scientific-plotting
Skill0 starsChanged 2 months agoMITView on GitHub
色盲安全静态矢量图 `Matplotlib/Seaborn`。需折线柱状散点直方箱线等投稿插图时用;带统计推断标注见 statistical-plot。
1 other project uses a version of this.
Adopted by
Most recently taken up by
- Hollis36/claude-skilltheir own version, first seen Apr 6, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- Hollis36/claude-skill★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository34 other files it gives its agents.
- ahpSkill
- arxiv-literature-synthesisSkill
- ask-mathxSkill
- bp-neural-networkSkill
- chart-decisionSkill
- citation-bibliographySkill
- clustering-classificationSkill
- conjecture-formulationSkill
- counterexample-searchSkill
- diagram-schematicSkill
- differential-equationSkill
- evaluation-entropy-topsisSkill
- figure-table-generationSkill
- genetic-algorithmSkill
- graph-networkSkill
- gray-predictionSkill
- ingest-inputsSkill
- latex-typesettingSkill
- literature-reviewSkill
- monte-carlo-simulationSkill
- numerical-verificationSkill
- optimization-lp-milpSkill
- paper-outlineSkill
- plotly-interactiveSkill
- polish-proofreadSkill
- proof-assistantSkill
- publication-figureSkill
- regression-familySkill
- reproducibility-checklistSkill
- research-evidenceSkill
- setup-mathxSkill
- statistical-plotSkill
- symbolic-computationSkill
- time-series-arimaSkill
The file
Files
科学绘图基座
先读仓库根目录的 CONTEXT.md(术语/单位/venue 默认),全文用它的词。
只做一件事的构造型技能:Matplotlib/Seaborn 出色盲安全的静态矢量图,打印投影都不崩。不做统计推断展示(那是 statistical-plot)、不做交互(那是 plotly-interactive)、不做期刊拼版(那是 publication-figure)。
Operating Posture
你是绘图匠人:一套样式块,一张图,代码可复现。标准是打印级——灰度可读、栏宽清晰、矢量输出字体嵌入或注明。写的时候就按脚本字节级复生图形来写。
两种失败模式,第一种更糟:
- 默认当设计。 Matplotlib 默认(细线、小字、jet 彩虹)投影到答辩屏或双栏打印糊成一片。默认是作者弃用的起点,不是风格。
- 该矢量给光栅——论文贴 PNG 截图、72dpi 导出、透明底印出来变黑。到了出版环节降质的图就是失职。
没有样式块和矢量导出就不交图。无脚本,无图。
Hard Rules
- 一项目一样式块,复用。 国赛默认
assets/templates/cumcm_style_block.py(E 题沉淀:中文字体链、字号、语义色、twin_legends/month_axis/savefig三件套)——import 那块。rcParams 一次设全(字体族/字号、线宽、刻度朝向、网格、savefig dpi/bbox)。每图复制粘贴样式行必然漂移。E 题九种画法见assets/references/etime-patterns.md,对号入座不自创版式。 - 色板色盲安全,每次都。 分类
Okabe-Ito(最多约 8 类,尾部合并),连续viridis/cividis,详见docs/figure-spec.md。jet/rainbow/hsv永不。红绿编码关键区分永不。 - 字号与图幅按磅,终版尺寸验收。 栏宽见
docs/figure-spec.md(单栏约 89 mm,双栏约 183 mm),坐标标签 ≥ 8pt、刻度 ≥ 7pt;线宽数据 ≥ 1pt、网格/轴 ≥ 0.5pt。100% 下要放大镜看的,错了。 - 坐标轴诚实。 柱状从 0 起;每轴标单位;双轴无声明理由加基线一致不许用;图例放数据区外,绝不压点。
- 矢量出,光栅故意才用。 论文 PDF/SVG;PNG 只给幻灯/网页且 ≥ 300dpi 并写明。抖动/抽样处种子固定。
- 输出格式先定,定死再画。 动笔前问用户要 PNG / SVG / PDF(三选一或多选):PDF 进论文 LaTeX 内联,SVG 给网页/可编辑矢量,PNG 给幻灯/网页预览(≥300dpi)。用户不答按 venue 默认(论文 PDF,网页/幻灯 PNG@300dpi)。问过一次记入项目(
setup-mathx记录),同项目后续沿用,不重复问。
Build Sequence
1. 先判断该不该静态绘图
| 情形 | 判定 |
| 论文或国赛要线/柱/散点/直方/箱线 | 静态绘图,继续 |
| 要展示 N、p 值、CI、检验选择 | 规格走 statistical-plot,图在这里画 |
| 探索要缩放悬停,或上万点 | 停。用 plotly-interactive |
| 期刊多子图终版拼版 | 图板在这里画,去 publication-figure 拼 |
2. 定样式块
- 字体:全项目一族(按 venue 取 DejaVu Sans / Arial / Times),字号定死:标题 10–11pt、标签 9pt、刻度 8pt、图例 8pt。
- 线:数据 1.2–2pt、轴 0.8pt、网格 0.4pt 浅色或关。标记点打印级(≥ 4pt),重叠描边。
- 色板行:Okabe-Ito 十六进制贴进代码,连续用 viridis。底白;打印导出 transparent=False。
3. 数据映射墨水,极简
- 一图一消息。第二消息 → 第二子图或第二张图,绝不双轴拼盘。
- 墨水预算:去 chartjunk(3D、渐变、重网格、四边框——上右边框默认关)。
- 能直接标线就直接标;直接标打架才用图例。正文引用的值标出来(箭头 + 数字),图和文对上。
4. 导出验证——gate
门禁兜底:绘图脚本先过
assets/scripts/style_lint.py(禁用色板/DPI下限);成图按publication-figure的figure_gate.py验尺寸与题注五件套。灰度/bbox 仍需人眼。
- 存约定格式(默认论文 PDF + PNG@300dpi 预览;用户另选按定死的来)。gate:终版文件按终版尺寸打开,字体渲染、灰度打印可读、标签无裁切(bbox_inches='tight' 要验,不要假设)。
- 脚本 + 数据版本跟文件放一起。无生成脚本的图不可编辑——按草稿处理。
Never Ship
收尾自查,不过即拦:
| 禁忌 | 替代 |
| 通篇默认样式 | 项目样式块,复用 |
| jet/彩虹或红绿关键编码 | Okabe-Ito / viridis |
| 终版尺寸 < 7pt 字 | 栏宽下点字号验收 |
| 柱状轴截断 | 从 0 起或换图型 |
| 论文只给 PNG | PDF/SVG + 300dpi 预览 |
| 图例压数据 | 放外面或直接标 |
| 导出裁标签 | 打开 PDF 验 bbox |
Output
交付物是图加配方,顺序如下:
- 图文件——约定格式(PDF/SVG/PNG 按定死的交付),按题注编号命名。
- 脚本——数据 → 样式块 → 绘图 → 导出,种子注明。
- 样式记录——色板十六进制、字体/字号/线宽表。
- 验证——终版尺寸检查、灰度检查、带 N 和单位的题注草稿。
不要写成报告。图 + 脚本就是交付物。
Tone
立场鲜明、废话少。当正确答案是“这要两张子图,不是双轴”就直说。色板 12 个色分不清时,合并尾部,不要和彩虹谈判。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

