Update app.py

app.py

@@ -5,28 +5,50 @@ 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
+import matplotlib.font_manager as fm # 新增导入
+import os # 新增导入
 
-# --- CONSTANTS ---
+# --- 全局常量 ---
 SPLIT_TIME = "17:30"
 BUSINESS_START = "09:30"
 BUSINESS_END = "01:30"
-BORDER_COLOR = '#A9A9A9'
+BORDER_COLOR = 'grey'
 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')
+FONT_PATH = 'SimHei.ttf' # 字体文件路径
+
+# --- 字体设置 ---
+FONT_BOLD = None
+FONT_REGULAR = None
+FONT_WARNING_ISSUED = False
+
+def get_font_properties():
+    """加载字体文件并返回 FontProperties 对象"""
+    global FONT_BOLD, FONT_REGULAR, FONT_WARNING_ISSUED
+    if FONT_BOLD is None: # 只加载一次
+        if os.path.exists(FONT_PATH):
+            try:
+                FONT_BOLD = fm.FontProperties(fname=FONT_PATH, weight='bold')
+                FONT_REGULAR = fm.FontProperties(fname=FONT_PATH, weight='normal')
+            except Exception as e:
+                if not FONT_WARNING_ISSUED:
+                    st.warning(f"加载字体 '{FONT_PATH}' 失败: {e}. 将使用默认字体。")
+                    FONT_WARNING_ISSUED = True
+        else:
+            if not FONT_WARNING_ISSUED:
+                st.warning(f"字体文件 '{FONT_PATH}' 未找到。请将其放置在代码同目录下。将使用默认字体。")
+                FONT_WARNING_ISSUED = True
+
+def get_font_args(style='regular'):
+    """根据请求的样式返回字体参数字典"""
+    get_font_properties() # 确保字体已加载
+    if style == 'bold' and FONT_BOLD:
+        return {'fontproperties': FONT_BOLD}
+    if style == 'regular' and FONT_REGULAR:
+        return {'fontproperties': FONT_REGULAR}
+    # Fallback
+    return {'fontfamily': 'sans-serif', 'weight': 'bold' if style == 'bold' else 'normal'}
 
 
 def process_schedule(file):
@@ -34,58 +56,82 @@ def process_schedule(file):
     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'])
-        df = df[df['Hall'].astype(str).str.contains(r'\d', na=False)] # Ensure Hall has a digit
-
+        
         # 转换影厅格式为 "#号" 格式
-        df['Hall'] = df['Hall'].astype(str).str.extract(r'(\d+)').astype(str) + ' '
-
+        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'], errors='coerce').dt.time
-        df['EndTime'] = pd.to_datetime(df['EndTime'], errors='coerce').dt.time
-
-        df = df.dropna(subset=['StartTime', '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
-
-        df['StartDateTime'] = df['StartTime'].apply(combine_date_time)
-        df['EndDateTime'] = df['EndTime'].apply(combine_date_time)
-
+        df['StartTime'] = pd.to_datetime(df['StartTime'])
+        df['EndTime'] = pd.to_datetime(df['EndTime'])
+        
+        # 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())))
+        )
+        
+        df = df[valid_times]
+        
         # 按散场时间排序
-        df = df.sort_values('EndDateTime')
-
+        df = df.sort_values('EndTime')
+        
         # 分割数据
-        split_dt = datetime.combine(base_date, datetime.strptime(SPLIT_TIME, "%H:%M").time())
+        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()
         
-        part1 = df[df['EndDateTime'] <= split_dt].copy()
-        part2 = df[df['EndDateTime'] > split_dt].copy()
-
         # 格式化时间显示
         for part in [part1, part2]:
-            part['EndTimeStr'] = part['EndDateTime'].dt.strftime('%-I:%M')
-
+            part['EndTime'] = part['EndTime'].dt.strftime('%-H:%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
+            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):
@@ -94,133 +140,142 @@ def process_schedule(file):
                 date_str = pd.to_datetime(date_cell).strftime('%Y-%m-%d')
         except:
             date_str = datetime.today().strftime('%Y-%m-%d')
-
-        return part1[['Hall', 'EndTimeStr']], part2[['Hall', 'EndTimeStr']], date_str
-
+        
+        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
 
-
 def create_print_layout(data, title, date_str):
-    """创建符合新要求的打印布局 (PNG 和 PDF)"""
+    """创建基于精确表格布局的打印页面 (PNG 和 PDF)"""
     if data.empty:
         return None
 
-    # --- 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']
+    # --- 1. 布局和尺寸计算 ---
+    A5_WIDTH_IN, A5_HEIGHT_IN = 5.83, 8.27
+    DPI = 300
+    MARGIN = 0.25 # 页边距 (英寸)
+    HEADER_HEIGHT_IN = 0.3 # 日期标题栏高度 (英寸)
 
-    # 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
+    content_width = A5_WIDTH_IN - (2 * MARGIN)
+    content_height = A5_HEIGHT_IN - (2 * MARGIN)
     
-    # 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)
+    total_items = len(data)
+    num_cols = 3
+    num_rows = math.ceil(total_items / num_cols)
     
-    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)
+    grid_height = content_height - HEADER_HEIGHT_IN
+    cell_width = content_width / num_cols
+    cell_height = grid_height / num_rows if num_rows > 0 else 0
+
+    # --- 2. 动态字体大小计算 ---
+    def calculate_font_size(fig):
+        # 使用一个典型的宽字符串来估算
+        sample_text = "8 88:88" 
+        target_width = cell_width * 0.9  # 目标宽度为单元格宽度的90%
         
