Files

T

Lucas WintherandClaude Opus 5.5 68fe9834c3 feat: module 06 parser

First Part II module. The cumulative reference adds a second engine
(src/engine/puzzle: grid, lexer, parser) beside Part I's; nouns and
properties are data in src/content/words.json, and the engine names only
IS/AND/NOT/TEXT and the base rule TEXT IS PUSH. Rules are recompiled from
the board on every change; rules() is a sorted set. Adds the M6 scenarios
and tests (rows, columns, shared IS, AND distribution, stacked text, base
rule), the HOT word as a data-only Expansion Drop, the lesson with the
Part II bridge, and the Rust LevelDef shape in harness.md.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

2026-10-29 03:33:53 +01:00

src

feat: module 06 parser

2026-10-29 03:33:53 +01:00

.course.json

feat: module 06 parser

2026-10-29 03:33:53 +01:00

.gitignore

feat: module 06 parser

2026-10-29 03:33:53 +01:00

package-lock.json

feat: module 06 parser

2026-10-29 03:33:53 +01:00

package.json

feat: module 06 parser

2026-10-29 03:33:53 +01:00

README.md

feat: module 06 parser

2026-10-29 03:33:53 +01:00

tsconfig.json

feat: module 06 parser

2026-10-29 03:33:53 +01:00

vitest.config.ts

feat: module 06 parser

2026-10-29 03:33:53 +01:00

README.md

Module 6 reference engine (TypeScript)

The course's own engine as of the end of Module 6, the first module of Part II. Part I's card engine is unchanged, and all of its tests still pass. Beside it, in src/engine/puzzle/, is a second engine for Baba Is You: a grid of objects and text, a lexer and a parser. Module 5's reference was the starting point, so a diff between the two shows what this module adds.

cards/ (at the package root; not in this directory): Part I's card data, a copy of course/cards/ staged by tools/verify-references.mjs. Part II doesn't read it.
src/engine/puzzle/: the Part II engine. It names no noun: the engine's own words are the operators IS, AND, NOT, the meta-noun TEXT and the property PUSH of the base rule TEXT IS PUSH.
- grid.ts: the level format (LevelDef, course/schema/level.schema.json), object kinds ("<noun>" or "text:<WORD>") and the board: objects on tiles, several per tile allowed.
- word.ts: the lexicon (every word the parser can read, with its role: noun, property or operator) and the lexer, which turns the text on the board into words by tile. Text whose word isn't in the lexicon spells nothing.
- parse.ts: the parser. From every IS, in a row and in a column, it collects the AND-joined subjects before it and the AND-joined complements after it, and emits one rule per pair. Stacked words each count. Base rules come first.
- level.ts: a level in play. Every change to the board recompiles the rules from scratch.
- index.ts: the module's exports.
The rest of src/engine/: Part I, unchanged from Module 5 except for the export that adds puzzle.
src/content/words.json: the nouns and properties the parser knows, as data ({ "name", "role" }). src/content/index.ts reads it with a JSON import as words() (tsconfig.json gains resolveJsonModule so tsc accepts the import; Vitest needs nothing). A new word is a new entry here.
src/harness.ts: the adapter the course tests call. Module 6 adds newBaba, and place and rules on Baba. rules() sorts the rules' normalised strings and keeps each distinct rule once; the engine itself keeps a rule spelled twice twice.

Run it with the module's tests: node tools/verify-references.mjs 06 from the course repository.