fluency-drill

Terminal practice tool for building muscle memory and fluency with the syntax and idioms of a programming language. Drill methods, work through progressively harder exercises, and resume mid-session.

npx fluency-drill

or install globally with your package manager of choice:

npm i -g fluency-drill        # or: pnpm add -g fluency-drill
fluency-drill

Requires Node.js ≥ 18. The Python curricula additionally need python3 on your PATH; the data-science / research curricula optionally need numpy, pandas, scipy, or matplotlib for their exercises (the drills work without them).

What it does

For each language, the game exposes a list of fluency entries — one per method, idiom, or quirk. For each entry you can:

1. Drill — type the entry's canonical syntax and 2–7 common variations, 10× per variation by default (5× in "quick" mode). Matching is fuzzy: whitespace, quote style, and trailing semicolons are normalized. Each variation shows an input and the resulting output, so you're not typing in the dark.

2. Exercises — a worked example, then progressively harder problems that end at "real-world repo usage". Your code is sandboxed, run, and compared to the expected output.

3. Both — drill, then exercises, in one go.

Progress saves per-curriculum after every successful rep. :quit mid-drill and resume from the same variation + rep number next session.

Languages and curricula

Out of the box:

• JavaScript — arrays, strings, objects, Map/Set, promises, JSON, quirks (sort default, == vs ===, closure capture, …), interview syntax drills (counter idioms, set-visited, grid init, bit tricks, binary search building blocks).

• Python — iterables / comprehensions, strings (f-strings), dicts, sequence slicing, collections (Counter / defaultdict / deque), itertools, functools, pathlib, datetime, json, asyncio.gather, quirks (mutable defaults, is vs ==, late-binding closures, …), interview syntax drills (Counter, set, grid init, bisect, heapq).

• Python · Data Science — numpy array basics; pandas DataFrame basics (filtering, new columns).

• Python · Data Engineering — stdlib csv (Reader / DictWriter); stdlib sqlite3 (parameterized queries, executemany, group-by).

• Python · Research — scipy.stats / scipy.optimize; matplotlib.pyplot.

The menu detects missing prerequisites and gracefully skips exercises that need an uninstalled library (drills always work).

In-game commands

In drill:

• :show peek the target syntax for the current variation
• :skip move on to the next variation
• :quit exit (state is saved — resume next session)

In exercises:

• :hint show the entry's hint, if any
• :solve reveal the expected output and move on
• :skip skip this exercise
• :quit exit

In menus:

• R reset all progress for the current language
• x (submenu) reset progress for one method
• L switch language

Adding a curriculum

Each language is one file in src/curriculum/. Drop a file and add an import in src/curriculum/index.js — the menu auto-detects. See CLAUDE.md for the full data shape and the requires schema for gating exercises behind installable prerequisites.

License

MIT