@stll/anonymize 1.3.0 → 1.4.1

package/dist/rolldown-runtime.mjs

@@ -0,0 +1,13 @@
+//#region \0rolldown/runtime.js
+var __defProp = Object.defineProperty;
+var __exportAll = (all, no_symbols) => {
+	let target = {};
+	for (var name in all) __defProp(target, name, {
+		get: all[name],
+		enumerable: true
+	});
+	if (!no_symbols) __defProp(target, Symbol.toStringTag, { value: "Module" });
+	return target;
+};
+//#endregion
+export { __exportAll as t };

package/dist/structural-single-cap-prefixes.mjs

@@ -0,0 +1,99 @@
+//#region src/data/structural-single-cap-prefixes.json
+var _comment = "Structural document-unit words that can precede a single-letter or letter-number marker before a legal form word, e.g. 'Schedule A LLC Members' or 'Exhibit A-1 LLC Agreement'. Lowercased; matched case-insensitively. Organised per-language; runtime unions all languages because match spans do not carry a language tag.";
+var en = [
+	"annex",
+	"appendix",
+	"article",
+	"attachment",
+	"clause",
+	"exhibit",
+	"paragraph",
+	"part",
+	"schedule",
+	"section"
+];
+var cs = [
+	"bod",
+	"paragraf",
+	"podbod",
+	"ustanovení",
+	"článek"
+];
+var sk = [
+	"bod",
+	"hlava",
+	"oddiel",
+	"odsek",
+	"pododdiel",
+	"sekcia",
+	"časť",
+	"článok"
+];
+var de = [
+	"absatz",
+	"abschnitt",
+	"abteilung",
+	"anlage",
+	"anhang",
+	"artikel",
+	"kapitel",
+	"klausel",
+	"punkt",
+	"sektion",
+	"unterabschnitt",
+	"unterpunkt"
+];
+var es = [
+	"anexo",
+	"apartado",
+	"artículo",
+	"cláusula",
+	"sección"
+];
+var fr = [
+	"annexe",
+	"article",
+	"clause",
+	"paragraphe",
+	"partie",
+	"section"
+];
+var it = [
+	"allegato",
+	"appendice",
+	"articolo",
+	"clausola",
+	"paragrafo",
+	"sezione"
+];
+var pl = [
+	"aneksem",
+	"artykuł",
+	"paragraf",
+	"punkt",
+	"sekcja",
+	"załącznik"
+];
+var pt_br = [
+	"anexo",
+	"artigo",
+	"cláusula",
+	"parágrafo",
+	"seção"
+];
+var structural_single_cap_prefixes_default = {
+	_comment,
+	en,
+	cs,
+	sk,
+	de,
+	es,
+	fr,
+	it,
+	pl,
+	"pt-br": pt_br
+};
+//#endregion
+export { _comment, cs, de, structural_single_cap_prefixes_default as default, en, es, fr, it, pl, pt_br as "pt-br", sk };
+
+//# sourceMappingURL=structural-single-cap-prefixes.mjs.map

package/dist/structural-single-cap-prefixes.mjs.map

@@ -0,0 +1 @@
+{"version":3,"file":"structural-single-cap-prefixes.mjs","names":[],"sources":["../src/data/structural-single-cap-prefixes.json"],"sourcesContent":[""],"mappings":""}

package/dist/triggers.en.mjs

@@ -33,7 +33,22 @@ var triggers_en_default = [
 	{
 		"id": "en-jurisdiction",
 		"label": "address",
-		"strategy": { "type": "to-next-comma" },
+		"strategy": {
+			"type": "to-next-comma",
+			"maxLength": 30,
+			"stopWords": [
+				"or",
+				"and",
+				"in",
+				"sitting",
+				"shall",
+				"that",
+				"which",
+				"without",
+				"for",
+				"to"
+			]
+		},
 		"triggers": [
 			"State of",
 			"Commonwealth of",
@@ -64,6 +79,38 @@ var triggers_en_default = [
 			"tax identification number",
 			"tax id"
 		]
+	},
+	{
+		"id": "en-uk-companies-house",
+		"label": "registration number",
+		"strategy": { "type": "company-id-value" },
+		"triggers": [
+			"Companies House",
+			"Company No.",
+			"Company No",
+			"CRN"
+		],
+		"extensions": ["add-colon", "add-colon-space"],
+		"validations": [{
+			"type": "matches-pattern",
+			"pattern": "^(?:SC|NI|OC|SO|R|FC|GE|GS|IC|IP|LP|NC|NF|NL|NO|NP|NR|NZ|RC|SA|SF|SI|SL|SP|SR|SZ)?\\d{6,8}$"
+		}]
+	},
+	{
+		"id": "en-uk-utr",
+		"label": "tax identification number",
+		"strategy": { "type": "company-id-value" },
+		"triggers": [
+			"UTR",
+			"Unique Tax Reference",
+			"Unique Taxpayer Reference",
+			"Self Assessment UTR"
+		],
+		"extensions": ["add-colon", "add-colon-space"],
+		"validations": [{
+			"type": "matches-pattern",
+			"pattern": "^\\d{10}$"
+		}]
 	}
 ];
 //#endregion

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@stll/anonymize",
-  "version": "1.3.0",
+  "version": "1.4.1",
   "description": "Multi-layer PII detection and anonymization pipeline with regex, NER, deny lists, and coreference resolution",
   "type": "module",
   "exports": {
@@ -8,6 +8,11 @@
       "types": "./dist/index.d.mts",
       "import": "./dist/index.mjs",
       "default": "./dist/index.mjs"
+    },
+    "./constants": {
+      "types": "./dist/constants.d.mts",
+      "import": "./dist/constants.mjs",
+      "default": "./dist/constants.mjs"
     }
   },
   "types": "dist/index.d.mts",
@@ -27,6 +32,7 @@
     "prepublishOnly": "bun run build",
     "typecheck": "tsc --noEmit -p tsconfig.json && tsc --noEmit -p tsconfig.test.json && tsc --noEmit -p tsconfig.wasm.json",
     "test": "bun test --preload ./src/__test__/setup.ts",
+    "test:fast": "bun test --preload ./src/__test__/setup.ts src/__test__/*.test.ts",
     "lint": "prettier --check . --ignore-path ../../.prettierignore",
     "format": "prettier --write . --ignore-path ../../.prettierignore"
   },