| name | pubfig |
| description | Publication-ready figure + LaTeX↔Excel table 切换。NeurIPS / CoRL / ICLR 模板内置。把 /auto-viz 的 PNG 升级成投稿质量的 PDF/SVG。Use when user says "pubfig", "publication figure", "投稿质量图", "latex table", "美化图", "出 PDF figure". |
| argument-hint | [--venue NeurIPS|CoRL|ICLR] [--input figures/loss.png] [--output paper/figures/loss.pdf] |
| allowed-tools | Bash(*), Read, Write |
/pubfig — Publication-Ready Figures + Tables
借鉴 claude-scholar pubfig + pubtab。
Overview
- Figure:matplotlib 用 venue-specific rcParams 重画(字体 / DPI / colorbar / linewidth)→ vector PDF
- Table:CSV/Markdown ↔ LaTeX (booktabs)
Venue presets:
| Venue | font | DPI | column width |
|---|
| NeurIPS | Times New Roman | 300 | 5.5 in (single) / 3.25 in (col) |
| CoRL | Times New Roman | 300 | 6 in / 3 in |
| ICLR | Computer Modern | 300 | 5.5 in |
| Nature | Helvetica | 600 | 89 mm / 183 mm |
Workflow
python3 tools/pubfig.py --venue $VENUE \
--input figures/loss.png \
--output paper/figures/loss.pdf
读 raw data 而非 PNG 重画(要求 figures/ 同名 .json data 文件由 /auto-viz 生成)。
Output
paper/figures/*.pdf vector
paper/tables/*.tex LaTeX
Composition
- 被
/paper-pipeline Stage 5 调
- 单独:「把 task_sr.png 出成 NeurIPS PDF」