Update app.py

app.py

@@ -4,66 +4,90 @@ from datetime import datetime, timedelta
 import matplotlib.pyplot as plt
 import io
 import base64
+import matplotlib.gridspec as gridspec
+import matplotlib.font_manager as fm
 import math
-from matplotlib.backends.backend_pdf import PdfPages
 
-# --- Constants ---
+# --- CONSTANTS ---
 SPLIT_TIME = "17:30"
 BUSINESS_START = "09:30"
 BUSINESS_END = "01:30"
-BORDER_COLOR = 'grey' # Changed to grey for the new design
+BORDER_COLOR = '#A9A9A9'
 DATE_COLOR = '#A9A9A9'
+A5_WIDTH_IN = 5.83
+A5_HEIGHT_IN = 8.27
+DPI = 300
+NUM_COLS = 3
+
+# --- FONT SETUP ---
+# Attempt to load the specified font, with a safe fallback.
+try:
+    # IMPORTANT: Place 'SimHei.ttf' in the same directory as the script.
+    FONT_PROP = fm.FontProperties(fname='SimHei.ttf')
+except FileNotFoundError:
+    st.warning("SimHei.ttf not found. Using a default sans-serif font. Chinese characters may not display correctly.")
+    FONT_PROP = fm.FontProperties(family='sans-serif')
+
 
 def process_schedule(file):
-    """
-    Processes the uploaded Excel file to extract, clean, and sort screening times.
-    """
+    """处理上传的 Excel 文件，生成排序和分组后的打印内容"""
     try:
-        # Read Excel, skipping header rows
+        # 读取 Excel，跳过前 8 行
         df = pd.read_excel(file, skiprows=8)
 
-        # Extract required columns (G, H, J)
-        df = df.iloc[:, [6, 7, 9]]
+        # 提取所需列 (G9, H9, J9)
+        df = df.iloc[:, [6, 7, 9]]  # G, H, J 列
         df.columns = ['Hall', 'StartTime', 'EndTime']
 
-        # Clean data: drop rows with missing values
+        # 清理数据
         df = df.dropna(subset=['Hall', 'StartTime', 'EndTime'])
+        df = df[df['Hall'].astype(str).str.contains(r'\d', na=False)] # Ensure Hall has a digit
 
-        # Format Hall to "# " format
-        df['Hall'] = df['Hall'].str.extract(r'(\d+)').astype(str) + ' '
+        # 转换影厅格式为 "#号" 格式
+        df['Hall'] = df['Hall'].astype(str).str.extract(r'(\d+)').astype(str) + ' '
 
-        # Convert time columns to datetime objects
+        # 转换时间为 datetime 对象
         base_date = datetime.today().date()
-        df['StartTime'] = pd.to_datetime(df['StartTime'])
-        df['EndTime'] = pd.to_datetime(df['EndTime'])
+        df['StartTime'] = pd.to_datetime(df['StartTime'], errors='coerce').dt.time
+        df['EndTime'] = pd.to_datetime(df['EndTime'], errors='coerce').dt.time
 
-        # --- Handle overnight screenings ---
-        # If a show ends after midnight (e.g., 1:30 AM), it belongs to the previous day's schedule.
-        # We handle this by adding a day to its datetime object.
-        for idx, row in df.iterrows():
-            if row['EndTime'].hour < 9: # Assuming any end time before 9 AM is part of the previous night
-                df.at[idx, 'EndTime'] = row['EndTime'] + timedelta(days=1)
+        df = df.dropna(subset=['StartTime', 'EndTime'])
 
-        # Create a comparable time column that correctly handles the business day logic
-        df['time_for_comparison'] = df['EndTime']
+        def combine_date_time(t):
+            dt = datetime.combine(base_date, t)
+            # Handle overnight shows (times past midnight like 1:30 AM are for the next day)
+            if t.hour < int(BUSINESS_START.split(':')[0]) - 1:
+                return dt + timedelta(days=1)
+            return dt
 
-        # Sort screenings by their end time
-        df = df.sort_values('EndTime')
+        df['StartDateTime'] = df['StartTime'].apply(combine_date_time)
+        df['EndDateTime'] = df['EndTime'].apply(combine_date_time)
 
-        # Split data into day and night shifts
-        split_datetime = datetime.combine(base_date, datetime.strptime(SPLIT_TIME, "%H:%M").time())
-        part1 = df[df['time_for_comparison'] <= split_datetime].copy()
-        part2 = df[df['time_for_comparison'] > split_datetime].copy()
+        # 按散场时间排序
+        df = df.sort_values('EndDateTime')
 
