图表设计详解与实战指南

plt.plot(epochs, train_loss, label='Train', linewidth=2)plt.plot(epochs, val_loss, label='Val', linewidth=2)plt.xlabel('Epoch', fontsize=14)plt.ylabel('Loss', fontsize=14)plt.legend(fontsize=12)plt.savefig('training_curve.pdf', bbox_inches='tight')

latex_code = "\\begin{tikzpicture}[node distance=2cm]\n\\node[rectangle,draw] (input) {Input Layer};\n\\node[rectangle,draw,right=of input] (hidden) {Hidden Layer};\n\\draw[->] (input) -- (hidden);\n\\end{tikzpicture}"

fig = go.Figure()fig.add_trace(go.Scatter(x=epochs, y=train_loss, name='Train'))fig.update_layout(hovermode='x unified', title='Training Progress')fig.write_html('interactive_chart.html')

import numpy as np class ChartDesignPrinciples:    """论文图表设计原则校验器"""    @staticmethod    def check_self_contained(chart_config):        """检查图表是否自包含(标题标签单位说明齐全)"""        required = ["title", "xlabel", "ylabel", "units"]        missing = [k for k in required if not chart_config.get(k)]        return {"self_contained": len(missing) == 0, "missing": missing}     @staticmethod    def check_resolution(dpi, format_type):        """检查分辨率是否满足发表要求"""        if format_type in ["pdf", "svg", "eps"]:            return {"vector": True, "suitable": True, "note": "矢量图无限缩放无失真"}        min_dpi = 300        return {"vector": False, "dpi": dpi, "suitable": dpi >= min_dpi,                "note": f"位图需{min_dpi}DPI以上,当前{dpi}DPI"}     @staticmethod    def get_colorblind_palette():        """获取色盲友好配色方案"""        return {            "blue": "#0072B2", "orange": "#E69F00", "green": "#009E73",            "red": "#D55E00", "purple": "#CC79A7", "cyan": "#56B4E9",            "yellow": "#F0E442", "grey": "#999999"        }  class MatplotlibChartBuilder:    """Matplotlib图表构建器: 模拟pyplot接口绘图"""    def __init__(self, style="seaborn-v0.8-whitegrid"):        self.style = style        self.figures = {}     def create_line_chart(self, title, x_data, y_data_list, labels, xlabel, ylabel):        """创建折线图: 适合展示训练曲线等时序数据"""        chart = {            "type": "line", "title": title,            "xlabel": xlabel, "ylabel": ylabel,            "style": self.style,            "series": [{"label": labels[i], "x": x_data, "y": y_data_list[i],                        "color": list(ChartDesignPrinciples.get_colorblind_palette().values())[i]}                       for i in range(len(y_data_list))],            "features": {"grid": True, "legend": True, "linewidth": 2,                         "fontsize": 12, "figsize": [8, 5]},            "export_format": "pdf",            "code_snippet": "plt.plot(x, y, label=label, linewidth=2)\nplt.xlabel('Epoch', fontsize=12)\nplt.ylabel('Loss', fontsize=12)\nplt.legend()\nplt.savefig('fig.pdf', bbox_inches='tight')"        }        return chart     def create_scatter_chart(self, title, x_data, y_data, xlabel, ylabel, color_by=None):        """创建散点图: 适合展示特征分布和相关性"""        return {            "type": "scatter", "title": title,            "xlabel": xlabel, "ylabel": ylabel,            "data_points": len(x_data),            "features": {"alpha": 0.6, "s": 30, "edgecolors": "black"},            "code_snippet": "plt.scatter(x, y, alpha=0.6, s=30, edgecolors='black')",            "export_format": "pdf"        }     def create_subplot_layout(self, n_rows, n_cols, chart_configs):        """管理多图布局: subplots控制对齐和间距"""        return {            "layout": f"{n_rows}x{n_cols}",            "total_charts": len(chart_configs),            "features": {"sharex": False, "sharey": False,                         "hspace": 0.3, "wspace": 0.3},            "code_snippet": f"fig, axes = plt.subplots({n_rows}, {n_cols}, figsize=(12, 8))\nfig.subplots_adjust(hspace=0.3, wspace=0.3)",            "charts": chart_configs        }     def apply_publication_style(self, chart):        """应用论文发表样式: 调整字体线宽和配色"""        chart["features"]["fontsize"] = 14        chart["features"]["linewidth"] = 2.0        chart["features"]["font_family"] = "serif"        chart["features"]["dpi"] = 300        chart["style_adjustments"] = [            "plt.rcParams['font.size'] = 14",            "plt.rcParams['font.family'] = 'serif'",            "plt.rcParams['axes.linewidth'] = 1.2",            "plt.rcParams['figure.dpi'] = 300"        ]        return chart  class TikZDiagramBuilder:    """TikZ矢量图形构建器: 模拟LaTeX内嵌绘图"""    def __init__(self):        self.diagrams = {}     def create_flowchart(self, nodes, edges):        """创建流程图: 适合展示算法流程和系统架构"""        tikz_nodes = []        for node in nodes:            tikz_nodes.append(                f"\\node[{node['style']}] ({node['id']}) at ({node['x']},{node['y']}) {{{node['label']}}};"            )        tikz_edges = []        for edge in edges:            style = edge.get("style", "->")            tikz_edges.append(f"\\draw[{style}] ({edge['from']}) -- ({edge['to']});")        full_code = "\\begin{tikzpicture}[node distance=2cm]\n" + \                    "\n".join(tikz_nodes) + "\n" + \                    "\n".join(tikz_edges) + "\n" + \                    "\\end{tikzpicture}"        return {            "type": "flowchart", "nodes": len(nodes), "edges": len(edges),            "latex_code": full_code,            "features": {"font_consistency": True, "vector": True,                         "editable": True, "latex_integrated": True}        }     def create_neural_network(self, layers):        """创建神经网络结构图"""        node_positions = []        edges = []        for layer_idx, layer_size in enumerate(layers):            x = layer_idx * 3            for node_idx in range(layer_size):                y = (node_idx - layer_size / 2) * 1.5                node_positions.append({"id": f"n{layer_idx}_{node_idx}",                                       "x": x, "y": y,                                       "label": "", "style": "circle,draw,minimum size=0.8cm"})        for i in range(len(layers) - 1):            for j in range(layers[i]):                for k in range(layers[i + 1]):                    edges.append({"from": f"n{i}_{j}", "to": f"n{i+1}_{k}", "style": "->,gray"})        return {            "type": "neural_network", "layers": layers,            "total_nodes": sum(layers), "total_edges": len(edges),            "features": {"vector": True, "latex_integrated": True,                         "font_consistency": True},            "code_preview": "\\begin{tikzpicture}\n% nodes and edges defined programmatically\n\\end{tikzpicture}"        }     def create_architecture_diagram(self, components, connections):        """创建架构图: 系统模块和数据流"""        return {            "type": "architecture",            "components": len(components),            "connections": len(connections),            "features": {"font_consistency": True, "vector": True,                         "seamless_latex": True,                         "advantage": "与LaTeX正文字体样式完全一致"},            "note": "TikZ架构图修改方便,调整只需改代码无需重新绘图"        }  class PlotlyInteractiveBuilder:    """Plotly交互式图表构建器"""    def __init__(self):        self.charts = {}     def create_interactive_line(self, title, x_data, y_data_list, labels):        """创建交互式折线图: 支持悬停缩放"""        return {            "type": "interactive_line", "title": title,            "series": len(y_data_list),            "features": {"hover": True, "zoom": True, "pan": True,                         "legend_click": True, "range_selector": True},            "code_snippet": "import plotly.graph_objects as go\nfig = go.Figure()\nfig.add_trace(go.Scatter(x=x, y=y, name=label))\nfig.update_layout(hovermode='x unified')",            "suitable_for": ["supplementary_material", "presentation", "demo"],            "not_suitable_for": ["paper_main_text"]        }     def create_interactive_heatmap(self, title, z_data, x_labels, y_labels):        """创建交互式热力图"""        return {            "type": "interactive_heatmap", "title": title,            "shape": f"{len(z_data)}x{len(z_data[0]) if z_data else 0}",            "features": {"hover_values": True, "color_scale": "Viridis",                         "zoom": True},            "code_snippet": "fig = go.Figure(data=go.Heatmap(z=z, x=x_labels, y=y_labels))",            "suitable_for": ["supplementary_material", "interactive_demo"]        }     def export_static(self, chart, format_type="pdf"):        """导出静态图片嵌入论文正文"""        chart["static_export"] = {            "format": format_type, "dpi": 300,            "note": "导出为静态图片后失去交互功能,但可嵌入论文正文"        }        return chart  class ChartToolSelector:    """图表工具选择器: 根据图表类型推荐最优工具"""    @staticmethod    def recommend(chart_type, use_case):        """根据图表类型和使用场景推荐工具"""        recommendations = {            ("data_chart", "paper_main_text"): {"tool": "Matplotlib",                "reason": "生成矢量PDF,精确控制样式,论文正文标准选择"},            ("data_chart", "supplementary"): {"tool": "Plotly",                "reason": "交互式探索数据细节,适合补充材料"},            ("data_chart", "presentation"): {"tool": "Plotly",                "reason": "交互展示效果好,适合演示文稿"},            ("structure_diagram", "paper_main_text"): {"tool": "TikZ",                "reason": "与LaTeX无缝集成,字体一致,矢量清晰"},            ("structure_diagram", "presentation"): {"tool": "TikZ/Draw.io",                "reason": "TikZ质量最高,Draw.io快速出图"},            ("architecture", "paper_main_text"): {"tool": "TikZ",                "reason": "复杂结构图矢量绘制,修改方便"},        }        key = (chart_type, use_case)        return recommendations.get(key, {"tool": "Matplotlib",                "reason": "通用数据图表的默认选择"})  if __name__ == "__main__":    print("=== 论文图表设计工具模拟器 ===\n")    mpl = MatplotlibChartBuilder()    epochs = list(range(1, 21))    train_loss = [2.5 * np.exp(-0.15 * e) + 0.1 for e in epochs]    val_loss = [2.5 * np.exp(-0.12 * e) + 0.15 + 0.02 * e for e in epochs]    line_chart = mpl.create_line_chart("Training Progress", epochs,                                       [train_loss, val_loss],                                       ["Train Loss", "Val Loss"],                                       "Epoch", "Loss")    line_chart = mpl.apply_publication_style(line_chart)    print("1. Matplotlib折线图:")    print(f"   类型: {line_chart['type']}, 标题: {line_chart['title']}")    print(f"   系列: {len(line_chart['series'])}, 导出: {line_chart['export_format']}")    print(f"   发表样式: 字体{line_chart['features']['fontsize']}pt, 线宽{line_chart['features']['linewidth']}")     tikz = TikZDiagramBuilder()    nn = tikz.create_neural_network([4, 6, 3])    print(f"\n2. TikZ神经网络图:")    print(f"   层结构: {nn['layers']}, 节点: {nn['total_nodes']}, 连接: {nn['total_edges']}")    print(f"   LaTeX集成: {nn['features']['latex_integrated']}, 矢量: {nn['features']['vector']}")     plotly = PlotlyInteractiveBuilder()    interactive = plotly.create_interactive_line("Interactive View", epochs,                                                  [train_loss, val_loss],                                                  ["Train", "Val"])    print(f"\n3. Plotly交互图:")    print(f"   悬停: {interactive['features']['hover']}, 缩放: {interactive['features']['zoom']}")    print(f"   适用: {interactive['suitable_for']}")     print(f"\n4. 工具推荐:")    selector = ChartToolSelector()    cases = [("data_chart", "paper_main_text"), ("structure_diagram", "paper_main_text"),             ("data_chart", "presentation")]    for ct, uc in cases:        rec = selector.recommend(ct, uc)        print(f"   {ct} + {uc} -> {rec['tool']}: {rec['reason'][:40]}...")