Update app3.py

app3.py

@@ -7,280 +7,259 @@ import base64
 import matplotlib.gridspec as gridspec
 import math
 from matplotlib.backends.backend_pdf import PdfPages
-from matplotlib.patches import FancyBboxPatch # 新增导入
+from matplotlib.patches import Rectangle # Replaced FancyBboxPatch
 
+# --- Constants ---
 SPLIT_TIME = "17:30"
 BUSINESS_START = "09:30"
 BUSINESS_END = "01:30"
 BORDER_COLOR = '#A9A9A9'
 DATE_COLOR = '#A9A9A9'
+SEQ_COLOR = '#A9A9A9' # Color for the new serial number
 
 def process_schedule(file):
     """处理上传的 Excel 文件，生成排序和分组后的打印内容"""
     try:
         # 读取 Excel，跳过前 8 行
         df = pd.read_excel(file, skiprows=8)
-        
+
         # 提取所需列 (G9, H9, J9)
         df = df.iloc[:, [6, 7, 9]]  # G, H, J 列
         df.columns = ['Hall', 'StartTime', 'EndTime']
-        
+
         # 清理数据
         df = df.dropna(subset=['Hall', 'StartTime', 'EndTime'])
-        
+
         # 转换影厅格式为 "#号" 格式
         df['Hall'] = df['Hall'].str.extract(r'(\d+)号').astype(str) + ' '
-        
+
         # 保存原始时间字符串用于诊断
         df['original_end'] = df['EndTime']
-        
+
         # 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'])
-        
+        # Using errors='coerce' will turn unparseable times into NaT (Not a Time)
+        df['StartTime'] = pd.to_datetime(df['StartTime'], errors='coerce')
+        df['EndTime'] = pd.to_datetime(df['EndTime'], errors='coerce')
+        df = df.dropna(subset=['StartTime', 'EndTime']) # Drop rows where time conversion failed
+
         # 设置基准时间
-        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")
-        
-        # 处理跨天情况
-        if business_end < business_start:
-            business_end += timedelta(days=1)
-        
-        # 标准化所有时间到同一天
-        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)
-        
-        # 筛选营业时间内的场次
-        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())))
+        business_start_time = datetime.strptime(BUSINESS_START, "%H:%M").time()
+        business_end_time = datetime.strptime(BUSINESS_END, "%H:%M").time()
+
+        # 处理跨天情况：结束时间小于开始时间，则结束时间加一天
+        # This logic handles cases like 9:30 AM to 1:30 AM (next day)
+        df['EndTime_adjusted'] = df.apply(
+            lambda row: row['EndTime'] + timedelta(days=1) if row['EndTime'].time() < row['StartTime'].time() else row['EndTime'],
+            axis=1
         )
         
-        df = df[valid_times]
-        
-        # 按散场时间排序
-        df = df.sort_values('EndTime')
-        
+        # 按散场时间排序 (using the adjusted time)
+        df = df.sort_values('EndTime_adjusted')
+
         # 分割数据
-        split_time = datetime.strptime(f"{base_date} {SPLIT_TIME}", "%Y-%m-%d %H:%M")
-        split_time_for_comparison = df['time_for_comparison'].apply(
-            lambda x: datetime.combine(base_date, split_time.time())
-        )
-        
-        part1 = df[df['time_for_comparison'] <= split_time_for_comparison].copy()
-        part2 = df[df['time_for_comparison'] > split_time_for_comparison].copy()
+        split_dt = datetime.strptime(SPLIT_TIME, "%H:%M").time()
         
-        # 格式化时间显示
+        part1 = df[df['EndTime_adjusted'].dt.time <= split_dt].copy()
+        part2 = df[df['EndTime_adjusted'].dt.time > split_dt].copy()
+
+        # 格式化时间显示 (use original EndTime for display)
         for part in [part1, part2]:
-            part['EndTime'] = part['EndTime'].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
-        )
+            part['EndTime_formatted'] = part['EndTime'].dt.strftime('%-I:%M')
+
+        # 读取日期单元格 C6
+        date_df = pd.read_excel(file, skiprows=5, nrows=1, usecols=[2], header=None)
         date_cell = date_df.iloc[0, 0]
-        
+
         try:
-            # 处理不同日期格式
             if isinstance(date_cell, str):
+                # Assuming format like '2023-10-27'
                 date_str = datetime.strptime(date_cell, '%Y-%m-%d').strftime('%Y-%m-%d')
             else:
+                # Assuming it's a datetime object
                 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
-    
+
+        return part1[['Hall', 'EndTime_formatted']], part2[['Hall', 'EndTime_formatted']], date_str
+
     except Exception as e:
         st.error(f"处理文件时出错: {str(e)}")
         return None, None, None
 
