In this repository I intend to keep track of some algorithms I've been learning throw the past year. The code consists on a set of algorithms written in p5.js (Processing), and three cryptography demos written in Python.
There is no server and no build step. The pages are plain HTML, the sketches are p5.js, and the cryptography is the original Python, running in the browser under Brython.
Any static file server will do:
python -m http.server 8000
Then open http://localhost:8000.
The site is plain files, so it can be served from anywhere. On GitHub Pages,
under Settings → Pages, set the source to Deploy from a branch and pick the
branch with / (root) as the folder. There is nothing to build.
The empty .nojekyll file at the root is load bearing. Without it GitHub runs
Jekyll over the repository first, and Jekyll drops two things this site needs:
vendor/, which is in its default exclude list, and the __init__.py files,
because it skips anything whose name starts with an underscore. The result
would be a site with no Bootstrap, no p5.js, no Brython, and no working
cryptography pages.
Links are relative throughout, so the site works both at a domain root and
under https://<user>.github.io/<repo>/.
Opening index.html from the filesystem works for the p5.js pages, but not for
the three cryptography ones: Brython fetches the Python modules over HTTP, which
file:// does not allow.
projects/ is the interesting part: the algorithms themselves, grouped the same
way the site groups them. The HTML beside it is mostly the same boilerplate
thirteen times over, and exists to put them on screen.
projects/algorithms mazes, path finding, regression, sorting, TSP (p5.js)
projects/ai flappy bird, rockets, steering, XOR (p5.js)
projects/cryptography AES, SHA-1, RSA (python)
index.html the menu
algorithms/ ai/ cryptography/ one page per project, grouped as above
css/style.css the stylesheet
vendor/p5 p5.js
vendor/bootstrap bootstrap, and jquery and popper beside it
vendor/brython the Brython runtime
tools/test_crypto.py known-answer tests for the cryptography
tools/check_heads.py checks the pages have not drifted apart
projects/cryptography is hand-written and has no dependencies, so it runs
under CPython as well as in the browser. tools/test_crypto.py checks it against
known answers: AES against the FIPS-197 vector, SHA-1 against hashlib, and RSA
round trips including messages longer than the modulus.
python tools/test_crypto.py
Run it after touching anything under projects/cryptography.
The asymmetric page keeps its keypairs and conversation in localStorage, so a
reload picks up where it left off. Regenerate clears them.
None of this is meant to secure anything. The keys are small enough to display on the page, and the interesting part is watching the algorithms work.
Put the code under projects/, copy the closest existing page, change
the title and the script it loads, and add a link to it in index.html. There is
no template to expand and nothing to rebuild.
Every page carries its own copy of the same <head>, since there is no template
engine to share one. If you change it, change it everywhere and confirm with:
python tools/check_heads.py