Back to Details

paper-figure

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Paper Figure: Publication-Quality Plots from Experiment Data

论文图表：基于实验数据生成出版级别的可视化图

Generate all figures and tables for a paper based on: $ARGUMENTS
基于**$ARGUMENTS**生成论文所需的所有图表与表格。

Scope: What This Skill Can and Cannot Do

适用范围：本技能能做什么，不能做什么

CategoryCan auto-generate?Examples
Data-driven plots✅ YesLine plots (training curves), bar charts (method comparison), scatter plots, heatmaps, box/violin plots
Comparison tables✅ YesLaTeX tables comparing prior bounds, method features, ablation results
Multi-panel figures✅ YesSubfigure grids combining multiple plots (e.g., 3×3 dataset × method)
Architecture/pipeline diagrams❌ No — manualModel architecture, data flow diagrams, system overviews. At best can generate a rough TikZ skeleton, but expect to draw these yourself using tools like draw.io, Figma, or TikZ
Generated image grids❌ No — manualGrids of generated samples (e.g., GAN/diffusion outputs). These come from running your model, not from this skill
Photographs / screenshots❌ No — manualReal-world images, UI screenshots, qualitative examples
In practice: For a typical ML paper, this skill handles ~60% of figures (all data plots + tables). The remaining ~40% (hero figure, architecture diagram, qualitative results) need to be created manually and placed in 
figures/
before running 
/paper-write
. The skill will detect these as "existing figures" and preserve them.
类别能否自动生成？示例
数据驱动型图表✅ 可以折线图（训练曲线）、柱状图（方法对比）、散点图、热力图、箱线图/小提琴图
对比表格✅ 可以用于对比现有边界、方法特性、消融实验结果的LaTeX表格
多面板组合图✅ 可以组合多个子图的网格布局（例如3×3的数据集×方法对比图）
架构/流程图❌ 不可以 — 需手动制作模型架构图、数据流图、系统概览图。最多只能生成一个粗略的TikZ框架，但建议你使用draw.io、Figma或TikZ等工具手动绘制
生成图像网格❌ 不可以 — 需手动制作生成样本的网格图（例如GAN/扩散模型的输出结果）。这类图来自你的模型运行结果，而非本技能生成
照片/截图❌ 不可以 — 需手动制作真实世界图像、UI截图、定性示例
实际应用： 对于一篇典型的机器学习论文，本技能可处理约60%的图表（所有数据类图表+表格）。剩余约40%的内容（核心展示图、架构图、定性结果图）需要手动制作并放入
figures/
目录后，再运行
/paper-write
。本技能会将这些检测为“已存在的图表”并保留它们。

Constants

常量参数

  • STYLE = 
    publication
     — Visual style preset. Options: 
    publication
    (default, clean for print), 
    poster
    (larger fonts), 
    slide
    (bold colors)
  • DPI = 300 — Output resolution
  • FORMAT = 
    pdf
     — Output format. Options: 
    pdf
    (vector, best for LaTeX), 
    png
    (raster fallback)
  • COLOR_PALETTE = 
    tab10
     — Default matplotlib color cycle. Options: 
    tab10
    , 
    Set2
    , 
    colorblind
    (deuteranopia-safe)
  • FONT_SIZE = 10 — Base font size (matches typical conference body text)
  • FIG_DIR = 
    figures/
     — Output directory for generated figures
  • REVIEWER_MODEL = 
    gpt-5.4
     — Model used via Codex MCP for figure quality review.
  • STYLE = 
    publication
     — 可视化风格预设。可选值：
    publication
    （默认，适合印刷的简洁风格）、
    poster
    （更大字体）、
    slide
    （醒目色彩）
  • DPI = 300 — 输出分辨率
  • FORMAT = 
    pdf
     — 输出格式。可选值：
    pdf
    （矢量图，最适合LaTeX）、
    png
    （栅格图备选）
  • COLOR_PALETTE = 
    tab10
     — 默认matplotlib配色循环。可选值：
    tab10
    、
    Set2
    、
    colorblind
    （适合红绿色盲人群）
  • FONT_SIZE = 10 — 基础字体大小（匹配常规会议论文的正文字号）
  • FIG_DIR = 
    figures/
     — 生成图表的输出目录
  • REVIEWER_MODEL = 
    gpt-5.4
     — 通过Codex MCP用于图表质量审核的模型。

