#!/usr/bin/env python3 """ 数据可视化图表生成脚本 - AI智能分析版本 使用Plotly生成精美、专业的数据可视化仪表板 支持多种图表类型，根据数据特征智能选择最佳可视化方案 """ import argparse import os import sys import pandas as pd import numpy as np import plotly.graph_objects as go from plotly.subplots import make_subplots from datetime import datetime def read_data(input_file): """ 读取输入数据文件，支持CSV和Excel格式 Args: input_file: 输入文件路径 Returns: pandas DataFrame """ if not os.path.exists(input_file): raise FileNotFoundError(f"输入文件不存在: {input_file}") file_ext = os.path.splitext(input_file)[1].lower() if file_ext == '.csv': df = pd.read_csv(input_file) elif file_ext in ['.xlsx', '.xls']: df = pd.read_excel(input_file) else: raise ValueError(f"不支持的文件格式: {file_ext}") return df def analyze_data_characteristics(df): """ 分析数据特征，为图表选择提供依据 Args: df: pandas DataFrame Returns: dict: 数据特征字典 """ numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist() categorical_cols = df.select_dtypes(include=['object', 'category']).columns.tolist() datetime_cols = df.select_dtypes(include=['datetime']).columns.tolist() # 尝试识别日期列 for col in df.columns: if col not in datetime_cols: try: pd.to_datetime(df[col]) if col not in categorical_cols: datetime_cols.append(col) except: pass characteristics = { 'total_rows': len(df), 'total_cols': len(df.columns), 'numeric_cols': numeric_cols, 'numeric_count': len(numeric_cols), 'categorical_cols': categorical_cols, 'categorical_count': len(categorical_cols), 'datetime_cols': datetime_cols, 'datetime_count': len(datetime_cols), 'numeric_stats': {}, 'categorical_stats': {}, 'correlations': None } # 数值列统计 for col in numeric_cols: characteristics['numeric_stats'][col] = { 'mean': df[col].mean(), 'median': df[col].median(), 'std': df[col].std(), 'min': df[col].min(), 'max': df[col].max(), 'skewness': df[col].skew(), 'unique_count': df[col].nunique(), 'missing_rate': df[col].isna().sum() / len(df) } # 分类列统计 for col in categorical_cols: characteristics['categorical_stats'][col] = { 'unique_count': df[col].nunique(), 'mode': df[col].mode().iloc[0] if len(df[col].mode()) > 0 else None, 'missing_rate': df[col].isna().sum() / len(df) } # 相关性分析 if len(numeric_cols) >= 2: try: correlations = df[numeric_cols].corr() characteristics['correlations'] = correlations except: pass return characteristics def recommend_charts(characteristics): """ 根据数据特征智能推荐图表类型 Args: characteristics: 数据特征字典 Returns: list: 推荐的图表列表 """ recommended = [] numeric_cols = characteristics['numeric_cols'] categorical_cols = characteristics['categorical_cols'] datetime_cols = characteristics['datetime_cols'] numeric_count = characteristics['numeric_count'] categorical_count = characteristics['categorical_count'] # 1. 数据概览表格（总是包含） recommended.append({ 'type': 'table', 'title': '数据概览', 'priority': 10 }) # 2. 数值列分布分析 if numeric_count > 0: for col in numeric_cols: stats = characteristics['numeric_stats'].get(col, {}) unique_count = stats.get('unique_count', 0) # 根据唯一值数量选择分布图类型 if unique_count <= 10: # 少量唯一值：使用柱状图 recommended.append({ 'type': 'value_counts_bar', 'col': col, 'title': f'{col}分布', 'priority': 8 }) else: # 大量唯一值：使用直方图 recommended.append({ 'type': 'histogram', 'col': col, 'title': f'{col}分布', 'priority': 8 }) # 如果数据量大且有偏态，添加箱线图 if characteristics['total_rows'] >= 10: recommended.append({ 'type': 'box_single', 'col': col, 'title': f'{col}箱线图', 'priority': 6 }) # 3. 