Add support for generating scientific illustrations using Nano Banan Pro and Flux.2 Pro

2026-01-26 16:58:56 +08:00 · 2025-11-30 09:30:58 -05:00
parent 8d82c83a1a
commit 90de96a99b
6 changed files with 561 additions and 6 deletions
--- a/docs/examples.md
+++ b/docs/examples.md
@@ -25,6 +25,7 @@ This document provides comprehensive, practical examples demonstrating how to co
 17. [Clinical Research & Real-World Evidence](#clinical-research--real-world-evidence)
 18. [Experimental Physics & Data Analysis](#experimental-physics--data-analysis)
 19. [Chemical Engineering & Process Optimization](#chemical-engineering--process-optimization)
+20. [Scientific Illustration & Visual Communication](#scientific-illustration--visual-communication)

 ---

@@ -2490,6 +2491,152 @@ Expected Output:

 ---

+## Scientific Illustration & Visual Communication
+
+### Example 20: Creating Publication-Ready Scientific Figures
+
+**Objective**: Generate and refine scientific illustrations, diagrams, and graphical abstracts for publications and presentations.
+
+**Skills Used**:
+- `generate-image` - AI image generation and editing
+- `matplotlib` - Data visualization
+- `scientific-visualization` - Best practices
+- `scientific-writing` - Figure caption creation
+- `reportlab` - PDF report generation
+
+**Workflow**:
+
+```bash
+Step 1: Plan visual communication strategy
+- Identify key concepts that need visual representation:
+  * Experimental workflow diagrams
+  * Molecular structures and interactions
+  * Data visualization (handled by matplotlib)
+  * Conceptual illustrations for mechanisms
+  * Graphical abstract for paper summary
+- Determine appropriate style for target journal/audience
+- Sketch rough layouts for each figure
+
+Step 2: Generate experimental workflow diagram
+- Use generate-image skill with detailed prompt:
+  "Scientific illustration showing a step-by-step experimental 
+  workflow for CRISPR gene editing: (1) guide RNA design at computer,
+  (2) cell culture in petri dish, (3) electroporation device,
+  (4) selection with antibiotics, (5) sequencing validation.
+  Clean, professional style with numbered steps, white background,
+  suitable for scientific publication."
+- Save as workflow_diagram.png
+- Review and iterate on prompt if needed
+
+Step 3: Create molecular interaction schematic
+- Generate detailed molecular visualization:
+  "Scientific diagram of protein-ligand binding mechanism:
+  show receptor protein (blue ribbon structure) with binding pocket,
+  small molecule ligand (ball-and-stick, orange) approaching,
+  key hydrogen bonds indicated with dashed lines, water molecules
+  in binding site. Professional biochemistry illustration style,
+  clean white background, publication quality."
+- Generate multiple versions with different angles/styles
+- Select best representation
+
+Step 4: Edit existing figures for consistency
+- Load existing figure that needs modification:
+  python scripts/generate_image.py "Change the background to white
+  and make the protein blue instead of green" --input figure1.png
+- Standardize color schemes across all figures
+- Edit to match journal style guidelines:
+  python scripts/generate_image.py "Remove the title text and
+  increase contrast for print publication" --input diagram.png
+
+Step 5: Generate graphical abstract
+- Create comprehensive visual summary:
+  "Graphical abstract for cancer immunotherapy paper: left side
+  shows tumor cells (irregular shapes, red) being attacked by
+  T cells (round, blue). Center shows the drug molecule structure.
+  Right side shows healthy tissue (green). Arrow flow from left
+  to right indicating treatment progression. Modern, clean style
+  with minimal text, high contrast, suitable for journal TOC."
+- Ensure dimensions meet journal requirements
+- Iterate to highlight key findings
+
+Step 6: Create conceptual mechanism illustrations
+- Generate mechanism diagrams:
+  "Scientific illustration of enzyme catalysis mechanism:
+  Show substrate entering active site (step 1), transition state
+  formation with electron movement arrows (step 2), product
+  release (step 3). Use standard biochemistry notation,
+  curved arrows for electron movement, clear labeling."
+- Generate alternative representations for supplementary materials
+
+Step 7: Produce presentation-ready figures
+- Create high-impact visuals for talks:
+  "Eye-catching scientific illustration of DNA double helix
+  unwinding during replication, with DNA polymerase (large
+  green structure) adding nucleotides. Dynamic composition,
+  vibrant but professional colors, dark background for
+  presentation slides."
+- Adjust style for poster vs slide format
+- Create versions at different resolutions
+
+Step 8: Generate figure panels for multi-part figures
+- Create consistent series of related images:
+  "Panel A: Normal cell with intact membrane (green outline)
+  Panel B: Cell under oxidative stress with damaged membrane
+  Panel C: Cell treated with antioxidant, membrane recovering
+  Consistent style across all panels, same scale, white background,
+  scientific illustration style suitable for publication."
+- Ensure visual consistency across panels
+- Annotate with panel labels
+
+Step 9: Edit for accessibility
+- Modify figures for colorblind accessibility:
+  python scripts/generate_image.py "Change the red and green
+  elements to blue and orange for colorblind accessibility,
+  maintain all other aspects" --input figure_v1.png
+- Add patterns or textures for additional differentiation
+- Verify contrast meets accessibility standards
+
+Step 10: Create supplementary visual materials
+- Generate additional context figures:
+  "Anatomical diagram showing location of pancreatic islets
+  within the pancreas, cross-section view with labeled structures:
+  alpha cells, beta cells, blood vessels. Medical illustration
+  style, educational, suitable for supplementary materials."
+- Create protocol flowcharts and decision trees
+- Generate equipment setup diagrams
+
+Step 11: Compile figure legends and captions
+- Use scientific-writing skill to create descriptions:
+  * Figure number and title
+  * Detailed description of what is shown
+  * Explanation of symbols, colors, and abbreviations
+  * Scale bars and measurement units
+  * Statistical information if applicable
+- Format according to journal guidelines
+
+Step 12: Assemble final publication package
+- Organize all figures in publication order
+- Create high-resolution exports (300+ DPI for print)
+- Generate both RGB (web) and CMYK (print) versions
+- Compile into PDF using ReportLab:
+  * Title page with graphical abstract
+  * All figures with captions
+  * Supplementary figures section
+- Create separate folder with individual figure files
+- Document all generation prompts for reproducibility
+
+Expected Output:
+- Complete set of publication-ready scientific illustrations
+- Graphical abstract for table of contents
+- Mechanism diagrams and workflow figures
+- Edited versions meeting journal style guidelines
+- Accessibility-compliant figure versions
+- Figure package with captions and metadata
+- Documentation of prompts used for reproducibility
+```
+
+---
+
 ## Summary

