diff --git a/packages/cli/src/utils/craftAudit.test.ts b/packages/cli/src/utils/craftAudit.test.ts index a1d8242e1a..25ed3ac5ff 100644 --- a/packages/cli/src/utils/craftAudit.test.ts +++ b/packages/cli/src/utils/craftAudit.test.ts @@ -81,6 +81,12 @@ describe("readingWords", () => { expect(readingWords("✓ — →")).toBe(0); expect(readingWords("今天天气很好")).toBeGreaterThan(1); }); + + it("counts a contraction as one word", () => { + expect(readingWords("a tooth can't get in")).toBe(5); + expect(readingWords("a tooth can’t get in")).toBe(5); + expect(readingWords("it's 'quoted'")).toBe(2); + }); }); describe("lineFloorSec", () => { diff --git a/packages/cli/src/utils/craftAudit.ts b/packages/cli/src/utils/craftAudit.ts index c76ed90484..b2456f64d5 100644 --- a/packages/cli/src/utils/craftAudit.ts +++ b/packages/cli/src/utils/craftAudit.ts @@ -29,7 +29,7 @@ const CRAFT_THRESHOLDS = { const T = CRAFT_THRESHOLDS; const ZERO_BBOX: CheckBbox = { x: 0, y: 0, width: 0, height: 0 }; -const WORD_RUNS = /[\p{L}\p{M}\p{N}%$]+/gu; +const WORD_RUNS = /[\p{L}\p{M}\p{N}%$]+(?:['’][\p{L}\p{M}\p{N}]+)*/gu; // Chinese and Japanese have no spaces; a dictionary segmenter counts their words like an English line's. const WORD_SEGMENTER = new Intl.Segmenter("ja", { granularity: "word" });