-        # Format the time display string (e.g., "5:30")
-        for part in [part1, part2]:
-            part['EndTime'] = part['EndTime'].dt.strftime('%-I:%M')
+        # 分割数据
+        split_dt = datetime.combine(base_date, datetime.strptime(SPLIT_TIME, "%H:%M").time())
+        
+        part1 = df[df['EndDateTime'] <= split_dt].copy()
+        part2 = df[df['EndDateTime'] > split_dt].copy()
 
-        # Precisely read the date from cell C6
-        date_df = pd.read_excel(file, skiprows=5, nrows=1, usecols=[2], header=None)
+        # 格式化时间显示
+        for part in [part1, part2]:
+            part['EndTimeStr'] = part['EndDateTime'].dt.strftime('%-I:%M')
+
+        # 关键修改：精确读取C6单元格
+        date_df = pd.read_excel(
+            file,
+            skiprows=5,  # 跳过前5行（0-4）
+            nrows=1,     # 只读1行
+            usecols=[2], # 第三列（C列）
+            header=None  # 无表头
+        )
         date_cell = date_df.iloc[0, 0]
 
         try:
+            # 处理不同日期格式
             if isinstance(date_cell, str):
                 date_str = datetime.strptime(date_cell, '%Y-%m-%d').strftime('%Y-%m-%d')
             else:
@@ -71,153 +95,132 @@ def process_schedule(file):
         except:
             date_str = datetime.today().strftime('%Y-%m-%d')
 
-        return part1[['Hall', 'EndTime']], part2[['Hall', 'EndTime']], date_str
+        return part1[['Hall', 'EndTimeStr']], part2[['Hall', 'EndTimeStr']], date_str
 
     except Exception as e:
-        st.error(f"Error processing file: {str(e)}")
+        st.error(f"处理文件时出错: {str(e)}")
         return None, None, None
 
-def _draw_grid_on_figure(fig, data, title, date_str):
-    """
-    Internal helper function to draw the new grid layout onto a Matplotlib figure.
-    """
-    plt.rcParams['font.family'] = 'sans-serif'
-    plt.rcParams['font.sans-serif'] = ['Arial Unicode MS'] # Font that supports Chinese characters
-
-    total_items = len(data)
-    if total_items == 0:
-        return
-
-    num_cols = 3
-    num_rows = math.ceil(total_items / num_cols)
-
-    A5_WIDTH_IN = 5.83
-    A5_HEIGHT_IN = 8.27
-
-    # 1. Redesign layout based on precise grid calculations
-    margin_y = (A5_HEIGHT_IN / num_rows) / 4
-    margin_x = margin_y # Use symmetric margins for a cleaner look
-
-    # Prevent margins from becoming too large on pages with few items
-    if A5_WIDTH_IN < 2 * margin_x or A5_HEIGHT_IN < 2 * margin_y:
-        margin_x = A5_WIDTH_IN / 10
-        margin_y = A5_HEIGHT_IN / 10
-
-    printable_width = A5_WIDTH_IN - 2 * margin_x
-    printable_height = A5_HEIGHT_IN - 2 * margin_y
-    cell_width = printable_width / num_cols
-    cell_height = printable_height / num_rows
-
-    # Prepare data: Sort into column-first (Z-style) order for layout
-    data_values = data.values.tolist()
-    while len(data_values) % num_cols != 0:
-        data_values.append(['', '']) # Pad data for a full grid
-    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
-
-    # --- Draw each cell onto the figure ---
-    item_counter = 0
-    for idx, (hall, end_time) in enumerate(sorted_data):
-        if not (hall and end_time):
-            continue
-
-        item_counter += 1
-        row_grid = idx // num_cols
-        col_grid = idx % num_cols
-
-        # Calculate position for each cell's axes in Figure coordinates [left, bottom, width, height]
-        ax_left_in = margin_x + col_grid * cell_width
-        ax_bottom_in = margin_y + (num_rows - 1 - row_grid) * cell_height # Y-axis from bottom
-        ax_pos = [
-            ax_left_in / A5_WIDTH_IN,
-            ax_bottom_in / A5_HEIGHT_IN,
-            cell_width / A5_WIDTH_IN,
-            cell_height / A5_HEIGHT_IN,
-        ]
-        ax = fig.add_axes(ax_pos)
-
-        # 2. Change Cell Border to a gray, dotted line
-        for spine in ax.spines.values():
-            spine.set_visible(True)
-            spine.set_linestyle(':') # Dotted line style
-            spine.set_edgecolor(BORDER_COLOR)
-            spine.set_linewidth(1)
-
-        # 4. Add Cell Index Number
-        ax.text(0.07, 0.93, str(item_counter),
-                transform=ax.transAxes,
-                fontsize=9,
-                color='grey',
-                ha='left',
-                va='top')
-
-        # 3. Adjust Cell Content
-        display_text = f"{hall}{end_time}"
-
-        # Dynamically estimate font size to fill 90% of the cell width
-        # This is a heuristic that provides a good balance of size and spacing.
-        font_scale_factor = 1.7
-        estimated_fontsize = (cell_width * 72 / len(display_text)) * font_scale_factor * 0.9
-        max_fontsize = cell_height * 72 * 0.6 # Cap font size to 60% of cell height
-        final_fontsize = min(estimated_fontsize, max_fontsize)
-
-        ax.text(0.5, 0.5, display_text,
-                fontsize=final_fontsize,
-                fontweight='bold',
-                ha='center',
-                va='center',
-                transform=ax.transAxes)
-
-        ax.set_xticks([])
-        ax.set_yticks([])
-
-    # Add date and title to the top margin of the figure
-    title_fontsize = margin_y * 72 * 0.3 # Scale title font size with margin
-    fig.text(margin_x / A5_WIDTH_IN, 1 - (margin_y * 0.5 / A5_HEIGHT_IN),
-             f"{date_str} {title}",
-             fontsize=title_fontsize,
-             color=DATE_COLOR,
-             fontweight='bold',
-             ha='left',
-             va='center')
-
 
