'EndTime_dt'] += timedelta(days=1)
        df = df.sort_values(['Hall', 'StartTime_dt'])

        # Merge consecutive screenings of the same movie
        merged_rows = []
        for hall, group in df.groupby('Hall'):
            group = group.sort_values('StartTime_dt')
            current = None
            for _, row in group.iterrows():
                if current is None:
                    current = row.copy()
                else:
                    if 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)

        # Adjust start and end times
        merged_df['StartTime_dt'] = merged_df['StartTime_dt'] - timedelta(minutes=10)
        merged_df['EndTime_dt'] = merged_df['EndTime_dt'] - timedelta(minutes=5)

        merged_df['StartTime_str'] = merged_df['StartTime_dt'].dt.strftime('%H:%M')
        merged_df['EndTime_str'] = merged_df['EndTime_dt'].dt.strftime('%H:%M')

        return merged_df[['Hall', 'Movie', 'StartTime_str', 'EndTime_str']], date_str
    except Exception as e:
        st.error(f"An error occurred during file processing: {e}")
        return None, date_str

# --- Layout generation for "LED Screen" (放映时间核对表) ---
def create_print_layout_led(data, date_str):
    """Generates PNG and PDF layouts for the 'LED Screen' schedule."""
    if data is None or data.empty:
        return None

    # Create figures for PNG and PDF output with A4 dimensions
    png_fig = plt.figure(figsize=(8.27, 11.69), dpi=300)
    png_ax = png_fig.add_subplot(111)
    png_ax.set_axis_off()
    png_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)

    pdf_fig = plt.figure(figsize=(8.27, 11.69), dpi=300)
    pdf_ax = pdf_fig.add_subplot(111)
    pdf_ax.set_axis_off()
    pdf_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)

    def process_figure(fig, ax):
        halls = sorted(data['Hall'].unique(), key=lambda h: int(h.replace('号','')) if h else 0)

        num_separators = len(halls) - 1
        total_layout_rows = len(data) + num_separators + 2
        
        available_height = 0.96
        row_height = available_height / total_layout_rows

        fig_height_inches = fig.get_figheight()
        row_height_points = row_height * fig_height_inches * 72
        font_size = row_height_points * 0.9

        date_font = get_font(font_size * 0.8)
        hall_font = get_font(font_size)
        movie_font = get_font(font_size)

        col_hall_left = 0.0
        col_movie_right = 0.50
        col_seq_left = 0.52
        col_pinyin_left = 0.62
        col_time_left = 0.75

        ax.text(col_hall_left, 0.99, date_str, color='#A9A9A9',
                ha='left', va='top', fontproperties=date_font, transform=ax.transAxes)
        
        y_position = 0.98 - row_height
        
        for i, hall in enumerate(halls):
            hall_data = data[data['Hall'] == hall]
            
            if i > 0:
                ax.axhline(y=y_position + row_height / 2, xmin=col_hall_left, xmax=0.97, color='black', linewidth=0.7)
                y_position -= row_height

            movie_count = 1
            for _, row in hall_data.iterrows():
                if movie_count == 1:
                    ax.text(col_hall_left, y_position, f"{hall.replace('号', '')}#",
                            ha='left', va='center', fontweight='bold',
                            fontproperties=hall_font, transform=ax.transAxes)
                
                ax.text(col_movie_right, y_position, row['Movie'],
                        ha='right', va='center', fontproperties=movie_font, transform=ax.transAxes)

                ax.text(col_seq_left, y_position, f"{movie_count}.",
                        ha='left', va='center', fontproperties=movie_font, transform=ax.transAxes)

                pinyin_abbr = get_pinyin_abbr(row['Movie'])
                ax.text(col_pinyin_left, y_position, pinyin_abbr,
                        ha='left', va='center', fontproperties=movie_font, transform=ax.transAxes)
                
                ax.text(col_time_left, y_position, f"{row['StartTime_str']}-{row['EndTime_str']}",
                        ha='left', va='center', fontproperties=movie_font, transform=ax.transAxes)

                y_position -= row_height
                movie_count += 1

    process_figure(png_fig, png_ax)
    process_figure(pdf_fig, pdf_ax)
    
    png_buffer = io.BytesIO()
    png_fig.savefig(png_buffer, format='png', bbox_inches='tight', pad_inches=0.05)
    png_buffer.seek(0)
    image_base64 = base64.b64encode(png_buffer.getvalue()).decode()
    plt.close(png_fig)
    
    pdf_buffer = io.BytesIO()
    with PdfPages(pdf_buffer) as pdf:
        pdf.savefig(pdf_fig, bbox_inches='tight', pad_inches=0.05)
    pdf_buffer.seek(0)
    pdf_base64 = base64.b64encode(pdf_buffer.getvalue()).decode()
    plt.close(pdf_fig)
    
    return {
        'png': f"data:image/png;base64,{image_base64}",
        'pdf': f"data:application/pdf;base64,{pdf_base64}"
    }

