latex-drawing-guide
安装方式
先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。
用 AI 助手安装
复制提示词,让助手先核对固定版本和依赖。
查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "latex-drawing-guide":https://skillpi.com/zh/skills/deepskill-647f25da2362175de99319187a452c7a4edef2fc12d4d61e7bde2b7272dff06c。安装前核对上游来源和版本:https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/writing/latex/latex-drawing-guide/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。
此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。
npx skills add https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/tree/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/43-wentorai-research-plugins/skills/writing/latex/latex-drawing-guide --skill latex-drawing-guide来源与配置
查看 GitHub 源码brycewang-stanford/Auto-Empirical-Research-Skills固定提交 bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba
配置方式待核实
请查看原始来源中的适用工具、依赖和权限要求。
安装实测待补充
本条目面向文献与学术工具场景,提供“latex-drawing-guide”技能的目录入口。公开标签:learning、TikZ、PGFPlots、LaTeX绘图。请通过固定上游版本核对实际功能、依赖与权限;此页为 Skillπ 独立编写的用途介绍和原站链接,不代表运行安全保证。原文与附件依各自许可开放。
文件浏览
---
name: latex-drawing-guide
description: "TikZ and PGFPlots techniques for publication-quality scientific figures"
metadata:
openclaw:
emoji: "🎨"
category: "writing"
subcategory: "latex"
keywords: ["LaTeX typesetting", "LaTeX figure insertion", "LaTeX custom style", "scientific figure creation"]
source: "https://github.com/xinychen/awesome-latex-drawing"
---
# LaTeX Drawing Guide
## Overview
Publication-quality figures are a critical component of scientific papers. While external tools like matplotlib or Inkscape can produce good results, drawing figures directly in LaTeX using TikZ and PGFPlots offers unique advantages: figures share the same fonts and styling as the document, scale perfectly at any resolution, and remain fully version-controllable as plain text.
This guide draws from the awesome-latex-drawing repository (2,000+ stars), which provides 30+ complete examples of LaTeX-drawn figures covering Bayesian networks, neural network architectures, function plots, tensor diagrams, and machine learning frameworks. The techniques here apply broadly to any discipline that needs diagrams, flowcharts, or data plots embedded in LaTeX documents.
Learning TikZ has a steep initial curve, but the investment pays off substantially for researchers who publish frequently. Once you build a library of reusable components, creating new figures becomes fast and consistent.
## TikZ Fundamentals
### Basic Setup
```latex
\usepackage{tikz}
\usetikzlibrary{arrows.meta, positioning, calc, shapes.geometric, fit}
```
### Coordinate System and Basic Shapes
```latex
\begin{tikzpicture}
% Rectangle
\draw[fill=blue!20, rounded corners] (0,0) rectangle (3,2);
% Circle
\draw[fill=red!20] (5,1) circle (1cm);
% Arrow
\draw[-{Stealth[length=3mm]}, thick] (3.2,1) -- (3.8,1);
% Text node
\node at (1.5,1) {Input};
\node at (5,1) {Output};
\end{tikzpicture}
```
### Node-Based Diagrams
Nodes are the building blocks of most scientific diagrams:
```latex
\begin{tikzpicture}[
block/.style={
rectangle, draw, fill=blue!10,
minimum width=2.5cm, minimum height=1cm,
rounded corners, font=\small
},
arrow/.style={-{Stealth[length=2.5mm]}, thick}
]
\node[block] (input) {Data Input};
\node[block, right=2cm of input] (process) {Processing};
\node[block, right=2cm of process] (output) {Results};
\draw[arrow] (input) -- (process);
\draw[arrow] (process) -- (output);
\end{tikzpicture}
```
## Neural Network Diagrams
### Fully Connected Layer
```latex
\begin{tikzpicture}[
neuron/.style={circle, draw, fill=orange!30, minimum size=8mm},
conn/.style={->, gray!70}
]
% Input layer
\foreach \i in {1,...,3}
\node[neuron] (I\i) at (0, -\i*1.2) {$x_{\i}$};
% Hidden layer
\foreach \j in {1,...,4}
\node[neuron, fill=blue!20] (H\j) at (3, -\j*1.2+0.6) {$h_{\j}$};