+
 def create_print_layout(data, title, date_str):
-    """创建打印布局 (PNG 和 PDF)"""
+    """
+    创建符合新要求的打印布局 (PNG 和 PDF)。
+    1. 动态计算边距。
+    2. 使用灰色虚线圆点作为单元格边框。
+    3. 单元格内容区域为单元格的90%。
+    4. 在左上角添加灰色序号。
+    """
     if data.empty:
         return None
 
-    # --- 创建 PNG 图形 ---
-    png_fig = plt.figure(figsize=(5.83, 8.27), dpi=300) # A5 竖向
-    png_ax_container = png_fig.add_subplot(111) # 创建一个容器轴，用于隐藏外部边框
-    png_ax_container.set_axis_off()
-    # 减小边距，例如从 0.05 减小到 0.02
-    png_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)
-
-    # --- 创建 PDF 图形 ---
-    pdf_fig = plt.figure(figsize=(5.83, 8.27), dpi=300) # A5 竖向
-    pdf_ax_container = pdf_fig.add_subplot(111)
-    pdf_ax_container.set_axis_off()
-    # 减小边距，例如从 0.05 减小到 0.02
-    pdf_fig.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)
-
-    # --- 内部绘图函数 ---
-    def process_figure(fig, is_pdf=False):
-        # 设置字体
-        plt.rcParams['font.family'] = 'sans-serif'
-        plt.rcParams['font.sans-serif'] = ['Arial Unicode MS'] # 确保字体可用
-
-        # 计算行数和总数
-        total_items = len(data)
-        num_cols = 3
-        num_rows = math.ceil(total_items / num_cols)
-
-        # 创建网格 (在 figure 内部创建)
-        # 减小子图间距 hspace/wspace，减小日期行高度比例 height_ratios
-        gs = gridspec.GridSpec(num_rows + 1, num_cols, hspace=0.05, wspace=0.05, height_ratios=[0.1] + [1] * num_rows, figure=fig) # 将日期行放在顶部
-
-        # 调整基础字体大小，避免过大或过小
-        # A5 宽度大约 1749 像素 @ 300dpi, 高度 2481
-        # 每列宽度约 1749 * 0.9 / 3 = 525 像素
-        # 每行高度约 (2481 * 0.9 * (1 / (1.2))) / num_rows
-        # 字体大小与单元格大小相关，这里用经验值调整
-        available_height_per_row = (8.27 * 0.9 * (1 / 1.2)) / num_rows if num_rows > 0 else 1
-        base_fontsize = min(40, max(10, available_height_per_row * 72 * 0.5)) # 72 points per inch, 估算系数
-
-        data_values = data.values.tolist()
-
-        # 补全空位，确保是3的倍数
-        while len(data_values) % num_cols != 0:
-            data_values.append(['', ''])
-
-        rows_per_col_layout = math.ceil(len(data_values) / num_cols) # 按列优先排列的行数
-
-        # 按列优先排序数据 (Z字形)
-        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
-
-        # 绘制数据单元格
-        for idx, (hall, end_time) in enumerate(sorted_data):
-            if hall and end_time:
-                row_grid = idx // num_cols + 1 # +1 因为日期占了第0行
-                col_grid = idx % num_cols
-
-                if row_grid < num_rows + 1: # 确保索引在网格内
-                    ax = fig.add_subplot(gs[row_grid, col_grid]) # 使用 fig.add_subplot
-
-                    # --- 修改开始：绘制圆角矩形 ---
-                    # 隐藏原始边框
-                    for spine in ax.spines.values():
-                        spine.set_visible(False)
-
-                    # 创建圆角矩形 Patch
-                    bbox = FancyBboxPatch(
-                        (0.01, 0.01),  # 左下角坐标 (稍微内缩一点避免接触边缘)
-                        0.98, 0.98,  # 宽度和高度 (占满大部分区域)
-                        boxstyle="round,pad=0,rounding_size=0.02", # 圆角样式，rounding_size 控制圆角程度
-                        edgecolor=BORDER_COLOR,
-                        facecolor='none', # 无填充色
-                        linewidth=0.5,
-                        transform=ax.transAxes, # 使用相对坐标
-                        clip_on=False # 避免被裁剪
-                    )
-                    # 添加 Patch 到 Axes
-                    ax.add_patch(bbox)
-                    # --- 修改结束 ---
-
-                    display_text = f"{hall}{end_time}"
-                    ax.text(0.5, 0.5, display_text,
-                            fontsize=base_fontsize,
-                            fontweight='bold',
-                            ha='center',
-                            va='center',
-                            transform=ax.transAxes) # 使用相对坐标
-
-                    ax.set_xticks([])
-                    ax.set_yticks([])
-                else:
-                     print(f"Warning: Index out of bounds - idx={idx}, row_grid={row_grid}, col_grid={col_grid}")
-
-
-        # 添加日期信息到第一个子图的顶部
-        ax_date = fig.add_subplot(gs[0, :]) # 跨越第一行的所有列
-        ax_date.text(0.01, 0.5, f"{date_str} {title}", # 调整位置和对齐
-                     fontsize=base_fontsize * 0.5, # 调整日期字体大小
-                     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') # 使背景透明
-
-
-    # --- 处理图形 ---
-    process_figure(png_fig)
-    process_figure(pdf_fig, is_pdf=True)
-
-    # --- 保存 PNG ---
+    # --- Constants ---
+    A5_WIDTH_IN = 5.83
+    A5_HEIGHT_IN = 8.27
+    DPI = 300
+    NUM_COLS = 3
+
+    # --- Setup Figure ---
+    fig = plt.figure(figsize=(A5_WIDTH_IN, A5_HEIGHT_IN), dpi=DPI)
+    
+    # --- Font Setup ---
+    plt.rcParams['font.family'] = 'sans-serif'
+    plt.rcParams['font.sans-serif'] = ['Arial Unicode MS', 'Heiti TC', 'sans-serif']
+
+
+    # --- Data Preparation ---