Inputs

输入要求

  1. PAPER_PLAN.md — figure plan table (from 
    /paper-plan
    )
  2. Experiment data — JSON files, CSV files, or screen logs in 
    figures/
    or project root
  3. Existing figures — any manually created figures to preserve
If no PAPER_PLAN.md exists, scan for data files and ask the user which figures to generate.
  1. PAPER_PLAN.md — 图表规划表（来自
    /paper-plan
    ）
  2. 实验数据 — JSON文件、CSV文件或屏幕日志，存放在
    figures/
    目录或项目根目录
  3. 已存在的图表 — 任何需要保留的手动制作图表
如果不存在PAPER_PLAN.md，则扫描数据文件并询问用户需要生成哪些图表。

Workflow

工作流程

Step 1: Read Figure Plan

步骤1：读取图表规划表

Parse the Figure Plan table from PAPER_PLAN.md:
markdown
| ID | Type | Description | Data Source | Priority |
|----|------|-------------|-------------|----------|
| Fig 1 | Architecture | ... | manual | HIGH |
| Fig 2 | Line plot | ... | figures/exp.json | HIGH |
Identify:
  • Which figures can be auto-generated from data
  • Which need manual creation (architecture diagrams, etc.)
  • Which are comparison tables (generate as LaTeX)
解析PAPER_PLAN.md中的图表规划表：
markdown
| ID | 类型 | 描述 | 数据源 | 优先级 |
|----|------|-------------|-------------|----------|
| Fig 1 | 架构图 | ... | 手动制作 | HIGH |
| Fig 2 | 折线图 | ... | figures/exp.json | HIGH |
识别：
  • 哪些图表可以从数据自动生成
  • 哪些需要手动制作（架构图等）
  • 哪些是对比表格（需生成为LaTeX格式）

Step 2: Set Up Plotting Environment

步骤2：配置绘图环境

Create a shared style configuration script:
python
undefined
创建共享的风格配置脚本：
python
undefined

paper_plot_style.py — shared across all figure scripts

paper_plot_style.py — 所有图表脚本共享的配置

import matplotlib.pyplot as plt import matplotlib matplotlib.rcParams.update({ 'font.size': FONT_SIZE, 'font.family': 'serif', 'font.serif': ['Times New Roman', 'Times', 'DejaVu Serif'], 'axes.labelsize': FONT_SIZE, 'axes.titlesize': FONT_SIZE + 1, 'xtick.labelsize': FONT_SIZE - 1, 'ytick.labelsize': FONT_SIZE - 1, 'legend.fontsize': FONT_SIZE - 1, 'figure.dpi': DPI, 'savefig.dpi': DPI, 'savefig.bbox': 'tight', 'savefig.pad_inches': 0.05, 'axes.grid': False, 'axes.spines.top': False, 'axes.spines.right': False, 'text.usetex': False, # set True if LaTeX is available 'mathtext.fontset': 'stix', })
import matplotlib.pyplot as plt import matplotlib matplotlib.rcParams.update({ 'font.size': FONT_SIZE, 'font.family': 'serif', 'font.serif': ['Times New Roman', 'Times', 'DejaVu Serif'], 'axes.labelsize': FONT_SIZE, 'axes.titlesize': FONT_SIZE + 1, 'xtick.labelsize': FONT_SIZE - 1, 'ytick.labelsize': FONT_SIZE - 1, 'legend.fontsize': FONT_SIZE - 1, 'figure.dpi': DPI, 'savefig.dpi': DPI, 'savefig.bbox': 'tight', 'savefig.pad_inches': 0.05, 'axes.grid': False, 'axes.spines.top': False, 'axes.spines.right': False, 'text.usetex': False, # 如果有LaTeX环境则设为True 'mathtext.fontset': 'stix', })

