Academic Figures Generator

Challenges in Academic Figure Preparation

Creating figures for scientific publications often involves meeting strict journal requirements, which can be time-consuming and error-prone. Key issues include:

• Complex Journal Standards: Journals like Nature and Lancet mandate specific dimensions (e.g., 89mm single-column, 183mm double-column), DPI (600+), fonts (Helvetica 7pt, minimum 5pt), and colorblind-safe palettes. Manual adjustment of matplotlib code to comply is inefficient and prone to errors, such as setting column widths or colors incorrectly.
• Lack of Data Validation: Invalid data structures can produce flawed outputs, delaying submissions. For example, mismatched series lengths or missing required keys may generate broken charts.
• Accessibility Concerns: Traditional color schemes may not be suitable for colorblind readers, reducing the paper's accessibility and potentially leading to journal rejections.

Core Capabilities and Workflow

This tool addresses these challenges with a single command, offering the following features:

• 14 Chart Types: Covers bar charts (-t bar), heatmaps (-t heatmap), forest plots (-t forest), KM survival curves (-t km), and more via the -t parameter, suitable for most academic analyses.
• 7 Color Schemes: Includes glm (default, elegant and colorblind-safe) and okabe-ito (Nature Methods gold standard), switchable with --theme. For instance, --theme okabe-ito ensures colorblind safety for journal submissions.
• Journal Presets: Parameters like --journal nature or --journal lancet automatically apply official column widths, fonts, and DPI (e.g., nature double-column at 183mm, 7pt font), eliminating manual configuration.
• Data Validation and Verification:
• Pre-run validate_data() checks data structure (e.g., box/violin labels must match group count), with fatal errors (e.g., empty series) returning exit code 1 without saving.
• For PDF output, --verify performs pixel-level overlap checks to prevent label obscuration.
• Chinese Language Support: --cjk auto-detects and loads Chinese fonts (e.g., Noto Sans CJK), recursively scanning data dictionary keys to avoid garbled text in bilingual charts.

Workflow: Users provide a JSON data file and run a command like python gen_figure.py -t bar --theme okabe-ito --journal nature -o figure.pdf. The tool executes data validation, color application, and layout adjustment, outputting a 600 DPI PDF that meets publication standards.

Important Considerations

• Data Format Requirements: Input must be JSON with required keys for the chart type. For example, bar charts need series (data arrays) and labels; error bars require the errors field.
• Journal Submission Tips: For submissions, use --theme okabe-ito and PDF output to comply with mainstream journal colorblind safety requirements. Avoid nature or lancet color schemes, as they are not colorblind-safe.
• CJK Font Issues: The system must have Chinese fonts installed (e.g., Noto Sans CJK); otherwise, run setup_env.py for automatic detection and cache clearing. Custom fonts can be specified with --cjk-font.
• Verification Process: Before delivering PDFs, run scripts/audit_pdf.py to check font sizes against journal minimums (e.g., nature 5pt) and use scripts/verify_overlap_pixel.py to verify no overlaps.
• Scope Limits: This tool is focused on academic figure generation and is not suitable for medical illustrations (e.g., SVG) or terminal CLI charts, which require specialized tools. For flowcharts (-t diagram), avoid white blocks to ensure visibility on white backgrounds.