+    total_items = len(data)
+    # Augment data with an original index for numbering
+    data_values_with_index = [(i, row) for i, row in enumerate(data.values.tolist())]
+    
+    # Pad data to be a multiple of NUM_COLS
+    padded_total = math.ceil(total_items / NUM_COLS) * NUM_COLS
+    while len(data_values_with_index) < padded_total:
+        data_values_with_index.append((None, ['', '']))
+
+    num_rows = padded_total // NUM_COLS
+
+    # --- Layout Calculation (Request 1) ---
+    if num_rows > 0:
+        # "A5 paper height / num_rows / 4 is the padding for all sides"
+        padding_in = (A5_HEIGHT_IN / num_rows / 4)
+        # Cap padding to prevent it from being excessively large
+        padding_in = min(padding_in, 0.5) 
+    else:
+        padding_in = 0.25 # Default padding if no rows
+
+    # Convert padding to relative figure coordinates for subplots_adjust
+    left_margin = padding_in / A5_WIDTH_IN
+    right_margin = 1 - left_margin
+    bottom_margin = padding_in / A5_HEIGHT_IN
+    top_margin = 1 - bottom_margin
+    
+    # Adjust overall figure margins
+    fig.subplots_adjust(left=left_margin, right=right_margin, top=top_margin, bottom=bottom_margin, hspace=0.4, wspace=0.4)
+
+    # --- Grid & Font Size ---
+    gs = gridspec.GridSpec(num_rows + 1, NUM_COLS, height_ratios=[0.2] + [1] * num_rows, figure=fig)
+
+    if num_rows > 0:
+        content_area_height_in = A5_HEIGHT_IN * (top_margin - bottom_margin)
+        cell_height_in = content_area_height_in / num_rows * (1 - fig.subplotpars.hspace)
+        base_fontsize = min(40, max(10, cell_height_in * 72 * 0.4)) # 72 pt/inch, 40% of cell height
+    else:
+        base_fontsize = 20
+
+    # --- Z-Sort (Column-major) Data for Layout ---
+    rows_per_col_layout = num_rows
+    sorted_data = [(None, ['',''])] * padded_total
+    for i, item_tuple in enumerate(data_values_with_index):
+        if item_tuple[0] is not None:
+            original_data_index = i # Index from the time-sorted list
+            row_in_col = original_data_index % rows_per_col_layout
+            col_idx = original_data_index // rows_per_col_layout
+            new_grid_index = row_in_col * NUM_COLS + col_idx
+            if new_grid_index < len(sorted_data):
+                sorted_data[new_grid_index] = item_tuple
+
+    # --- Drawing Logic ---
+    for grid_idx, item_tuple in enumerate(sorted_data):
+        original_index, (hall, end_time) = item_tuple
+        
+        if original_index is not None:
+            row_grid = grid_idx // NUM_COLS + 1 # +1 because date is in row 0
+            col_grid = grid_idx % NUM_COLS
+            
+            ax = fig.add_subplot(gs[row_grid, col_grid])
+            ax.set_axis_off()
+
+            # --- Cell Border (Request 2) & Content Area (Request 3) ---
+            # Draw a dotted rectangle. Content will be placed inside this.
+            # Making the rect slightly smaller creates a visual 90% area.
+            cell_border = Rectangle((0.05, 0.05), 0.9, 0.9,
+                                    edgecolor=BORDER_COLOR,
+                                    facecolor='none',
+                                    linestyle=(0, (1, 1.5)), # Dotted line with round caps
+                                    linewidth=1,
+                                    transform=ax.transAxes,
+                                    clip_on=False)
+            ax.add_patch(cell_border)
+
+            # --- Cell Content ---
+            display_text = f"{hall}{end_time}"
+            ax.text(0.5, 0.5, display_text,
+                    fontsize=base_fontsize,
+                    fontweight='bold',
+                    ha='center', va='center',
+                    transform=ax.transAxes)
+
+            # --- Cell Numbering (Request 4) ---
+            # Serial number is original_index + 1
+            ax.text(0.12, 0.82, str(original_index + 1),
+                    fontsize=base_fontsize * 0.5,
+                    color=SEQ_COLOR,
+                    fontweight='normal',
+                    ha='center', va='center',
+                    transform=ax.transAxes)
+
+    # --- Date Header ---
+    ax_date = fig.add_subplot(gs[0, :])
+    ax_date.set_axis_off()
+    ax_date.text(0, 0.5, f"{date_str} {title}",
+                 fontsize=base_fontsize * 0.6,
+                 color=DATE_COLOR,
+                 fontweight='bold',
+                 ha='left', va='center',
+                 transform=ax_date.transAxes)
+
+    # --- Save to Buffers ---
+    # Save PNG
     png_buffer = io.BytesIO()
-    # 可以尝试减小 pad_inches， even set to 0
-    png_fig.savefig(png_buffer, format='png', bbox_inches='tight', pad_inches=0.02)
+    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)
 