Color palette

配色方案

COLORS = plt.cm.tab10.colors # or Set2, or colorblind-safe
def save_fig(fig, name, fmt=FORMAT): """Save figure to FIG_DIR with consistent naming.""" fig.savefig(f'{FIG_DIR}/{name}.{fmt}') print(f'Saved: {FIG_DIR}/{name}.{fmt}')
undefined
COLORS = plt.cm.tab10.colors # 或Set2，或色盲友好配色
def save_fig(fig, name, fmt=FORMAT): """将图表保存到FIG_DIR目录，命名保持一致。""" fig.savefig(f'{FIG_DIR}/{name}.{fmt}') print(f'已保存: {FIG_DIR}/{name}.{fmt}')
undefined

Step 3: Auto-Select Figure Type

步骤3：自动选择图表类型

Use this decision tree for data-driven figures (inspired by Imbad0202/academic-research-skills):
Data PatternRecommended TypeSize
X=time/steps, Y=metricLine plot0.48\textwidth
Methods × 1 metricBar chart0.48\textwidth
Methods × multiple metricsGrouped bar / radar0.95\textwidth
Two continuous variablesScatter plot0.48\textwidth
Matrix / grid valuesHeatmap0.48\textwidth
Distribution comparisonBox/violin plot0.48\textwidth
Multi-dataset resultsMulti-panel (subfigure)0.95\textwidth
Prior work comparisonLaTeX table—
使用以下决策树选择数据驱动型图表的类型（灵感来自Imbad0202/academic-research-skills）：
数据模式推荐图表类型尺寸
X=时间/步数，Y=指标折线图0.48\textwidth
方法 × 1个指标柱状图0.48\textwidth
方法 × 多个指标分组柱状图 / 雷达图0.95\textwidth
两个连续变量散点图0.48\textwidth
矩阵/网格值热力图0.48\textwidth
分布对比箱线图/小提琴图0.48\textwidth
多数据集结果多面板组合图（子图）0.95\textwidth
现有工作对比LaTeX表格—

Step 4: Generate Each Figure

步骤4：生成每个图表

For each figure in the plan, create a standalone Python script:
Line plots (training curves, scaling):
python
undefined
为规划表中的每个图表创建独立的Python脚本：
折线图（训练曲线、缩放趋势）：
python
undefined

gen_fig2_training_curves.py

gen_fig2_training_curves.py

from paper_plot_style import * import json
with open('figures/exp_results.json') as f: data = json.load(f)
fig, ax = plt.subplots(1, 1, figsize=(5, 3.5)) ax.plot(data['steps'], data['fac_loss'], label='Factorized', color=COLORS[0]) ax.plot(data['steps'], data['crf_loss'], label='CRF-LR', color=COLORS[1]) ax.set_xlabel('Training Steps') ax.set_ylabel('Cross-Entropy Loss') ax.legend(frameon=False) save_fig(fig, 'fig2_training_curves')