分类数据分析 if categorical_count > 0 and numeric_count > 0: for cat_col in categorical_cols[:2]: # 最多取前2个分类列 cat_stats = characteristics['categorical_stats'].get(cat_col, {}) unique_count = cat_stats.get('unique_count', 0) # 只处理合理数量的分类 if 2 <= unique_count <= 15: # 柱状图对比 for num_col in numeric_cols[:2]: # 最多取前2个数值列 recommended.append({ 'type': 'grouped_bar', 'cat_col': cat_col, 'num_col': num_col, 'title': f'{num_col}按{cat_col}', 'priority': 9 }) # 饼图（分类数量适中） if unique_count <= 8: recommended.append({ 'type': 'pie', 'cat_col': cat_col, 'num_col': num_col, 'title': f'{num_col}占比', 'priority': 7 }) # 箱线图（展示分布差异） if characteristics['total_rows'] >= 10: recommended.append({ 'type': 'box_grouped', 'cat_col': cat_col, 'num_col': num_col, 'title': f'{num_col}分布差异', 'priority': 7 }) # 小提琴图（更详细的分布展示） if unique_count <= 6 and characteristics['total_rows'] >= 20: recommended.append({ 'type': 'violin', 'cat_col': cat_col, 'num_col': num_col, 'title': f'{num_col}分布密度', 'priority': 5 }) # 4. 时间序列分析 if len(datetime_cols) > 0 and numeric_count > 0: for dt_col in datetime_cols[:1]: for num_col in numeric_cols[:2]: recommended.append({ 'type': 'line', 'x_col': dt_col, 'y_col': num_col, 'title': f'{num_col}趋势', 'priority': 9 }) # 面积图 recommended.append({ 'type': 'area', 'x_col': dt_col, 'y_col': num_col, 'title': f'{num_col}面积图', 'priority': 6 }) # 5. 相关性分析 if numeric_count >= 2: # 散点图矩阵（取前几个重要数值列） main_numeric_cols = numeric_cols[:min(4, numeric_count)] for i in range(len(main_numeric_cols)): for j in range(i+1, len(main_numeric_cols)): x_col = main_numeric_cols[i] y_col = main_numeric_cols[j] # 检查相关性强度 if characteristics['correlations'] is not None: corr_value = abs(characteristics['correlations'].loc[x_col, y_col]) priority = 8 if corr_value > 0.3 else 5 else: priority = 6 recommended.append({ 'type': 'scatter', 'x_col': x_col, 'y_col': y_col, 'title': f'{x_col} vs {y_col}', 'priority': priority }) # 相关性热力图 if numeric_count >= 3: recommended.append({ 'type': 'heatmap', 'cols': numeric_cols, 'title': '相关性热力图', 'priority': 7 }) # 6. 多变量分析 if numeric_count >= 3: # 平行坐标图 recommended.append({ 'type': 'parallel_coordinates', 'cols': numeric_cols[:5], 'title': '多维度分析', 'priority': 5 }) # 7. 分类列之间的关系 if categorical_count >= 2: for i in range(min(2, categorical_count)): for j in range(i+1, min(3, categorical_count)): recommended.append({ 'type': 'sunburst', 'cat_cols': [categorical_cols[i], categorical_cols[j]], 'num_col': numeric_cols[0] if numeric_count > 0 else None, 'title': '层级分析', 'priority': 4 }) # 按优先级排序 recommended.sort(key=lambda x: x['priority'], reverse=True) # 限制图表数量（最多15个） return recommended[:15] def create_metric_card(title, value, change=None, icon="📊", color="#667eea"): """ 创建指标卡片 Args: title: 标题 value: 数值 change: 变化值 icon: 图标 color: 颜色 Returns: HTML字符串 """ change_html = "" if change is not None: change_color = "#10b981" if change >= 0 else "#ef4444" change_icon = "↑" if change >= 0 else "↓" change_html = f'
{change_icon} {abs(change):.1f}% vs 上期
' return f"""
{icon}
{title}
{value}
{change_html}
