Scientific Diagrams Collection

The janosh/diagrams repository contains a collection of high-quality scientific diagrams. These diagrams are available in multiple formats, including:

• Web View: Interactive or static views hosted on GitHub Pages.
• LaTeX: Source files (.tex) for LaTeX users.
• Typst: Source files (.typ) for Typst users.
• PNG: Pre-rendered image assets.

Each diagram in the collection covers various scientific and mathematical topics such as 2D Convolution, Amplitude Modulation, Autoencoders, Bloch Spheres, and more. You can find the source files for a specific diagram within its corresponding directory in the assets/ folder.

The janosh/diagrams repository contains a collection of scientific diagrams available in multiple formats. Most diagrams are provided as both LaTeX (.tex) and Typst (.typ) source files, allowing users to use them in their own scientific documents.

Each diagram entry in the collection typically includes:

• A hosted web view of the diagram.
• Source files for LaTeX.
• Source files for Typst.
• A rendered PNG image for quick previewing.

Examples of available diagrams include:

• Machine Learning: Random Forest, Self Attention, Skip Connection, Supervised Learning Setup, Train-Test Split.
• Physics/Quantum Mechanics: Cost vs Accuracy in Quantum Mechanical Simulations, RG Flow in Theory Space, Sabatier Principle, Spontaneous Magnetization, Thermodynamic Ensemble transforms.
• Mathematics/Geometry: Sierpinski Triangle, Tori, Torus, Torus Fundamental Domain.

The repository contains a collection of scientific diagrams available in multiple formats, including web views, LaTeX (.tex), and Typst (.typ). Each diagram is accompanied by a PNG preview.

Available diagrams include:

• Variational Autoencoder
• Web Graph
• Wetterich Equation
• Wyckoff Positions
• zT vs n

To contribute a new diagram to the collection, follow these steps:

1. Create a .tex (for TikZ) or .typ (for Typst/CeTZ) file containing your diagram.
2. Create a corresponding metadata .yml file for the diagram.
3. Place both files in the assets/ directory.
4. Add your name to the citation.cff file to ensure proper attribution.
5. Submit your changes via a Pull Request.

To run the development server for this SvelteKit-based site, clone the repository, install the dependencies in the site directory, and start the dev server. The site supports hot-reloading; editing files in src/ will automatically update the browser.

git clone https://github.com/janosh/diagrams
cd diagrams/site
npm install
npm run dev

To use pdf-compressor directly or as part of the render-(typst|tikz).py pipeline, you must provide a free public API key from iLovePDF. Use the --set-api-key flag to configure it.

pdf-compressor --set-api-key project_public_7c854a9db0...

To generate an optimized production build and preview it locally, use the build and preview scripts via npm.

npm run build && npm run preview

The scripts in /scripts convert .typ and .tex files from /assets into PNG (low/high-res), PDF, and SVG formats. To use these scripts, you must install the following dependencies:

• pdf-compressor: Install via pip install pdf-compressor.
• gs (GhostScript): Optional; provides compression without an API key.
• pdf2svg: Install via brew install pdf2svg.
• convert: Part of ImageMagick.
• pngquant: Install via brew install pngquant.
• zopflipng: Install via brew install zopfli.

If you use these diagrams in your work, cite the Zenodo record using the following BibTeX entry:

@software{riebesell_diagrams_2020,
  title = {Collection of scientific diagrams},
  author = {Riebesell, Janosh and Bringuier, Stefan},
  date = {2020-08-09},
  year = {2020},
  doi = {10.5281/zenodo.7486911},
  url = {https://github.com/janosh/diagrams},
  note = {10.5281/zenodo.7486911 - https://github.com/janosh/diagrams},
  version = {0.2.0},
  urldate = {2023-01-01}, % optional, replace with your date of access,
}