Python Lab
Python Lab is a free static study studio for Python. It is meant to stay useful from a first print through interview practice. The pages are HTML, CSS, and JavaScript. There is no build step. GitHub Pages can serve the repository root as-is.
Lucky, a lifelong student and an AI enthusiast moving into AI omni-channel agents, shares what he is learning. Facts are cited in the Sources section. The lab does not estimate pay.
Run it locally
python3 -m http.server 41731 --bind 127.0.0.1
Open http://127.0.0.1:41731
What’s in the page
- Lessons from syntax through objects, tests, async, typing, the standard library, Flask and FastAPI, NumPy and pandas, models and agents, automation, tooling, licensing, certifications, career study order, and projects.
- A browser Python runtime: Pyodide 314.0.7, which is CPython 3.14.2, plus vendored NumPy 2.4.6 and pandas 3.0.2. Python 3.15.0 (9 October 2026) is newer than this runtime. Lessons say so when syntax is 3.15-only.
- Checks on the exercises, quizzes, and three games: guess, snake, and a text adventure.
- CSS 3D rooms, with a flat list when the browser asks for reduced motion.
- Diagrams drawn with vendored Mermaid, plus 8 interactive Archify showcase diagrams under
diagrams/ (specs in diagrams/specs/; see diagrams/NOTICE.md).
- 9 slide decks (92 slides, arrow keys or swipe, full screen, print) and 11 guides. Code samples that show an output were run on Pyodide 314 first.
- A first-visit how-to popup (Help button,
? key), a colourful helper bot, and the lab footer with the shared disclaimer.
- Read-aloud through the browser’s own speech synthesis. The voice depends on the device.
- Short original GIF, WebM, and MP4 clips under
media/.
- Cheat sheets in
downloads/ as PDF and plain text, the slides as Markdown, and a Jupyter notebook of code samples. Regenerate those two with node scripts/make_downloads.mjs.
Regenerate the clips with python3 scripts/make_media.py (needs ffmpeg). Regenerate the sheets with python3 scripts/make_pdfs.py (needs fpdf2).
Checks
The site itself does not need Node. The Playwright suite does:
npm install
npx playwright install chromium
npx playwright test
Test matrix
npx playwright test runs the end-to-end suite. python3 tests/matrix.py <base-url> <out-dir> (Python Playwright) runs the theme and viewport matrix: 4 themes x desktop 1440x900 and touch 390x844, covering every route, menu item, help, search, bot, footer and slides, and writes results.json plus screenshots.
Publish
The files are meant to be served as a directory, including under a longer path such as https://cyphermonkeylabs.com/python-lab-free-university-mesmerizing/. Links to CSS, scripts, clips, and downloads are relative, so that address needs its trailing slash. .nojekyll is included so GitHub Pages serves the runtime files. sw.js is registered with scope ./, which is this directory and not the rest of the host. No CDN is required at runtime.
3D studio, Py teacher, Downloads
- Worlds (
#/worlds, #/worlds/0 through #/worlds/10): each Level has its own three.js scene. The lobby keeps a shared still. three.js is vendored (vendor/three/, MIT) and loads only on pages with a scene. Tap or click a scene and it reacts. Reduced motion keeps the still picture.
- Quality tiers (Display menu): Auto, Off (still), Flat (CSS parallax), Light (WebGL, 30 fps cap, DPR 1.5), Full (60 fps cap, DPR 2, aurora). Auto honours reduced motion and Save-Data and runs a short GPU probe.
?quality=full overrides for testing. Read-only numbers: window.__labStudio.
- Py has a four-rung hint ladder per exercise (hand-written for 13 exercises, generic for the rest), Socratic prompts, and tap/drag/drop/win/fail reactions. No stored solutions, no AI service.
- Downloads. The page route is
#/downloads. #/download-center is the same page (top bar and footer use that alias). GitHub renders a markdown link whose target starts with #/ as an in-file heading anchor, so this README names both routes in code spans instead of a markdown link. A module appears when content/<slug>/module.md exists (Levels 0–10 in this tree). Rebuild with python3 pipeline/build_downloads.py --content content --out downloads/modules --pack-b64.
- Binary files. PDF, DOCX, PPTX and PNG are not committed as binaries.
pipeline/build_downloads.py --pack-b64 writes base64 parts under _b64/ (about 100000 characters each, named <path>.b64.001, .002, …) and refreshes the matching lines in SHA256SUMS. HTML, JSON and TXT stay as text. Locally, python3 pipeline/restore_binaries.py decodes the parts and keeps _b64. Pass --check to compare hashes without writing. The GitHub Action .github/workflows/restore-binaries.yml runs on a push to main that changes _b64/** (and on workflow_dispatch): it joins the parts, decodes them, deletes _b64, checks SHA256SUMS, and commits the binaries to main with message Restore binary assets [skip ci]. Do not push those binaries from a feature branch.
