nature-figure

verdict: PASS | 4/4 tasks passed | 2026-06-15

Overview

Nature-quality scientific figure generation with matplotlib/ggplot2 dual backend, figure contract methodology, and strict export standards.

Core Capabilities

• Publication-quality figures (150/300 DPI)
• Multi-panel layouts
• Nature-style formatting (Arial, no borders, 7pt)
• Dual DPI export (preview + publication)
• Figure contract workflow (claim → evidence → chart)
• Python + R dual backend

Test Results

All 4 tasks passed L3 deep testing across 4 disciplines:

Discipline	Task	Chart Type	Status
Medical	task-medical-drug-response	Dose-response curve	PASS
Physics	task-physics-spectrum	UV-Vis absorption spectrum	PASS
Chemistry	task-chemistry-catalyst-bar	Catalyst activity bar chart	PASS
Biology	task-biology-gene-heatmap	Gene expression heatmap	PASS

Limitations

• Arial font dependency (system installation required)
• Figure contract must be written manually
• No built-in interactive/dashboard mode
• Backend must be chosen before plotting