-    # --- 保存 PDF ---
+    # Save PDF
     pdf_buffer = io.BytesIO()
-    with PdfPages(pdf_buffer) as pdf:
-        # 可以尝试减小 pad_inches， even set to 0
-        pdf.savefig(pdf_fig, bbox_inches='tight', pad_inches=0.02)
+    fig.savefig(pdf_buffer, format='pdf', bbox_inches='tight', pad_inches=0.02)
     pdf_buffer.seek(0)
     pdf_base64 = base64.b64encode(pdf_buffer.getvalue()).decode()
-    plt.close(pdf_fig)
+    
+    plt.close(fig)
 
     return {
         'png': f'data:image/png;base64,{png_base64}',
         'pdf': f'data:application/pdf;base64,{pdf_base64}'
     }
 
-# --- 新增 PDF 显示函数 ---
 def display_pdf(base64_pdf):
     """在Streamlit中嵌入显示PDF"""
     pdf_display = f'<iframe src="{base64_pdf}" width="100%" height="800" type="application/pdf"></iframe>'
     return pdf_display
 
-# Streamlit 界面
+# --- 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"])
+uploaded_file = st.file_uploader("上传【放映场次核对表.xls】文件", type=["xls", "xlsx"])
 
 if uploaded_file:
+    # Use new column name 'EndTime_formatted' for display
     part1, part2, date_str = process_schedule(uploaded_file)
-
     if part1 is not None and part2 is not None:
-        # 生成包含 PNG 和 PDF 的字典
-        part1_output = create_print_layout(part1, "A", date_str)
-        part2_output = create_print_layout(part2, "C", date_str)
+        part1_data_for_layout = part1[['Hall', 'EndTime_formatted']]
+        part2_data_for_layout = part2[['Hall', 'EndTime_formatted']]
+        
+        part1_output = create_print_layout(part1_data_for_layout, "A", date_str)
+        part2_output = create_print_layout(part2_data_for_layout, "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（散场时间 ≤ 17:30）")
             if part1_output:
                 tab1_1, tab1_2 = st.tabs(["PDF 预览", "PNG 预览"])
                 with tab1_1:
@@ -291,7 +270,7 @@ 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（散场时间 > 17:30）")
             if part2_output:
                 tab2_1, tab2_2 = st.tabs(["PDF 预览", "PNG 预览"])
                 with tab2_1:
@@ -299,6 +278,4 @@ if uploaded_file:
                 with tab2_2:
                     st.image(part2_output['png'])
             else:
-                st.info("夜班部分没有数据")
-
-                
+                st.info("夜班部分没有数据")