@mailwoman/codex 8.5.0 → 9.0.0

package/postcode-systems.ts

@@ -27,7 +27,7 @@ import { normalizePLZ } from "./de/index.ts"
 import { normalizeCodePostal } from "./fr/index.ts"
 import { normalizeUkPostcode } from "./gb/index.ts"
 import { normalizeJpPostalCode } from "./jp/index.ts"
-import { normalizeNzPostcode } from "./nz/index.ts"
+import { normalizeNZPostcode } from "./nz/index.ts"
 import { isZipCode } from "./us/index.ts"
 
 /**
@@ -48,7 +48,7 @@ const SYSTEM_ACCEPTS: ReadonlyArray<readonly [SystemCode, (s: string) => boolean
 	["gb", (s) => normalizeUkPostcode(s) !== null],
 	["jp", (s) => normalizeJpPostalCode(s) !== null],
 	["au", (s) => normalizeAuPostcode(s) !== null],
-	["nz", (s) => normalizeNzPostcode(s) !== null],
+	["nz", (s) => normalizeNZPostcode(s) !== null],
 ]
 
 /**

package/tools/build-country-surface-lexicon.ts

@@ -44,6 +44,7 @@ import { mkdirSync, writeFileSync } from "node:fs"
 import { dirname, resolve } from "node:path"
 
 import { COUNTRY_SURFACE_FORMS, ISO2_TO_NAME } from "../country/country.ts"
+import { wordNorm, wordNormLower } from "../normalize.ts"
 import { US_STATE_ABBREVIATIONS, US_STATE_NAMES } from "../us/state.ts"
 
 /**
@@ -71,15 +72,6 @@ const OUTPUT = resolve(import.meta.dirname, "../../data/gazetteer/country-surfac
  * letters or digits (keep internal ones: "u.s.a", "timor-leste"), rejoin single-spaced. Entry keys and scanned tokens
  * both pass through it, so "U.S.A." ≡ "u.s.a".
  */
-const wordNorm = (s: string): string =>
-	s
-		.split(/\s+/)
-		.map((w) => w.replaceAll(/^[^\p{L}\p{N}]+|[^\p{L}\p{N}]+$/gu, ""))
-		.filter(Boolean)
-		.join(" ")
-
-const norm = (s: string): string => wordNorm(s).toLowerCase()
-
 /**
  * Short alphabetic code (≤3 letters once punctuation is dropped) → exact-uppercase matching.
  */
@@ -125,7 +117,7 @@ function add(surface: string): void {
 		return
 	}
 
-	const key = norm(s)
+	const key = wordNormLower(s)
 
 	if (!key) return
 	const words = key.split(" ")

package/tools/generate-official-languages.ts

@@ -23,6 +23,8 @@ import { readFileSync, writeFileSync } from "node:fs"
 import { join } from "node:path"
 import { fileURLToPath } from "node:url"
 
+import { parseJSONStrict } from "@mailwoman/core/objects"
+
 /**
  * The committed output path, resolved relative to this module (codex/tools/ → codex/country/). The codegen is
  * repo-only, and in the repo `@mailwoman/codex/tools` always loads from source via the `node` exports condition, so
@@ -62,7 +64,10 @@ interface LanguagePopulation {
 }
 
 async function loadCLDR(file: string, cldrDir: string | undefined, cldrVersion: string): Promise<unknown> {
-	if (cldrDir) return JSON.parse(readFileSync(join(cldrDir, `cldr-${file}.json`), "utf8"))
+	if (cldrDir) {
+		return parseJSONStrict(readFileSync(join(cldrDir, `cldr-${file}.json`), "utf8"))
+	}
+
 	const url = `https://cdn.jsdelivr.net/npm/cldr-core@${cldrVersion}/supplemental/${file}.json`
 	const res = await fetch(url)