This repository contains a template for blueprint-driven formalization projects in Lean 4.
This is the video recording of the tutorial talk I presented at the Hausdorff Research Institute for Mathematics (HIM) in Bonn.
It was designed for mathematicians at all levels to provide a comprehensive introduction to the design, management, and implementation of blueprint-driven formalisation projects in Lean, with almost no prerequisite knowledge of Git, GitHub, continuous integration systems, and other technical tools.
Ensure that you have a functioning Lean 4 installation. If you do not, please follow the Lean installation guide.
To create a new repository using this template, ensure you are on the correct repository page (LeanProject) and then follow these steps:
- Click the Use this template button located at the top right of the repository page.
- Click the Create a new repository button.
- Select the account or organization where you want to create it, choose a name for the new repository, and click the Create repository button.
To clone this repository to your local machine, please refer to the relevant section of the GitHub documentation here.
To tailor this template to your specific project, follow these steps:
- If you don't have a Python environment, you can install one by following the instructions in the Python installation guide.
- Verify your Python installation by running:
python3 --version
- Verify your Pip installation by running:
pip3 --version
- Ensure your terminal is in the project directory by running the following command:
cd path/to/your/project
- Execute the customization script by running:
where
scripts/customize_template.py NewProject
NewProject
must be replaced by the name of your project.
The script customize_template.py
will automatically rename the
project folder and update the necessary files and configurations to match the new project name.
To set up GitHub Pages for your repository, follow these steps:
- Go to the Settings tab of your repository.
- In the left sidebar, click on the Pages section.
- In the Source dropdown, select
GitHub Actions
.
The template repository is organized as follows (listing the main folders and files):
.github
contains GitHub-specific configuration files and workflows.workflows
contains GitHub Actions workflow files.build-project.yml
defines the workflow for building the Lean project on pushes, pull requests, and manual triggers. This is a minimalistic build workflow which is not necessary if you decide to generate a blueprint (see instructions below) and can be manually disabled by clicking on the Actions tab, selecting Build Project in the left sidebar, then clicking the horizontal triple dots (⋯) on the right, and choosing Disable workflow.create-release.yml
: defines the workflow for creating a new Git tag and GitHub release when thelean-toolchain
file is updated in themain
branch.update-dependencies.yml
is the dependency update workflow to be triggered manually by default. [It's not documented yet, but it will be soon.]
dependabot.yml
is the configuration file to automate CI dependency updates.
.vscode
contains Visual Studio Code configuration filesextensions.json
recommends VS Code extensions for the project.settings.json
defines the project-specific settings for VS Code.
Project
should contain the Lean code files.Mathlib
should contain.lean
files with declarations missing from existing Mathlib developments.ForMathlib
should contain.lean
files with new declarations to be upstreamed to Mathlib.Example.lean
is a sample Lean file.
scripts
contains scripts to update Mathlib ensuring that the latest version is fetched and integrated into the development environment..gitignore
specifies files and folders to be ignored by Git. and environment.CODE_OF_CONDUCT.md
should contain the code of conduct for the project.CONTRIBUTING.md
should provide the guidelines for contributing to the project.lakefile.toml
is the configuration file for the Lake build system used in Lean projects.lean-toolchain
specifies the Lean version and toolchain used for the project.
- Fermat's Last Theorem for Exponent 3 by Riccardo Brasca et al.
- Polynomial Freiman-Ruzsa Conjecture by Terence Tao et al.
- Fermat's Last Theorem by Kevin Buzzard et al.
- Carleson Operators on Doubling Metric Measure Spaces by Floris van Doorn et al.
- Bonn Collaborative Formalization Seminar Series in Analysis by Floris van Doorn et al.
- Prime Number Theorem and More by Alex Kontorovich et al.
- Infinity Cosmos by Emily Riehl et al.
- Analytic Number Theory Exponent Database by Terence Tao et al.
- Groupoid Model of Homotopy Type Theory by Sina Hazratpour et al.
- Equational Theories by Terence Tao et al.
For more examples of completed and ongoing Lean projects and libraries, please see the Lean Reservoir.
To install the necessary dependencies, follow the instructions in the PyGraphViz installation guide.
Assuming you have a properly configured Python environment, install LeanBlueprint by running:
pip install leanblueprint
If you have an existing installation of LeanBlueprint, you can upgrade to the latest version by running:
pip install -U leanblueprint
To set up the blueprint for your project, run:
leanblueprint new
Then, follow the prompts and answer the questions as you like, except for a few specific questions which should be answered as indicated below to ensure compatibility with this template.
Respond affirmatively with y
to the following prompt:
Proceed with blueprint creation? [y/n]
Respond affirmatively with y
to the following prompt:
Modify lakefile and lake-manifest to allow checking declarations exist? [y/n] (y)
Respond affirmatively with y
to the following prompt:
Modify lakefile and lake-manifest to allow building the documentation? [y/n] (y):
If you want to generate a Jekyll-based home page for the project, respond
affirmatively with y
to the following prompt:
Do you want to create a home page for the project, with links to the blueprint, the API documentation and the repository? [y/n]:
Respond affirmatively with y
to the following prompt:
Configure continuous integration to compile blueprint? [y/n] (y):
For more details about the LeanBlueprint package and its commands, please refer to its documentation.
After configuring the blueprint, please wait for the GitHub Action workflow to finish. You can keep track of the progress in the Actions tab of your repository.
If you have used this template to create your own Lean project and would like to share it with the community, please consider opening a PR to add your project to this list:
- Infinity Cosmos by Emily Riehl et al.
- Analytic Number Theory Exponent Database by Terence Tao et al.
- Equational Theories led by Terence Tao.
- Groupoid Model of Homotopy Type Theory by Sina Hazratpour et al.
- Soundness of FRI by Bolton Bailey et al.
- Weil's Converse Theorem by Chris Birkbeck et al.
- Proofs from THE BOOK by Moritz Firsching et al.
- Automata Theory by Stefan Hetzl et al.
- Dirichlet Nonvanishing by Chris Birkbeck et al.
- Seymour's Decomposition Theorem by Ivan Sergeyev et al.