% Output layer
\foreach \k in {1,...,2}
\node[neuron, fill=green!20] (O\k) at (6, -\k*1.2-0.6) {$y_{\k}$};
% Connections
\foreach \i in {1,...,3}
\foreach \j in {1,...,4}
\draw[conn] (I\i) -- (H\j);
\foreach \j in {1,...,4}
\foreach \k in {1,...,2}
\draw[conn] (H\j) -- (O\k);
% Labels
\node[above=0.3cm of I1] {\small Input};
\node[above=0.3cm of H1] {\small Hidden};
\node[above=0.3cm of O1] {\small Output};
\end{tikzpicture}
```
### Transformer Block
```latex
\begin{tikzpicture}[
block/.style={rectangle, draw, rounded corners, minimum width=3cm,
minimum height=0.8cm, fill=#1, font=\small},
block/.default=gray!10,
arr/.style={-{Stealth}, thick}
]
\node[block=yellow!20] (attn) at (0,0) {Multi-Head Attention};
\node[block=blue!10] (norm1) at (0,1.3) {Add \& LayerNorm};
\node[block=green!20] (ffn) at (0,2.6) {Feed-Forward Network};
\node[block=blue!10] (norm2) at (0,3.9) {Add \& LayerNorm};
\draw[arr] (attn) -- (norm1);
\draw[arr] (norm1) -- (ffn);
\draw[arr] (ffn) -- (norm2);
% Residual connections
\draw[arr, dashed, gray] (attn.west) -- ++(-0.8,0) |- (norm1.west);
\draw[arr, dashed, gray] (ffn.west) -- ++(-0.8,0) |- (norm2.west);
\end{tikzpicture}
```
## PGFPlots for Data Visualization
### Setup
```latex
\usepackage{pgfplots}
\pgfplotsset{compat=1.18}
```
### Line Plot with Error Bars
```latex
\begin{tikzpicture}
\begin{axis}[
width=0.8\textwidth,
height=6cm,
xlabel={Epoch},
ylabel={Accuracy (\%)},
legend pos=south east,
grid=major,
grid style={gray!30},
tick label style={font=\small}
]
\addplot+[mark=o, thick, error bars/.cd, y dir=both, y explicit]
coordinates {
(1,72) +- (0,1.5)
(5,85) +- (0,1.2)
(10,91) +- (0,0.8)
(20,94) +- (0,0.5)
(50,96) +- (0,0.3)
};
\addlegendentry{Our Method}
\addplot+[mark=square, thick, dashed]
coordinates {(1,68) (5,79) (10,85) (20,89) (50,91)};
\addlegendentry{Baseline}
\end{axis}
\end{tikzpicture}
```
### Bar Chart Comparing Methods
```latex
\begin{tikzpicture}
\begin{axis}[
ybar,
width=10cm, height=6cm,
symbolic x coords={BLEU, ROUGE-L, METEOR},
xtick=data,
ylabel={Score},
ymin=0, ymax=100,
bar width=12pt,
legend style={at={(0.5,1.05)}, anchor=south, legend columns=3},
nodes near coords,
nodes near coords style={font=\tiny}
]
\addplot coordinates {(BLEU,45.2) (ROUGE-L,62.1) (METEOR,38.7)};
\addplot coordinates {(BLEU,52.8) (ROUGE-L,68.4) (METEOR,44.3)};
\addplot coordinates {(BLEU,58.1) (ROUGE-L,71.9) (METEOR,49.6)};
\legend{Baseline, +Pretraining, +Fine-tuning}
\end{axis}
\end{tikzpicture}
```
## Bayesian Network and Graphical Models
```latex
\begin{tikzpicture}[
latent/.style={circle, draw, minimum size=1cm, fill=gray!20},
observed/.style={circle, draw, minimum size=1cm, fill=white, thick},
plate/.style={rectangle, draw, dashed, rounded corners, inner sep=10pt},
arr/.style={-{Stealth}, thick}
]
\node[latent] (theta) at (0,2) {$\theta$};
\node[latent] (z) at (2,2) {$z_n$};
\node[observed] (x) at (2,0) {$x_n$};
\node[latent] (alpha) at (-1.5,2) {$\alpha$};
\draw[arr] (alpha) -- (theta);
\draw[arr] (theta) -- (z);
\draw[arr] (z) -- (x);
\node[plate, fit=(z)(x), label=below right:$N$] {};
\end{tikzpicture}
```
## Best Practices
- **Define styles globally.** Use `\tikzset{}` in the preamble so all figures share consistent colors and shapes.
- **Use relative positioning.** `right=2cm of nodeA` is more maintainable than absolute coordinates.
- **Externalize figures.** For large documents, use `\usetikzlibrary{external}` to cache compiled figures and speed up builds.
- **Match document fonts.** TikZ inherits the document font automatically -- this is a key advantage over external tools.
- **Export standalone figures.** Use the `standalone` document class to compile figures individually for reuse in presentations.
- **Keep source readable.** One node or drawing command per line, with comments explaining the visual structure.