-        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(['', ''])
+        # 用初始字号10来测量
+        initial_fontsize = 10
+        t = fig.text(0.5, 0.5, sample_text, fontsize=initial_fontsize, **get_font_args('bold'))
+        # 渲染并获取边界框（以英寸为单位）
+        bbox = t.get_window_extent(renderer=fig.canvas.get_renderer())
+        text_width_in = bbox.width / fig.get_dpi()
+        t.remove() # 移除临时文本
         
-        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()
-
+        # 根据宽度比例调整字体大小
+        if text_width_in > 0:
+            scale_factor = target_width / text_width_in
+            return initial_fontsize * scale_factor
+        return initial_fontsize # 备用
+
+    # --- 3. 数据准备 (Z字形排序) ---
+    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
+
+    # --- 4. 绘图 ---
+    fig = plt.figure(figsize=(A5_WIDTH_IN, A5_HEIGHT_IN), dpi=DPI)
+    final_fontsize = calculate_font_size(fig)
+    index_fontsize = final_fontsize * 0.35
+    date_fontsize = final_fontsize * 0.5
+    
+    # 绘制日期头
+    header_ax = fig.add_axes([MARGIN / A5_WIDTH_IN, (A5_HEIGHT_IN - MARGIN - HEADER_HEIGHT_IN) / A5_HEIGHT_IN, content_width / A5_WIDTH_IN, HEADER_HEIGHT_IN / A5_HEIGHT_IN])
+    header_ax.text(0.01, 0.5, f"{date_str} {title}",
+                     fontsize=date_fontsize,
+                     color=DATE_COLOR,
+                     ha='left',
+                     va='center',
+                     transform=header_ax.transAxes,
+                     **get_font_args('bold'))
+    header_ax.set_axis_off()
+    
+    # 绘制每个单元格
+    for idx, (hall, end_time) in enumerate(sorted_data):
+        if hall and end_time:
+            row_grid = idx // num_cols
+            col_grid = idx % num_cols
+
+            # 计算单元格的绝对位置（从左下角开始，单位：英寸）
+            x_pos_in = MARGIN + (col_grid * cell_width)
+            y_pos_in = MARGIN + ( (num_rows - 1 - row_grid) * cell_height )
+            
+            # 转换为 Figure 的相对坐标 [left, bottom, width, height]
+            ax_rect = [x_pos_in / A5_WIDTH_IN, y_pos_in / A5_HEIGHT_IN, cell_width / A5_WIDTH_IN, cell_height / A5_HEIGHT_IN]
+            ax = fig.add_axes(ax_rect)
+
+            # 绘制主要文本
+            display_text = f"{hall}{end_time}"
+            ax.text(0.5, 0.5, display_text,
+                      fontsize=final_fontsize,
+                      ha='center',
+                      va='center',
+                      transform=ax.transAxes,
+                      **get_font_args('bold'))
+            
+            # 绘制左上角序号
+            ax.text(0.08, 0.92, str(idx + 1),
+                      fontsize=index_fontsize,
+                      color=BORDER_COLOR,
+                      ha='left',
+                      va='top',
+                      transform=ax.transAxes,
+                      **get_font_args('regular'))
+
+            # 设置边框为灰色点状虚线
+            for spine in ax.spines.values():
+                spine.set_color(BORDER_COLOR)
+                spine.set_linestyle(':')
+                spine.set_linewidth(1)
+            
+            ax.set_xticks([])
+            ax.set_yticks([])
+
+    # --- 5. 保存到内存 ---
     # 保存 PNG
     png_buffer = io.BytesIO()
-    fig_for_output.savefig(png_buffer, format='png', dpi=DPI)
+    fig.savefig(png_buffer, format='png', pad_inches=0)
     png_buffer.seek(0)
     png_base64 = base64.b64encode(png_buffer.getvalue()).decode()
     
     # 保存 PDF
     pdf_buffer = io.BytesIO()
-    fig_for_output.savefig(pdf_buffer, format='pdf', dpi=DPI)
+    fig.savefig(pdf_buffer, format='pdf', pad_inches=0)
     pdf_buffer.seek(0)
     pdf_base64 = base64.b64encode(pdf_buffer.getvalue()).decode()
-
-    plt.close(fig_for_output)
+    
+    plt.close(fig)
 
     return {
         'png': f'data:image/png;base64,{png_base64}',
@@ -229,28 +284,30 @@ def create_print_layout(data, title, date_str):
 
 def display_pdf(base64_pdf):
     """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>'
+    pdf_display = f'<iframe src="{base64_pdf}" width="100%" height="800" type="application/pdf"></iframe>'
     return pdf_display
 
-# --- Streamlit UI ---
+# --- Streamlit 界面 ---
 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"])
+# 检查并加载字体，如有需要会显示警告
+get_font_properties()
+
+uploaded_file = st.file_uploader("上传【放映场次核对表.xls】文件", type=["xls"])
 
 if uploaded_file:
-    # Rename columns in process_schedule call to match new names
-    part1_data, part2_data, date_str = process_schedule(uploaded_file)
+    part1, part2, date_str = process_schedule(uploaded_file)
 
-    if part1_data is not None and part2_data is not None:
-        # 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)
+    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)
 
         col1, col2 = st.columns(2)
 
         with col1:
-            st.subheader("白班散场预览（结束时间 ≤ 17:30）")
+            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"])
                 with tab1_1:
@@ -261,7 +318,7 @@ if uploaded_file:
                 st.info("白班部分没有数据")
 
         with col2:
-            st.subheader("夜班散场预览（结束时间 > 17:30）")
+            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"])
                 with tab2_1: