@stll/anonymize-wasm 1.0.8 → 1.1.0

package/dist/wasm.mjs

@@ -1,7 +1,6 @@
 import { TextSearch } from "@stll/text-search-wasm";
 import { at, be, bg, cy, cz, de, dk, ee, es, fi, fr, gb, gr, hr, hu, ie, it, lt, lu, lv, mt, nl, pl, pt, ro, se, si, sk } from "@stll/stdnum";
 import { toRegex } from "@stll/stdnum/patterns";
-import charGroupsJson from "@stll/anonymize-data/config/char-groups.json";
 //#region src/search-engine.ts
 let _TextSearch;
 const initTextSearch = (ctor) => {
@@ -125,7 +124,7 @@ const loadManifest = () => {
 	if (_manifestPromise) return _manifestPromise;
 	_manifestPromise = (async () => {
 		try {
-			const mod = await import("@stll/anonymize-data/config/manifest.json");
+			const mod = await import("./manifest.mjs");
 			const parsed = mod.default ?? mod;
 			if (!parsed || typeof parsed.languages !== "object" || parsed.languages === null || Array.isArray(parsed.languages)) {
 				console.warn("[anonymize] lang-loader: manifest has unexpected structure, falling back to hardcoded list");
@@ -144,23 +143,23 @@ const loadManifest = () => {
 };
 const LOADER_REGISTRIES = {
 	triggers: {
-		cs: () => import("@stll/anonymize-data/config/triggers.cs.json"),
-		de: () => import("@stll/anonymize-data/config/triggers.de.json"),
-		en: () => import("@stll/anonymize-data/config/triggers.en.json"),
-		es: () => import("@stll/anonymize-data/config/triggers.es.json"),
-		fr: () => import("@stll/anonymize-data/config/triggers.fr.json"),
-		hu: () => import("@stll/anonymize-data/config/triggers.hu.json"),
-		it: () => import("@stll/anonymize-data/config/triggers.it.json"),
-		pl: () => import("@stll/anonymize-data/config/triggers.pl.json"),
-		ro: () => import("@stll/anonymize-data/config/triggers.ro.json"),
-		sk: () => import("@stll/anonymize-data/config/triggers.sk.json"),
-		sv: () => import("@stll/anonymize-data/config/triggers.sv.json")
+		cs: () => import("./triggers.cs.mjs"),
+		de: () => import("./triggers.de.mjs"),
+		en: () => import("./triggers.en.mjs"),
+		es: () => import("./triggers.es.mjs"),
+		fr: () => import("./triggers.fr.mjs"),
+		hu: () => import("./triggers.hu.mjs"),
+		it: () => import("./triggers.it.mjs"),
+		pl: () => import("./triggers.pl.mjs"),
+		ro: () => import("./triggers.ro.mjs"),
+		sk: () => import("./triggers.sk.mjs"),
+		sv: () => import("./triggers.sv.mjs")
 	},
 	coreference: {
-		cs: () => import("@stll/anonymize-data/config/coreference.cs.json"),
-		de: () => import("@stll/anonymize-data/config/coreference.de.json"),
-		en: () => import("@stll/anonymize-data/config/coreference.en.json"),
-		sk: () => import("@stll/anonymize-data/config/coreference.sk.json")
+		cs: () => import("./coreference.cs.mjs"),
+		de: () => import("./coreference.de.mjs"),
+		en: () => import("./coreference.en.mjs"),
+		sk: () => import("./coreference.sk.mjs")
 	}
 };
 const FALLBACK_LANGUAGES = {
@@ -231,7 +230,7 @@ const loadLanguageConfigs = async (configType, mapFn) => {
 //#region src/detectors/coreference.ts
 /**
 * Load coreference definition patterns from per-language
-* JSON configs in @stll/anonymize-data. Uses the
+* JSON configs in src/data/. Uses the
 * language manifest for auto-discovery.
 */
 const loadDefinitionPatterns = async () => {
@@ -264,7 +263,7 @@ const getRoleStopSet = async (ctx) => {
 	const promise = (async () => {
 		let result;
 		try {
-			const mod = await import("@stll/anonymize-data/config/generic-roles.json");
+			const mod = await import("./generic-roles.mjs");
 			const data = mod.default ?? mod;
 			result = new Set(data.roles.map((r) => r.toLowerCase()));
 		} catch {
@@ -607,58 +606,31 @@ const getNameCorpusFirstNames = (ctx = defaultContext) => ctx.nameCorpus?.firstN
 const getNameCorpusSurnames = (ctx = defaultContext) => ctx.nameCorpus?.surnamesList ?? [];
 const getNameCorpusTitles = (ctx = defaultContext) => ctx.nameCorpus?.titlesList ?? [];
 /**
-* Languages with per-language first/surname
-* dictionaries in @stll/anonymize-data.
-*/
-const NAME_LANGUAGES = [
-	"cs",
-	"sk",
-	"de",
-	"pl",
-	"hu",
-	"ro",
-	"fr",
-	"es",
-	"it",
-	"en",
-	"sv"
-];
-/**
-* Try importing a JSON module; return empty array
-* if not found.
-*/
-const tryImportArray = async (path) => {
-	try {
-		return (await import(path)).default;
-	} catch {
-		return [];
-	}
-};
-/**
-* Load name corpus data from per-language dictionary
-* files and legacy config files. Merges all sources.
+* Load name corpus data from injected dictionaries
+* and legacy config files. Merges all sources.
 *
 * Safe to call multiple times; only loads once per
 * context. Must be called before detectNameCorpus or
 * the getNameCorpus*() accessors are used.