**Bar charts** (comparison, ablation):
```python
fig, ax = plt.subplots(1, 1, figsize=(5, 3))
methods = ['Baseline', 'Method A', 'Method B', 'Ours']
values = [82.3, 85.1, 86.7, 89.2]
bars = ax.bar(methods, values, color=[COLORS[i] for i in range(len(methods))])
ax.set_ylabel('Accuracy (%)')
from paper_plot_style import * import json
with open('figures/exp_results.json') as f: data = json.load(f)
fig, ax = plt.subplots(1, 1, figsize=(5, 3.5)) ax.plot(data['steps'], data['fac_loss'], label='Factorized', color=COLORS[0]) ax.plot(data['steps'], data['crf_loss'], label='CRF-LR', color=COLORS[1]) ax.set_xlabel('训练步数') ax.set_ylabel('交叉熵损失') ax.legend(frameon=False) save_fig(fig, 'fig2_training_curves')

**柱状图**（方法对比、消融实验）：
```python
fig, ax = plt.subplots(1, 1, figsize=(5, 3))
methods = ['Baseline', 'Method A', 'Method B', 'Ours']
values = [82.3, 85.1, 86.7, 89.2]
bars = ax.bar(methods, values, color=[COLORS[i] for i in range(len(methods))])
ax.set_ylabel('准确率 (%)')

Add value labels on bars

在柱子上添加数值标签

for bar, val in zip(bars, values): ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 0.3, f'{val:.1f}', ha='center', va='bottom', fontsize=FONT_SIZE-1) save_fig(fig, 'fig3_comparison')

**Comparison tables** (LaTeX, for theory papers):
```latex
\begin{table}[t]
\centering
\caption{Comparison of estimation error bounds. $n$: sample size, $D$: ambient dim, $d$: latent dim, $K$: subspaces, $n_k$: modes.}
\label{tab:bounds}
\begin{tabular}{lccc}
\toprule
Method & Rate & Depends on $D$? & Multi-modal? \\
\midrule
\citet{MinimaxOkoAS23} & $n^{-s'/D}$ & Yes (curse) & No \\
\citet{ScoreMatchingdistributionrecovery} & $n^{-2/d}$ & No & No \\
\textbf{Ours} & $\sqrt{\sum n_k d_k / n}$ & No & Yes \\
\bottomrule
\end{tabular}
\end{table}
Architecture/pipeline diagrams (MANUAL — outside this skill's scope):
  • These require manual creation using draw.io, Figma, Keynote, or TikZ
  • This skill can generate a rough TikZ skeleton as a starting point, but do not expect publication-quality results
  • If the figure already exists in 
    figures/
    , preserve it and generate only the LaTeX 
    \includegraphics
    snippet
  • Flag as 
    [MANUAL]
    in the figure plan and 
    latex_includes.tex
for bar, val in zip(bars, values): ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 0.3, f'{val:.1f}', ha='center', va='bottom', fontsize=FONT_SIZE-1) save_fig(fig, 'fig3_comparison')

**对比表格**（LaTeX格式，适用于理论论文）：
```latex
\begin{table}[t]
\centering
\caption{估计误差边界对比。$n$: 样本量, $D$: 环境维度, $d$: 潜在维度, $K$: 子空间, $n_k$: 模式。}
\label{tab:bounds}
\begin{tabular}{lccc}
\toprule
方法 & 速率 & 是否依赖$D$? & 多模式支持? \\
\midrule
\citet{MinimaxOkoAS23} & $n^{-s'/D}$ & 是（维度灾难） & 否 \\
\citet{ScoreMatchingdistributionrecovery} & $n^{-2/d}$ & 否 & 否 \\
\textbf{本文方法} & $\sqrt{\sum n_k d_k / n}$ & 否 & 是 \\
\bottomrule
\end{tabular}
\end{table}
架构/流程图（手动制作 — 超出本技能范围）：
  • 需要使用draw.io、Figma、Keynote或TikZ手动制作
  • 本技能可生成一个粗略的TikZ框架作为起点，但无法直接生成达到出版级别的结果
  • 如果图表已存在于
    figures/
    目录中，则保留该图表并仅生成LaTeX的
    \includegraphics
    代码片段
  • 在图表规划表和
    latex_includes.tex
    中标记为
    [MANUAL]

Step 5: Run All Scripts

步骤5：运行所有脚本

bash
undefined
bash
undefined

Run all figure generation scripts

运行所有图表生成脚本

for script in gen_fig*.py; do python "$script" done

Verify all output files exist and are non-empty.
for script in gen_fig*.py; do python "$script" done

