Skip to main content

lean4_jupyter: A Lean 4 Jupyter kernel via REPL

Project description

lean4_jupyter

A Lean 4 Jupyter kernel via repl.

PyPI PyPI - Python Version PyPI - License PyPI - Status Python CI lint - flake8 Static Badge Maintainability

What's already working

🔥 See it in action: Tutorial notebook.

The kernel can:

  • execute Lean 4 commands
  • execute Lean 4 tatics with magic like % proof immediately after a sorryed theorem
  • backtrack to earlier environment or proof states with magic like % env 1 or % prove 3
  • support magics like %cd or %load (loading a file)
  • support for importing modules from projects and their dependencies, e.g. Mathlib ( demo ).

Output:

  • In jupyter notebook and alike: echos the input annotated in alectryon style, at the corresponding line (not columns yet), with messages, proof states
  • In jupyter console and alike: echos the input annotated in codespan style, at the corresponding line:column, with messages, proof states

What's next

  • Fix repl#40 (PRed as repl#41)
  • Improve the alectryon annotation to support annotations in the middle of a line
  • Make magics like %cd and %load work more robustly
  • Support show tactics after %load
  • Add all repl test cases to the CI and set up coverage
  • Improve UI in Jupyter Dark themes
  • Minor code refactor to smooth things out

Installation

First, you need a working Lean 4 installation. You can install it via elan.

Verify that lean and lake is in your PATH:

lean --version
lake --help|head -n 1

they should output Lean/Lake versions, respectively.

Then, you need to have a working repl in your PATH.

You can build it from source (please read and adjust them before executing):

# If you need to clean up before reinstalling
# rm -rf ~/.lean4_jupyter/repl
# rm /usr/local/bin/repl

# Prepare a directory for lean4_jupyter where we install repl to
mkdir -p ~/.lean4_jupyter

# Before repl#41 merge, you might need to use this branch instead
git clone -b fix-dup https://github.com/utensil/repl ~/.lean4_jupyter/repl
# git clone https://github.com/leanprover-community/repl ~/.lean4_jupyter/repl

# Build repl
(cd ~/.lean4_jupyter/repl && lake build)

# Install repl to a place in your PATH, so less hassle of fiddling with PATH
ln -s ~/.lean4_jupyter/repl/.lake/build/bin/repl /usr/local/bin/repl

Verify that repl is working:

echo '{"cmd": "#eval Lean.versionString"}'|repl

In case repl hangs, you could kill it with

ps aux|grep repl|grep -v grep|awk '{print $2}'|xargs kill -9

Then, ensure that you have python, pip, jupyter installed, and run:

pip install jupyterlab

Then, install the kernel:

pip install lean4_jupyter
# or in development mode, check out the repo then run
# pip install -e .
python -m lean4_jupyter.install

To use it, run one of:

# Web UI
jupyter notebook
jupyter lab

# Console UI
# hint: use Ctrl-D and confirm with y to exit
jupyter console --kernel lean4
# hint: you need to run `pip install PyQt5 qtconsole` to install it
jupyter qtconsole --kernel lean4

Inspired by

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

lean4_jupyter-0.0.1rc8.tar.gz (29.5 kB view details)

Uploaded Source

Built Distribution

lean4_jupyter-0.0.1rc8-py2.py3-none-any.whl (15.1 kB view details)

Uploaded Python 2 Python 3

File details

Details for the file lean4_jupyter-0.0.1rc8.tar.gz.

File metadata

  • Download URL: lean4_jupyter-0.0.1rc8.tar.gz
  • Upload date:
  • Size: 29.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/5.0.0 CPython/3.12.3

File hashes

Hashes for lean4_jupyter-0.0.1rc8.tar.gz
Algorithm Hash digest
SHA256 28e9212ab4fe745e6d7d5950c8724374361898194a93dc236f46b84dae4e0c20
MD5 69e0874928dfe4b5d4edb3c4f4c1ef38
BLAKE2b-256 70d204562ff792ff27d504c2f30101512ce2335058df11e58a61b63fbf624d9e

See more details on using hashes here.

File details

Details for the file lean4_jupyter-0.0.1rc8-py2.py3-none-any.whl.

File metadata

File hashes

Hashes for lean4_jupyter-0.0.1rc8-py2.py3-none-any.whl
Algorithm Hash digest
SHA256 8d8ed38e952a431bf90f339957cdd29d051b833d8d32a926a2af062cb9007c11
MD5 1eba0b75b2b5127101d944fd70d0ad0a
BLAKE2b-256 bea028be6e132c56352cea14889b56444cfb4d1910b32f5545d9dd66c1296e80

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page