""" def generate_unified_dashboard(df, output_file="visualization_dashboard.html"): """ 生成精美、专业的数据可视化仪表板 Args: df: pandas DataFrame output_file: 输出文件路径 """ # 分析数据特征 characteristics = analyze_data_characteristics(df) # 智能推荐图表 recommended_charts = recommend_charts(characteristics) numeric_cols = characteristics['numeric_cols'] categorical_cols = characteristics['categorical_cols'] # 专业配色方案 color_list = [ '#6366F1', '#8B5CF6', '#EC4899', '#F59E0B', '#10B981', '#3B82F6', '#14B8A6', '#EF4444', '#F97316', '#84CC16', '#06B6D4', '#A855F7' ] # 创建指标卡片HTML metrics_html = "" metrics_html += create_metric_card("总数据量", f"{characteristics['total_rows']:,}", icon="📈", color="#6366F1") if characteristics['numeric_count'] > 0: avg_value = df[numeric_cols[0]].mean() metrics_html += create_metric_card(f"{numeric_cols[0]}平均值", f"{avg_value:,.0f}", icon="📊", color="#8B5CF6") if characteristics['categorical_count'] > 0 and characteristics['numeric_count'] > 0: max_cat = df[categorical_cols[0]].value_counts().idxmax() metrics_html += create_metric_card(f"最常见{categorical_cols[0]}", f"{max_cat}", icon="🏷️", color="#10B981") if characteristics['numeric_count'] > 0: total_sum = df[numeric_cols[0]].sum() metrics_html += create_metric_card(f"{numeric_cols[0]}总计", f"{total_sum:,.0f}", icon="💰", color="#F59E0B") # 构建图表列表（根据推荐生成） chart_list = [] for chart_info in recommended_charts: chart_type = chart_info['type'] if chart_type == 'table': chart_list.append({ 'title': f'📋 {chart_info["title"]}', 'type': 'table', 'full_width': True }) elif chart_type == 'histogram': chart_list.append({ 'title': f'📊 {chart_info["title"]}', 'type': 'histogram', 'col': chart_info['col'] }) elif chart_type == 'value_counts_bar': chart_list.append({ 'title': f'📊 {chart_info["title"]}', 'type': 'value_counts_bar', 'col': chart_info['col'] }) elif chart_type == 'box_single': chart_list.append({ 'title': f'📦 {chart_info["title"]}', 'type': 'box_single', 'col': chart_info['col'] }) elif chart_type == 'grouped_bar': chart_list.append({ 'title': f'📈 {chart_info["title"]}', 'type': 'grouped_bar', 'cat_col': chart_info['cat_col'], 'num_col': chart_info['num_col'] }) elif chart_type == 'pie': chart_list.append({ 'title': f'🥧 {chart_info["title"]}', 'type': 'pie', 'cat_col': chart_info['cat_col'], 'num_col': chart_info['num_col'] }) elif chart_type == 'box_grouped': chart_list.append({ 'title': f'📦 {chart_info["title"]}', 'type': 'box_grouped', 'cat_col': chart_info['cat_col'], 'num_col': chart_info['num_col'] }) elif chart_type == 'violin': chart_list.append({ 'title': f'🎻 {chart_info["title"]}', 'type': 'violin', 'cat_col': chart_info['cat_col'], 'num_col': chart_info['num_col'] }) elif chart_type == 'line': chart_list.append({ 'title': f'📉 {chart_info["title"]}', 'type': 'line', 'x_col': chart_info['x_col'], 'y_col': chart_info['y_col'] }) elif chart_type == 'area': chart_list.append({ 'title': f'🏔️ {chart_info["title"]}', 'type': 'area', 'x_col': chart_info['x_col'], 'y_col': chart_info['y_col'] }) elif chart_type == 'scatter': chart_list.append({ 'title': f'🔵 {chart_info["title"]}', 'type': 'scatter', 'x_col': chart_info['x_col'], 'y_col': chart_info['y_col'] }) elif chart_type == 'heatmap': chart_list.append({ 'title': f'🌡️ {chart_info["title"]}', 'type': 'heatmap', 'cols': chart_info['cols'], 'full_width': True }) elif chart_type == 'parallel_coordinates': chart_list.append({ 'title': f'🧭 {chart_info["title"]}', 'type': 'parallel_coordinates', 'cols': chart_info['cols'], 'full_width': True }) elif chart_type == 'sunburst': chart_list.append({ 'title': f'☀️ {chart_info["title"]}', 'type': 'sunburst', 'cat_cols': chart_info['cat_cols'], 'num_col': chart_info['num_col'] }) # 创建子图布局 n_charts = len(chart_list) cols = 2 subplot_titles = [] specs = [] i = 0 while i < n_charts: chart = chart_list[i] if chart.get('full_width'): subplot_titles.append(chart['title']) specs.append([{"type": "domain" if