# --- Processing logic for "Quick Print" (放映场次核对表) ---
def process_schedule_quick(file):
    """Processes the '放映场次核对表.xls' file."""
    try:
        df = pd.read_excel(file, skiprows=8)
        df = df.iloc[:, [6, 7, 9]]
        df.columns = ['Hall', 'StartTime', 'EndTime']
        df = df.dropna(subset=['Hall', 'StartTime', 'EndTime'])
        df['Hall'] = df['Hall'].str.extract(r'(\d+)号').astype(str) + ' '
        
        base_date = datetime.today().date()
        df['StartTime'] = pd.to_datetime(df['StartTime'])
        df['EndTime'] = pd.to_datetime(df['EndTime'])

        business_start = datetime.strptime(f"{base_date} {BUSINESS_START}", "%Y-%m-%d %H:%M")
        business_end = datetime.strptime(f"{base_date} {BUSINESS_END}", "%Y-%m-%d %H:%M")
        if business_end < business_start:
            business_end += timedelta(days=1)

        for idx, row in df.iterrows():
            end_time = row['EndTime']
            if end_time.hour < 9:
                df.at[idx, 'EndTime'] = end_time + timedelta(days=1)
            if row['StartTime'].hour >= 21 and end_time.hour < 9:
                df.at[idx, 'EndTime'] = end_time + timedelta(days=1)
        
        df['time_for_comparison'] = df['EndTime'].apply(lambda x: datetime.combine(base_date, x.time()))
        df.loc[df['time_for_comparison'].dt.hour < 9, 'time_for_comparison'] += timedelta(days=1)
        
        valid_times = (
            (df['time_for_comparison'] >= datetime.combine(base_date, business_start.time())) &
            (df['time_for_comparison'] <= datetime.combine(base_date + timedelta(days=1), business_end.time()))
        )
        df = df[valid_times]
        df = df.sort_values('EndTime')
        
        split_time_dt = datetime.strptime(f"{base_date} {SPLIT_TIME}", "%Y-%m-%d %H:%M")
        
        part1 = df[df['time_for_comparison'] <= split_time_dt].copy()
        part2 = df[df['time_for_comparison'] > split_time_dt].copy()

        for part in [part1, part2]:
            part['EndTime'] = part['EndTime'].dt.strftime('%-H:%M')
        
        date_df = pd.read_excel(file, skiprows=5, nrows=1, usecols=[2], header=None)
        date_cell = date_df.iloc[0, 0]
        
        try:
            if isinstance(date_cell, str):
                date_str = datetime.strptime(date_cell, '%Y-%m-%d').strftime('%Y-%m-%d')
            else:
                date_str = pd.to_datetime(date_cell).strftime('%Y-%m-%d')
        except:
            date_str = datetime.today().strftime('%Y-%m-%d')
            
        return part1[['Hall', 'EndTime']], part2[['Hall', 'EndTime']], date_str

    except Exception as e:
        st.error(f"处理文件时出错: {str(e)}")
        return None, None, None

# --- Layout generation for "Quick Print" (放映场次核对表) ---
def create_print_layout_quick(data, title, date_str):
    """Creates print layout for the 'Quick Print' schedule."""
    if data.empty:
        return None

    png_fig = plt.figure(figsize=(5.83, 8.27), dpi=300) # A5
    png_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)

    pdf_fig = plt.figure(figsize=(5.83, 8.27), dpi=300) # A5
    pdf_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)

    def process_figure(fig, is_pdf=False):
        plt.rcParams['font.family'] = 'sans-serif'
        plt.rcParams['font.sans-serif'] = ['Arial Unicode MS', 'Heiti TC', 'SimHei']

        total_items = len(data)
        num_cols = 3
        num_rows = math.ceil(total_items / num_cols)

        gs = gridspec.GridSpec(num_rows + 1, num_cols, hspace=0.05, wspace=0.05, height_ratios=[0.1] + [1] * num_rows, figure=fig)

        target_width_px = 1
        if total_items > 0:
            ax_temp = fig.add_subplot(gs[1, 0])
            fig.canvas.draw()
            target_width_px = ax_temp.get_window_extent().width * 0.90
            ax_temp.remove()

        available_height_per_row = (8.27 * 0.9 * (1 / 1.2)) / num_rows if num_rows > 0 else 1
        date_fontsize = min(40, max(10, available_height_per_row * 72 * 0.5))