## References
- [awesome-latex-drawing](https://github.com/xinychen/awesome-latex-drawing) -- 30+ LaTeX drawing examples (2,000+ stars)
- [TikZ and PGF Manual](https://tikz.dev/) -- Official documentation
- [PGFPlots Manual](https://pgfplots.net/) -- Data visualization in LaTeX
- [TikZ Examples](https://texample.net/tikz/examples/) -- Community gallery
- [LaTeX Drawing Tutorial](https://www.overleaf.com/learn/latex/TikZ_package) -- Overleaf tutorial
版本历史
-
v2026.07
GitHub ↗
原版更新说明
## Highlights - **First tagged release.** Every gate is green: `make check` (validate + 183 unit tests + eval-harness + numeric benchmarks), validate-catalog, quality-evals, and OpenSSF Scorecard on `main`. - **Rigor double loop at 13 method families.** Each family closes taxonomy tag → eval scenario (does the agent avoid the method's known traps?) → deterministic numeric benchmark (does it recover the true effect?). New this cycle: heterogeneous treatment effects (CATE) and quantile treatment effects (QTE). - **Paper-WorkFlow flagship at 29/29 executable rigor gates**, now merged to its `main` with a drift-gated README badge — the credibility artifact is checked by CI, so it can never silently go stale ([RIGOR.md](https://github.com/brycewang-stanford/Paper-WorkFlow/blob/main/RIGOR.md)). - **1,150 vendored skills / 69 collections**, all license-audited, security-scanned, tracked in a machine-readable catalog with CI-enforced freshness, and importable as a single whole-repo router skill. - **Hardened weekly upstream syncs**: synced trees must pass the full gate before merge; this cycle both sync PRs were verified, patched (StatsPAI's SkillOpt execution-gate card restored), and merged. - **First-party plugin marketplace** (Python / R / Stata empirical-analysis plugins + guide) shipped in the previous cycle, now under a tagged version. ## Known Follow-Ups - Ecosystem benchmark comparison (AERS vs econometrics agents on the shared benchmark suite) — [PLAN-2026-07 week 3](https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/main/docs/PLAN-2026-07.md). - Decomposition (Oaxaca–Blinder) and shift-share (Bartik) method families (weeks 3–4). - GitHub Pages catalog site, README demo screenshots, five-language README consistency script (week 4). ## Catalog - Top-level collections: **69** - `SKILL.md` files: **1150** - Cataloged agent tools: **335** ([`tools/CATALOG.md`](../tools/CATALOG.md)) ## Methodological rigor - Method families in the coverage map: **13** ([`RIGOR_COVERAGE.md`](RIGOR_COVERAGE.md)) - Eval scenarios: **26** (138 rubric items, 122 auto-checkable) ([`../eval-harness/`](../eval-harness/README.md)) - Benchmark tasks: **13** (70 gold checks, 58 required) ([`../benchmark/`](../benchmark/README.md)) ### Method families (skills tagged) - `iv`: 27 - `rdd`: 25 - `panel-fe`: 19 - `staggered-did`: 17 - `bayesian`: 13 - `did`: 10 - `event-study`: 10 - `matching`: 10 - `synthetic-control`: 10 - `cate`: 8 - `dml`: 8 - `survival`: 3 - `quantile`: 2 ### Eval scenarios by category - `causal-identification`: 14 - `citation-hygiene`: 1 - `reproducibility`: 2 - `research-integrity`: 1 - `runtime-safety`: 1 - `statistical-validity`: 2 - `writing-compliance`: 3 - `writing-style`: 2 ## License buckets - AGPL-3.0: 1 - Apache-2.0: 1 - CC-BY-NC-4.0 (non-commercial): 1 - CC-BY-SA-4.0: 1 - CC-BY-SA-4.0 (repository default): 4 - GPL-3.0: 1 - MIT: 35 - MIT Non-Commercial: 1 - Mixed (per-skill; see folder): 2 - UNKNOWN - check upstream: 22 ## Quality gates - `make check` — catalog/provenance/audit/eval freshness, repo hygiene, Paper-WorkFlow gates, stdlib unit tests, eval-harness lint + graded smoke, and the strict numeric benchmark (reference-gold drift fails the build). - External links and OpenSSF Scorecard run on their scheduled workflows.
当前收录版本
0.0.1
质量评测
满分 100。前四项来自 DeepSkill 收录的质量报告。
评分依据
提供出版级科学图表绘制指南,解决研究人员核心需求
评分依据
关键词精准覆盖 LaTeX 排版、图表插入、科学图表创建等核心场景
评分依据
结构清晰,分模块讲解 TikZ 基础、神经网络、PGFPlots 等
评分依据
代码示例完整且可直接使用
计算方式
取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。
权限与使用前检查
根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。
未从主文档识别到上述常见操作,不能据此判断无需权限或没有风险。
文档引用的域名或地址
这些地址可能用于文档链接或示例,不是已批准的联网清单。使用前核对实际目的地及发送的数据。
- github.com
- pgfplots.net
- texample.net
- tikz.dev
- www.overleaf.com
安全评级与访问范围
risk · 静态证据核验,不代表运行安全保证。
使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。
用户讨论
还没有已核实的点评。