 def create_print_layout(data, title, date_str):
-    """
-    Creates the final print-ready output in both PNG and PDF formats using the new grid layout.
-    """
+    """创建符合新要求的打印布局 (PNG 和 PDF)"""
     if data.empty:
         return None
 
-    # --- Create separate figures for PNG and PDF to ensure no cross-contamination ---
-    png_fig = plt.figure(figsize=(5.83, 8.27), dpi=300)
-    pdf_fig = plt.figure(figsize=(5.83, 8.27), dpi=300)
-
-    # --- Draw the layout on both figures ---
-    _draw_grid_on_figure(png_fig, data, title, date_str)
-    _draw_grid_on_figure(pdf_fig, data, title, date_str)
-
-    # --- Save PNG to a memory buffer ---
+    # --- 1. 动态字体大小计算 ---
+    # 目标：让最长的文本占单元格宽度的90%
+    longest_text = ""
+    if not data.empty:
+        data['combined_text'] = data['Hall'] + data['EndTimeStr']
+        if not data['combined_text'].empty:
+            longest_text = data.loc[data['combined_text'].str.len().idxmax(), 'combined_text']
+
+    # 估算单元格宽度 (A5纸张，减去边距和间距)
+    fig_width_in = A5_WIDTH_IN
+    margin_frac = 0.04  # 4% margin on each side
+    wspace_frac = 0.05  # 5% space between columns
+    
+    # 可用绘图宽度
+    drawable_width_in = fig_width_in * (1 - 2 * margin_frac)
+    # 宽度被3个子图和2个间隙瓜分
+    # 3*cell_w + 2*(cell_w*wspace) = drawable_width
+    cell_width_in = drawable_width_in / (NUM_COLS + (NUM_COLS - 1) * wspace_frac)
+    
+    main_fontsize = 30  # Default start size
+    if longest_text:
+        # 估算文本宽度 (这是一个经验法则，0.7是中英混合字体的估算因子)
+        # 文本宽度(点) ≈ 字符数 * 字体大小(点) * 因子
+        estimated_text_width_pt = len(longest_text) * main_fontsize * 0.7
+        target_width_pt = (cell_width_in * 0.9) * 72  # 目标宽度 (英寸*0.9 -> 点)
+        
+        if estimated_text_width_pt > 0:
+            ratio = target_width_pt / estimated_text_width_pt
+            main_fontsize *= ratio
+    
+    main_fontsize = max(10, min(main_fontsize, 50))  # 限制字体大小在合理范围
+    index_fontsize = main_fontsize * 0.45 # 序号字体大小
+
+    # --- 2. 准备绘图 ---
+    def generate_figure():
+        fig = plt.figure(figsize=(A5_WIDTH_IN, A5_HEIGHT_IN), dpi=DPI)
+        fig.subplots_adjust(left=margin_frac, right=1-margin_frac, top=0.95, bottom=margin_frac, hspace=0.2, wspace=wspace_frac)
+
+        total_items = len(data)
+        num_rows = math.ceil(total_items / NUM_COLS)
+
+        # 创建网格
+        gs = gridspec.GridSpec(num_rows + 1, NUM_COLS, height_ratios=[0.2] + [1] * num_rows, figure=fig)
+
+        # 绘制日期标题
+        ax_date = fig.add_subplot(gs[0, :])
+        ax_date.text(0, 0.5, f"{date_str} {title}", fontproperties=FONT_PROP,
+                     fontsize=main_fontsize * 0.5, color=DATE_COLOR, fontweight='bold',
+                     ha='left', va='center', transform=ax_date.transAxes)
+        ax_date.set_axis_off()
+
+        # 准备Z字形排列的数据
+        data_values = data[['Hall', 'EndTimeStr']].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
+        
+        item_counter = 0
+        for idx, (hall, end_time) in enumerate(sorted_data):
+            if hall and end_time:
+                item_counter += 1
+                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])
+                    
+                    # --- 3. 绘制单元格 ---
+                    # a. 设置点状虚线边框
+                    for spine in ax.spines.values():
+                        spine.set_linestyle((0, (1, 2)))  # (offset, (on, off)) tuple for dotted
+                        spine.set_edgecolor(BORDER_COLOR)
+                        spine.set_linewidth(1)
+
+                    # b. 居中绘制主要文本
+                    display_text = f"{hall}{end_time}"
+                    ax.text(0.5, 0.5, display_text, fontproperties=FONT_PROP,
+                            fontsize=main_fontsize, fontweight='bold',
+                            ha='center', va='center', transform=ax.transAxes)
+                    
+                    # c. 在左上角添加序号
+                    ax.text(0.05, 0.95, str(item_counter), fontproperties=FONT_PROP,
+                            fontsize=index_fontsize, color='grey',
+                            ha='left', va='top', transform=ax.transAxes)
+
+                    ax.set_xticks([])
+                    ax.set_yticks([])
+        return fig
+
+    # --- 4. 保存为 PNG 和 PDF ---
+    fig_for_output = generate_figure()
+
+    # 保存 PNG
     png_buffer = io.BytesIO()
-    png_fig.savefig(png_buffer, format='png', bbox_inches='tight', pad_inches=0.02)
+    fig_for_output.savefig(png_buffer, format='png', dpi=DPI)
     png_buffer.seek(0)
     png_base64 = base64.b64encode(png_buffer.getvalue()).decode()
-    plt.close(png_fig)
-
-    # --- Save PDF to a memory buffer ---
+    
+    # 保存 PDF
     pdf_buffer = io.BytesIO()
-    with PdfPages(pdf_buffer) as pdf:
-        pdf.savefig(pdf_fig, bbox_inches='tight', pad_inches=0.02)
+    fig_for_output.savefig(pdf_buffer, format='pdf', dpi=DPI)
     pdf_buffer.seek(0)
     pdf_base64 = base64.b64encode(pdf_buffer.getvalue()).decode()
-    plt.close(pdf_fig)
+
+    plt.close(fig_for_output)
 