验证所有输出文件存在且非空。

Step 6: Generate LaTeX Include Snippets

步骤6：生成LaTeX引用代码片段

For each figure, output the LaTeX code to include it:
latex
% === Fig 2: Training Curves ===
\begin{figure}[t]
    \centering
    \includegraphics[width=0.48\textwidth]{figures/fig2_training_curves.pdf}
    \caption{Training curves comparing factorized and CRF-LR denoising.}
    \label{fig:training_curves}
\end{figure}
Save all snippets to 
figures/latex_includes.tex
for easy copy-paste into the paper.
为每个图表输出用于引用的LaTeX代码：
latex
% === Fig 2: 训练曲线 ===
\begin{figure}[t]
    \centering
    \includegraphics[width=0.48\textwidth]{figures/fig2_training_curves.pdf}
    \caption{对比分解式和CRF-LR去噪的训练曲线。}
    \label{fig:training_curves}
\end{figure}
将所有代码片段保存到
figures/latex_includes.tex
，以便轻松复制粘贴到论文中。

Step 7: Figure Quality Review with REVIEWER_MODEL

步骤7：使用REVIEWER_MODEL进行图表质量审核

Send figure descriptions and captions to GPT-5.4 for review:
mcp__codex__codex:
  model: gpt-5.4
  config: {"model_reasoning_effort": "xhigh"}
  prompt: |
    Review these figure/table plans for a [VENUE] submission.

    For each figure:
    1. Is the caption informative and self-contained?
    2. Does the figure type match the data being shown?
    3. Is the comparison fair and clear?
    4. Any missing baselines or ablations?
    5. Would a different visualization be more effective?

    [list all figures with captions and descriptions]
将图表描述和标题发送给GPT-5.4进行审核：
mcp__codex__codex:
  model: gpt-5.4
  config: {"model_reasoning_effort": "xhigh"}
  prompt: |
    审核这些用于[VENUE]投稿的图表/表格规划。

    针对每个图表：
    1. 标题是否信息丰富且独立完整？
    2. 图表类型是否与展示的数据匹配？
    3. 对比是否公平清晰？
    4. 是否缺少基线或消融实验？
    5. 是否有更有效的可视化方式？

    [列出所有图表的标题和描述]

Step 8: Quality Checklist

步骤8：质量检查清单

Before finishing, verify each figure (from pedrohcgs/claude-code-my-workflow):
  • Font size readable at printed paper size (not too small)
  • Colors distinguishable in grayscale (print-friendly)
  • No title inside figures — titles go only in LaTeX 
    \caption{}
    (from pedrohcgs)
  • Legend does not overlap data
  • Axis labels have units where applicable
  • Axis labels are publication-quality (not variable names like 
    emp_rate
    )
  • Figure width fits single column (0.48\textwidth) or full width (0.95\textwidth)
  • PDF output is vector (not rasterized text)
  • No matplotlib default title (remove 
    plt.title
    for publications)
  • Serif font matches paper body text (Times / Computer Modern)
  • Colorblind-accessible (if using colorblind palette)
完成前，验证每个图表（参考pedrohcgs/claude-code-my-workflow）：
  • 字体大小在打印后可读（不要过小）
  • 颜色在灰度模式下可区分（适合印刷）
  • 图表内部不要加标题 — 标题仅放在LaTeX的
    \caption{}
    中（来自pedrohcgs）
  • 图例不与数据重叠
  • 坐标轴标签在必要时包含单位
  • 坐标轴标签达到出版级别（不要使用
    emp_rate
    这类变量名）
  • 图表宽度适配单栏（0.48\textwidth）或全宽度（0.95\textwidth）
  • PDF输出为矢量图（不是栅格化的文本）
  • 没有matplotlib默认标题（移除
    plt.title
    以符合出版要求）
  • 衬线字体与论文正文字体匹配（Times / Computer Modern）
  • 色盲友好（如果使用色盲友好配色）