chart['type'] in ['pie', 'sunburst'] else "table" if chart['type'] == 'table' else "xy", "colspan": 2}, None]) i += 1 else: subplot_titles.append(chart['title']) if i + 1 < n_charts: next_chart = chart_list[i + 1] subplot_titles.append(next_chart['title']) # 判断图表类型 def get_trace_type(chart_type): if chart_type in ['pie', 'sunburst']: return "domain" elif chart_type == 'heatmap': return "xy" else: return "xy" type1 = get_trace_type(chart['type']) type2 = get_trace_type(next_chart['type']) specs.append([{"type": type1}, {"type": type2}]) i += 2 else: type1 = get_trace_type(chart['type']) specs.append([{"type": type1}, {"type": "xy"}]) i += 1 rows = len(specs) # 创建子图 fig = make_subplots( rows=rows, cols=cols, subplot_titles=subplot_titles, specs=specs, vertical_spacing=0.08, horizontal_spacing=0.10 ) # 添加图表 current_row = 1 current_col = 1 color_idx = 0 # 定义位置推进函数 def advance_position(): nonlocal current_row, current_col, color_idx if current_col == 2: current_col = 1 current_row += 1 else: current_col = 2 color_idx += 1 for chart in chart_list: chart_type = chart['type'] if chart_type == 'table': # 数据表格 fig.add_trace( go.Table( columnwidth=[200] + [150] * (len(df.columns) - 1), header=dict( values=[f"{col}" for col in df.columns], fill_color=color_list[color_idx % len(color_list)], font=dict(color='white', size=14, family='Arial'), align='center', height=45, line=dict(color='white', width=2) ), cells=dict( values=[df[col] for col in df.columns], fill_color=[['#F8FAFC', '#EEF2FF'] * (len(df) // 2 + 1)], font=dict(color='#1E293B', size=13, family='Arial'), align='center', height=38, line=dict(color='#E2E8F0', width=1) ) ), row=current_row, col=1 ) current_row += 1 elif chart_type == 'histogram': # 直方图 fig.add_trace( go.Histogram( x=df[chart['col']], name=chart['col'], marker_color=color_list[color_idx % len(color_list)], opacity=0.8, marker_line=dict(color='white', width=2), hovertemplate=f'{chart["col"]}: %{{x}}
频数: %{{y}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'value_counts_bar': # 值计数柱状图 value_counts = df[chart['col']].value_counts() fig.add_trace( go.Bar( x=value_counts.index.astype(str), y=value_counts.values, name=chart['col'], marker_color=color_list[color_idx % len(color_list)], opacity=0.9, marker_line=dict(color='white', width=2), text=value_counts.values, textposition='outside', hovertemplate=f'{chart["col"]}: %{{x}}
数量: %{{y}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'box_single': # 单变量箱线图 fig.add_trace( go.Box( y=df[chart['col']], name=chart['col'], marker_color=color_list[color_idx % len(color_list)], boxmean='sd', marker=dict(size=5), line=dict(width=2) ), row=current_row, col=current_col ) advance_position() elif chart_type == 'grouped_bar': # 分组柱状图 grouped = df.groupby(chart['cat_col'])[chart['num_col']].sum().reset_index() fig.add_trace( go.Bar( x=grouped[chart['cat_col']], y=grouped[chart['num_col']], name=chart['num_col'], marker_color=color_list[color_idx % len(color_list)], opacity=0.9, marker_line=dict(color='white', width=2), text=[f'{v:,.0f}' for v in grouped[chart['num_col']]], textposition='outside', textfont=dict(size=11), hovertemplate=f'{chart["cat_col"]}: %{{x}}
{chart["num_col"]}: %{{y:,.0f}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'pie': # 饼图 grouped = df.groupby(chart['cat_col'])[chart['num_col']].sum().reset_index() fig.add_trace( go.Pie( labels=grouped[chart['cat_col']], values=grouped[chart['num_col']], hole=0.6, marker_colors=color_list[:len(grouped)], textinfo='percent+label', textposition='outside', textfont=dict(size=13), hovertemplate='%{label}