 These examples demonstrate:
--- a/docs/scientific-skills.md
+++ b/docs/scientific-skills.md
@@ -154,6 +154,7 @@
 - **HypoGeniC** - Automated hypothesis generation and testing using large language models to accelerate scientific discovery. Provides three frameworks: HypoGeniC (data-driven hypothesis generation from observational data), HypoRefine (synergistic approach combining literature insights with empirical patterns through an agentic system), and Union methods (mechanistic combination of literature and data-driven hypotheses). Features iterative refinement that improves hypotheses by learning from challenging examples, Redis caching for API cost reduction, and customizable YAML-based prompt templates. Includes command-line tools for generation (hypogenic_generation) and testing (hypogenic_inference). Research applications have demonstrated 14.19% accuracy improvement in AI-content detection and 7.44% in deception detection. Use cases: deception detection in reviews, AI-generated content identification, mental stress detection, exploratory research without existing literature, hypothesis-driven analysis in novel domains, and systematic exploration of competing explanations

 ### Scientific Communication & Publishing
+- **Generate Image** - AI-powered image generation and editing for scientific illustrations, schematics, and visualizations using OpenRouter's image generation models. Supports multiple models including google/gemini-3-pro-image-preview (high quality, recommended default) and black-forest-labs/flux.2-pro (fast, high quality). Key features include: text-to-image generation from detailed prompts, image editing capabilities (modify existing images with natural language instructions), automatic base64 encoding/decoding, PNG output with configurable paths, and comprehensive error handling. Requires OpenRouter API key (via .env file or environment variable). Use cases: generating scientific diagrams and illustrations, creating publication-quality figures, editing existing images (changing colors, adding elements, removing backgrounds), producing schematics for papers and presentations, visualizing experimental setups, creating graphical abstracts, and generating conceptual illustrations for scientific communication
 - **Paper-2-Web** - Autonomous pipeline for transforming academic papers into multiple promotional formats using the Paper2All system. Converts LaTeX or PDF papers into: (1) Paper2Web - interactive, layout-aware academic homepages with responsive design, interactive figures, and mobile support; (2) Paper2Video - professional presentation videos with slides, narration, cursor movements, and optional talking-head generation using Hallo2; (3) Paper2Poster - print-ready conference posters with custom dimensions, professional layouts, and institution branding. Supports GPT-4/GPT-4.1 models, batch processing, QR code generation, multi-language content, and quality assessment metrics. Use cases: conference materials, video abstracts, preprint enhancement, research promotion, poster sessions, and academic website creation
 - **Perplexity Search** - AI-powered web search using Perplexity models via LiteLLM and OpenRouter for real-time, web-grounded answers with source citations. Provides access to multiple Perplexity models: Sonar Pro (general-purpose, best cost-quality balance), Sonar Pro Search (most advanced agentic search with multi-step reasoning), Sonar (cost-effective for simple queries), Sonar Reasoning Pro (advanced step-by-step analysis), and Sonar Reasoning (basic reasoning). Key features include: single OpenRouter API key setup (no separate Perplexity account), real-time access to current information beyond training data cutoff, comprehensive query design guidance (domain-specific patterns, time constraints, source preferences), cost optimization strategies with usage monitoring, programmatic and CLI interfaces, batch processing support, and integration with other scientific skills. Installation uses uv pip for LiteLLM, with detailed setup, troubleshooting, and security documentation. Use cases: finding recent scientific publications and research, conducting literature searches across domains, verifying facts with source citations, accessing current developments in any field, comparing technologies and approaches, performing domain-specific research (biomedical, clinical, technical), supplementing PubMed searches with real-time web results, and discovering latest developments post-database indexing