Output

输出目录结构

figures/
├── paper_plot_style.py          # shared style config
├── gen_fig1_architecture.py     # per-figure scripts
├── gen_fig2_training_curves.py
├── gen_fig3_comparison.py
├── fig1_architecture.pdf        # generated figures
├── fig2_training_curves.pdf
├── fig3_comparison.pdf
├── latex_includes.tex           # LaTeX snippets for all figures
└── TABLE_*.tex                  # standalone table LaTeX files
figures/
├── paper_plot_style.py          # 共享风格配置
├── gen_fig1_architecture.py     # 每个图表对应的脚本
├── gen_fig2_training_curves.py
├── gen_fig3_comparison.py
├── fig1_architecture.pdf        # 生成的图表
├── fig2_training_curves.pdf
├── fig3_comparison.pdf
├── latex_includes.tex           # 所有图表的LaTeX引用片段
└── TABLE_*.tex                  # 独立的LaTeX表格文件

Key Rules

核心规则

  • Every figure must be reproducible — save the generation script alongside the output
  • Do NOT hardcode data — always read from JSON/CSV files
  • Use vector format (PDF) for all plots — PNG only as fallback
  • No decorative elements — no background colors, no 3D effects, no chart junk
  • Consistent style across all figures — same fonts, colors, line widths
  • Colorblind-safe — verify with https://davidmathlogic.com/colorblind/ if needed
  • One script per figure — easy to re-run individual figures when data changes
  • No titles inside figures — captions are in LaTeX only
  • Comparison tables count as figures — generate them as standalone .tex files
  • 每个图表必须可复现 — 将生成脚本与输出文件保存在一起
  • 不要硬编码数据 — 始终从JSON/CSV文件读取数据
  • 所有图表使用矢量格式（PDF） — PNG仅作为备选
  • 不要添加装饰元素 — 无背景色、无3D效果、无冗余图表元素
  • 所有图表风格一致 — 相同字体、颜色、线条宽度
  • 色盲友好 — 必要时使用https://davidmathlogic.com/colorblind/验证
  • 一个图表对应一个脚本 — 数据变化时可轻松重新生成单个图表
  • 图表内部不要加标题 — 标题仅放在LaTeX中
  • 对比表格视为图表的一部分 — 生成为独立的.tex文件

Figure Type Reference

图表类型参考

TypeWhen to UseTypical Size
Line plotTraining curves, scaling trends0.48\textwidth
Bar chartMethod comparison, ablation0.48\textwidth
Grouped barMulti-metric comparison0.95\textwidth
Scatter plotCorrelation analysis0.48\textwidth
HeatmapAttention, confusion matrix0.48\textwidth
Box/violinDistribution comparison0.48\textwidth
ArchitectureSystem overview0.95\textwidth
Multi-panelCombined results (subfigures)0.95\textwidth
Comparison tablePrior bounds vs. ours (theory)full width
类型使用场景典型尺寸
折线图训练曲线、缩放趋势0.48\textwidth
柱状图方法对比、消融实验0.48\textwidth
分组柱状图多指标对比0.95\textwidth
散点图相关性分析0.48\textwidth
热力图注意力机制、混淆矩阵0.48\textwidth
箱线图/小提琴图分布对比0.48\textwidth
架构图系统概览0.95\textwidth
多面板组合图组合结果（子图）0.95\textwidth
对比表格现有边界与本文方法对比（理论论文）全宽度

Acknowledgements

致谢

Design pattern (type × style matrix) inspired by baoyu-skills. Publication style defaults and figure rules from pedrohcgs/claude-code-my-workflow. Visualization decision tree from Imbad0202/academic-research-skills.
设计模式（类型×风格矩阵）灵感来自baoyu-skills。出版风格默认设置和图表规则来自pedrohcgs/claude-code-my-workflow。可视化决策树来自Imbad0202/academic-research-skills。