数值: %{value:,.0f}
占比: %{percent}', pull=[0.08 if i == 0 else 0.02 for i in range(len(grouped))], marker=dict(line=dict(color='white', width=3)), rotation=45 ), row=current_row, col=current_col ) advance_position() elif chart_type == 'box_grouped': # 分组箱线图 fig.add_trace( go.Box( x=df[chart['cat_col']], y=df[chart['num_col']], name=chart['num_col'], marker_color=color_list[color_idx % len(color_list)], boxmean='sd', boxpoints='outliers', marker=dict(size=6), line=dict(width=3), opacity=0.7 ), row=current_row, col=current_col ) advance_position() elif chart_type == 'violin': # 小提琴图 fig.add_trace( go.Violin( x=df[chart['cat_col']], y=df[chart['num_col']], name=chart['num_col'], box_visible=True, meanline_visible=True, fillcolor=color_list[color_idx % len(color_list)], opacity=0.6, marker=dict(size=3) ), row=current_row, col=current_col ) advance_position() elif chart_type == 'line': # 折线图 x_data = df[chart['x_col']] if chart['x_col'] in df.columns else list(range(len(df))) fig.add_trace( go.Scatter( x=x_data, y=df[chart['y_col']], mode='lines+markers', name=chart['y_col'], line=dict( color=color_list[color_idx % len(color_list)], width=3, shape='spline', smoothing=1.3 ), marker=dict(size=8, color='white', line=dict(color=color_list[color_idx % len(color_list)], width=2)), hovertemplate=f'{chart["y_col"]}: %{{y:,.0f}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'area': # 面积图 x_data = df[chart['x_col']] if chart['x_col'] in df.columns else list(range(len(df))) fig.add_trace( go.Scatter( x=x_data, y=df[chart['y_col']], mode='lines', name=chart['y_col'], fill='tozeroy', fillcolor=f'rgba(99, 102, 241, 0.2)', line=dict(color=color_list[color_idx % len(color_list)], width=3), hovertemplate=f'{chart["y_col"]}: %{{y:,.0f}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'scatter': # 散点图 fig.add_trace( go.Scatter( x=df[chart['x_col']], y=df[chart['y_col']], mode='markers', name=f'{chart["x_col"]} vs {chart["y_col"]}', marker=dict( color=color_list[color_idx % len(color_list)], size=12, opacity=0.6, line=dict(color='white', width=2) ), hovertemplate=f'{chart["x_col"]}: %{{x:,.0f}}
{chart["y_col"]}: %{{y:,.0f}}' ), row=current_row, col=current_col ) advance_position() elif chart_type == 'heatmap': # 相关性热力图 if characteristics['correlations'] is not None: corr = characteristics['correlations'] fig.add_trace( go.Heatmap( z=corr.values, x=corr.columns, y=corr.index, colorscale='RdBu', zmid=0, text=[[f'{val:.2f}' for val in row] for row in corr.values], texttemplate='%{text}', textfont=dict(size=10), hovertemplate='%{x} vs %{y}: %{z:.3f}', colorbar=dict(title='相关系数', len=0.8) ), row=current_row, col=1 ) current_row += 1 elif chart_type == 'parallel_coordinates': # 平行坐标图 dimensions = [] for col in chart['cols']: dimensions.append(dict( label=col, values=df[col] )) fig.add_trace( go.Parcoords( line=dict( color=df[chart['cols'][0]], colorscale='Viridis', showscale=True ), dimensions=dimensions ), row=current_row, col=1 ) current_row += 1 elif chart_type == 'sunburst': # 旭日图 if chart['num_col']: path = chart['cat_cols'] + [chart['num_col']] fig.add_trace( go.Sunburst( labels=df[chart['cat_cols'][0]], parents=[''] * len(df), values=df[chart['num_col']] if chart['num_col'] else None, marker_colors=color_list[:len(df)], hovertemplate='%{label}