        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 = i % rows_per_col_layout
                col_idx = 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

        for idx, (hall, end_time) in enumerate(sorted_data):
            if hall and end_time:
                row_grid = idx // num_cols + 1
                col_grid = idx % num_cols

                if row_grid < num_rows + 1:
                    ax = fig.add_subplot(gs[row_grid, col_grid])
                    for spine in ax.spines.values():
                        spine.set_visible(False)

                    bbox = FancyBboxPatch(
                        (0.01, 0.01), 0.98, 0.98,
                        boxstyle="round,pad=0,rounding_size=0.02",
                        edgecolor=BORDER_COLOR, facecolor='none',
                        linewidth=0.5, transform=ax.transAxes, clip_on=False
                    )
                    ax.add_patch(bbox)

                    display_text = f"{hall}{end_time}"
                    t = ax.text(0.5, 0.5, display_text,
                                fontweight='bold', ha='center', va='center',
                                transform=ax.transAxes)
                    
                    current_size = 120
                    while current_size > 1:
                        t.set_fontsize(current_size)
                        text_bbox = t.get_window_extent(renderer=fig.canvas.get_renderer())
                        if text_bbox.width <= target_width_px:
                            break
                        current_size -= 2

                    ax.set_xticks([])
                    ax.set_yticks([])

        ax_date = fig.add_subplot(gs[0, :])
        ax_date.text(0.01, 0.5, f"{date_str} {title}",
                     fontsize=date_fontsize * 0.5,
                     color=DATE_COLOR, fontweight='bold',
                     ha='left', va='center', transform=ax_date.transAxes)
        for spine in ax_date.spines.values():
            spine.set_visible(False)
        ax_date.set_xticks([])
        ax_date.set_yticks([])
        ax_date.set_facecolor('none')

    process_figure(png_fig)
    process_figure(pdf_fig, is_pdf=True)

    png_buffer = io.BytesIO()
    png_fig.savefig(png_buffer, format='png', bbox_inches='tight', pad_inches=0.02)
    png_buffer.seek(0)
    png_base64 = base64.b64encode(png_buffer.getvalue()).decode()
    plt.close(png_fig)

    pdf_buffer = io.BytesIO()
    with PdfPages(pdf_buffer) as pdf:
        pdf.savefig(pdf_fig, bbox_inches='tight', pad_inches=0.02)
    pdf_buffer.seek(0)
    pdf_base64 = base64.b64encode(pdf_buffer.getvalue()).decode()
    plt.close(pdf_fig)

    return {
        'png': f'data:image/png;base64,{png_base64}',
        'pdf': f'data:application/pdf;base64,{pdf_base64}'
    }

# --- Generic Helper to Display PDF ---
def display_pdf(base64_pdf):
    """Generates the HTML to embed and display a PDF in Streamlit."""
    pdf_display = f"""
    <iframe src="{base64_pdf}" width="100%" height="800" type="application/pdf"></iframe>
    """
    return pdf_display

# --- Main Streamlit App ---
st.set_page_config(page_title="影院排期打印工具", layout="wide")
st.title("影院排期打印工具")

uploaded_file = st.file_uploader(
    "选择【放映时间核对表.xls】或【放映场次核对表.xls】文件",
    accept_multiple_files=False,
    type=["xls"]
)

if uploaded_file:
    with st.spinner("文件正在处理中，请稍候..."):
        # --- Route to the correct processor based on filename ---

        # 1. Logic for "LED 屏幕时间表打印"
        if "放映时间核对表" in uploaded_file.name:
            st.subheader("LED 屏幕时间表")
            schedule, date_str = process_schedule_led(uploaded_file)
            if schedule is not None:
                output = create_print_layout_led(schedule, date_str)
                if output:
                    tab1, tab2 = st.tabs(["PDF 预览", "PNG 预览"])
                    with tab1:
                        st.markdown(display_pdf(output['pdf']), unsafe_allow_html=True)
                    with tab2:
                        st.image(output['png'], use_container_width=True)
                else:
                    st.info("没有可显示的数据。")
            else:
                st.error("无法处理文件，请检查文件格式或内容是否正确。")

        # 2. Logic for "散厅时间快捷打印"
        elif "放映场次核对表" in uploaded_file.name:
            part1_data, part2_data, date_str = process_schedule_quick(uploaded_file)

            if part1_data is not None and part2_data is not None:
                part1_output = create_print_layout_quick(part1_data, "A", date_str)
                part2_output = create_print_layout_quick(part2_data, "C", date_str)

                col1, col2 = st.columns(2)

                with col1:
                    st.subheader("白班散场预览（时间 ≤ 17:30）")
                    if part1_output:
                        tab1_1, tab1_2 = st.tabs(["PDF 预览 ", "PNG 预览 "]) # Added space to make keys unique
                        with tab1_1:
                            st.markdown(display_pdf(part1_output['pdf']), unsafe_allow_html=True)
                        with tab1_2:
                            st.image(part1_output['png'])
                    else:
                        st.info("白班部分没有数据")

                with col2:
                    st.subheader("夜班散场预览（时间 > 17:30）")
                    if part2_output:
                        tab2_1, tab2_2 = st.tabs(["PDF 预览  ", "PNG 预览  "]) # Added spaces to make keys unique
                        with tab2_1:
                            st.markdown(display_pdf(part2_output['pdf']), unsafe_allow_html=True)
                        with tab2_2:
                            st.image(part2_output['png'])
                    else:
                        st.info("夜班部分没有数据")
            else:
                st.error("无法处理文件，请检查文件格式或内容是否正确。")
        
        # 3. Fallback for incorrect file
        else:
            st.warning("文件名不匹配。请上传名为【放映时间核对表.xls】或【放映场次核对表.xls】的文件。")