     return {
         'png': f'data:image/png;base64,{png_base64}',
@@ -225,32 +228,30 @@ def create_print_layout(data, title, date_str):
     }
 
 def display_pdf(base64_pdf):
-    """
-    Generates the HTML to embed and display a PDF in Streamlit.
-    """
+    """在Streamlit中嵌入显示PDF"""
     pdf_display = f'<iframe src="data:application/pdf;base64,{base64_pdf}" width="100%" height="800" type="application/pdf"></iframe>'
     return pdf_display
 
-# --- Streamlit User Interface ---
+# --- Streamlit UI ---
 st.set_page_config(page_title="散厅时间快捷打印", layout="wide")
 st.title("散厅时间快捷打印")
 
 uploaded_file = st.file_uploader("上传【放映场次核对表.xls】文件", type=["xls", "xlsx"])
 
 if uploaded_file:
-    part1_data, part2_data, date_string = process_schedule(uploaded_file)
+    # Rename columns in process_schedule call to match new names
+    part1_data, part2_data, date_str = process_schedule(uploaded_file)
 
     if part1_data is not None and part2_data is not None:
-        # Generate layouts for both day and night shifts
-        part1_output = create_print_layout(part1_data, "A", date_string)
-        part2_output = create_print_layout(part2_data, "C", date_string)
+        # Pass the dataframes with renamed 'EndTimeStr' column
+        part1_output = create_print_layout(part1_data, "A", date_str)
+        part2_output = create_print_layout(part2_data, "C", date_str)
 
         col1, col2 = st.columns(2)
 
         with col1:
-            st.subheader("白班散场预览 (时间 ≤ 17:30)")
+            st.subheader("白班散场预览（结束时间 ≤ 17:30）")
             if part1_output:
-                # Use tabs for PDF and PNG previews
                 tab1_1, tab1_2 = st.tabs(["PDF 预览", "PNG 预览"])
                 with tab1_1:
                     st.markdown(display_pdf(part1_output['pdf']), unsafe_allow_html=True)
@@ -260,9 +261,8 @@ if uploaded_file:
                 st.info("白班部分没有数据")
 
         with col2:
-            st.subheader("夜班散场预览 (时间 > 17:30)")
+            st.subheader("夜班散场预览（结束时间 > 17:30）")
             if part2_output:
-                # Use tabs for PDF and PNG previews
                 tab2_1, tab2_2 = st.tabs(["PDF 预览", "PNG 预览"])
                 with tab2_1:
                     st.markdown(display_pdf(part2_output['pdf']), unsafe_allow_html=True)