数值: %{value}' ), row=current_row, col=current_col ) advance_position() # 更新布局 fig.update_layout( height=420 * rows, title=dict(text="", x=0.5, xanchor='center'), paper_bgcolor='rgba(0,0,0,0)', plot_bgcolor='#FFFFFF', showlegend=False, hovermode='closest', margin=dict(l=70, r=70, t=30, b=70), font=dict(family='Arial', size=13) ) # 更新坐标轴样式 fig.update_xaxes( showgrid=True, gridwidth=1, gridcolor='#F1F5F9', linecolor='#CBD5E1', linewidth=2, ticks='outside', tickwidth=2, ticklen=6, tickfont=dict(size=12, family='Arial', color='#475569'), title_font=dict(size=14, family='Arial', color='#1E293B'), title_standoff=15, mirror=False ) fig.update_yaxes( showgrid=True, gridwidth=1, gridcolor='#F1F5F9', linecolor='#CBD5E1', linewidth=2, ticks='outside', tickwidth=2, ticklen=6, tickfont=dict(size=12, family='Arial', color='#475569'), title_font=dict(size=14, family='Arial', color='#1E293B'), title_standoff=15, mirror=False ) # 更新子图标题样式 fig.update_annotations(font=dict(size=16, family='Arial', color='#1E293B'), borderpad=10) # 生成图表HTML plotly_html = fig.to_html( include_plotlyjs='cdn', full_html=False, config={ 'displayModeBar': False, 'displaylogo': False, 'scrollZoom': True, 'editable': False } ) # 完整HTML模板 html_template = f""" 数据可视化仪表板

📊 智能数据可视化仪表板

AI驱动的数据分析与可视化
生成时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}
{metrics_html}
{plotly_html}

💡 AI智能数据可视化 · 自动选择最佳图表类型

""" # 写入文件 with open(output_file, 'w', encoding='utf-8') as f: f.write(html_template) def generate_auto_charts(df, output_dir="auto_charts"): """ 自动生成图表 Args: df: pandas DataFrame output_dir: 输出目录 Returns: 生成的图表文件列表 """ os.makedirs(output_dir, exist_ok=True) dashboard_file = os.path.join(output_dir, "visualization_dashboard.html") generate_unified_dashboard(df, output_file=dashboard_file) return [dashboard_file] def main(): parser = argparse.ArgumentParser(description='数据可视化图表生成工具 - AI智能分析版本') parser.add_argument('--input', '-i', required=True, help='输入数据文件路径 (CSV或Excel)') parser.add_argument('--output', '-o', default='visualization_dashboard.html', help='输出文件路径') parser.add_argument('--output-dir', help='输出目录') args = parser.parse_args() try: # 读取数据 df = read_data(args.input) print(f"✅ 成功读取数据，共 {len(df)} 行，{len(df.columns)} 列") print(f"📋 列名: {list(df.columns)}") # 分析数据特征 print(f"\n🔍 正在分析数据特征...") characteristics = analyze_data_characteristics(df) print(f" - 数值列: {characteristics['numeric_count']} 个") print(f" - 分类列: {characteristics['categorical_count']} 个") print(f" - 日期列: {characteristics['datetime_count']} 个") # 智能推荐图表 print(f"\n🤖 AI正在推荐最佳图表组合...") recommended = recommend_charts(characteristics) print(f" 推荐了 {len(recommended)} 种图表类型:") chart_type_names = { 'table': '数据表格', 'histogram': '直方图', 'value_counts_bar': '值计数柱状图', 'box_single': '箱线图', 'grouped_bar': '分组柱状图', 'pie': '饼图', 'box_grouped': '分组箱线图', 'violin': '小提琴图', 'line': '折线图', 'area': '面积图', 'scatter': '散点图', 'heatmap': '热力图', 'parallel_coordinates': '平行坐标图', 'sunburst': '旭日图' } for chart in recommended: chart_name = chart_type_names.get(chart['type'], chart['type']) print(f" - {chart_name} (优先级: {chart['priority']})") # 生成仪表板 output_file = args.output generate_unified_dashboard(df, output_file) print(f"\n✨ AI智能可视化仪表板已成功生成!") print(f"📁 文件路径: {output_file}") print(f"\n💡 提示: 图表类型由AI根据数据特征自动选择") except Exception as e: print(f"❌ 错误: {str(e)}", file=sys.stderr) import traceback traceback.print_exc() sys.exit(1) if __name__ == "__main__": main()