+*
+* @param dictionaries Optional pre-loaded dictionaries
+*   with per-language first names and surnames. When
+*   omitted, only legacy config files are used.
 */
-const initNameCorpus = (ctx = defaultContext) => {
+const initNameCorpus = (ctx = defaultContext, dictionaries) => {
 	if (ctx.nameCorpusPromise) return ctx.nameCorpusPromise;
 	const promise = (async () => {
 		try {
 			const [legacyFirstMod, legacySurnameMod, titleMod, exclusionMod] = await Promise.all([
-				import("@stll/anonymize-data/config/names-first.json"),
-				import("@stll/anonymize-data/config/names-surnames.json"),
-				import("@stll/anonymize-data/config/names-title-tokens.json"),
-				import("@stll/anonymize-data/config/names-exclusions.json")
+				import("./names-first.mjs"),
+				import("./names-surnames.mjs"),
+				import("./names-title-tokens.mjs"),
+				import("./names-exclusions.mjs")
 			]);
-			const firstImports = NAME_LANGUAGES.map((lang) => tryImportArray(`@stll/anonymize-data/dictionaries/names/first/${lang}.json`));
-			const surnameImports = NAME_LANGUAGES.map((lang) => tryImportArray(`@stll/anonymize-data/dictionaries/names/surnames/${lang}.json`));
-			const [firstResults, surnameResults] = await Promise.all([Promise.all(firstImports), Promise.all(surnameImports)]);
 			const firstNames = [...legacyFirstMod.default.names];
-			for (const names of firstResults) for (const name of names) firstNames.push(name);
+			if (dictionaries?.firstNames) for (const names of Object.values(dictionaries.firstNames)) for (const name of names) firstNames.push(name);
 			const surnames = [...legacySurnameMod.default.names];
-			for (const names of surnameResults) for (const name of names) surnames.push(name);
+			if (dictionaries?.surnames) for (const names of Object.values(dictionaries.surnames)) for (const name of names) surnames.push(name);
 			const dedup = (arr) => {
 				const seen = /* @__PURE__ */ new Set();
 				const result = [];
@@ -1087,12 +1059,515 @@ const POST_NOMINALS = [
 	"CIPP"
 ];
 //#endregion
+//#region src/data/char-groups.json
+var char_groups_default = {
+	_comment: "Unicode character equivalence groups. Used to build regex character classes that match all variants of a character type.",
+	groups: {
+		"dash": {
+			"description": "Dash-like characters (hyphens, en-dash, em-dash, minus)",
+			"chars": [
+				{
+					"char": "-",
+					"name": "hyphen-minus",
+					"code": "U+002D"
+				},
+				{
+					"char": "–",
+					"name": "en-dash",
+					"code": "U+2013"
+				},
+				{
+					"char": "—",
+					"name": "em-dash",
+					"code": "U+2014"
+				},
+				{
+					"char": "‐",
+					"name": "hyphen",
+					"code": "U+2010"
+				},
+				{
+					"char": "‑",
+					"name": "non-breaking hyphen",
+					"code": "U+2011"
+				},
+				{
+					"char": "−",
+					"name": "minus sign",
+					"code": "U+2212"
+				},
+				{
+					"char": "⁃",
+					"name": "hyphen bullet",
+					"code": "U+2043"
+				},
+				{
+					"char": "‒",
+					"name": "figure dash",
+					"code": "U+2012"
+				},
+				{
+					"char": "―",
+					"name": "horizontal bar",
+					"code": "U+2015"
+				},
+				{
+					"char": "⸺",
+					"name": "two-em dash",
+					"code": "U+2E3A"
+				},
+				{
+					"char": "⸻",
+					"name": "three-em dash",
+					"code": "U+2E3B"
+				},
+				{
+					"char": "־",
+					"name": "Hebrew maqaf",
+					"code": "U+05BE"
+				}
+			]
+		},
+		"space": {
+			"description": "Space-like characters",
+			"chars": [
+				{
+					"char": " ",
+					"name": "space",
+					"code": "U+0020"
+				},
+				{
+					"char": "\xA0",
+					"name": "no-break space",
+					"code": "U+00A0"
+				},
+				{
+					"char": " ",
+					"name": "en space",
+					"code": "U+2002"
+				},
+				{
+					"char": " ",
+					"name": "em space",
+					"code": "U+2003"
+				},
+				{
+					"char": " ",
+					"name": "thin space",
+					"code": "U+2009"
+				},
+				{
+					"char": " ",
+					"name": "hair space",
+					"code": "U+200A"
+				},
+				{
+					"char": " ",
+					"name": "narrow no-break space",
+					"code": "U+202F"
+				},
+				{
+					"char": " ",
+					"name": "Ogham space mark",
+					"code": "U+1680"
+				},
+				{
+					"char": " ",
+					"name": "en quad",
+					"code": "U+2000"
+				},
+				{
+					"char": " ",
+					"name": "em quad",
+					"code": "U+2001"
+				},
+				{
+					"char": " ",
+					"name": "three-per-em space",
+					"code": "U+2004"
+				},
+				{
+					"char": " ",
+					"name": "four-per-em space",
+					"code": "U+2005"
+				},
+				{
+					"char": " ",
+					"name": "six-per-em space",
+					"code": "U+2006"
+				},
+				{
+					"char": " ",
+					"name": "figure space",
+					"code": "U+2007"
+				},
+				{
+					"char": " ",
+					"name": "punctuation space",
+					"code": "U+2008"
+				},
+				{
+					"char": " ",
+					"name": "medium mathematical space",
+					"code": "U+205F"
+				},
+				{
+					"char": "　",
+					"name": "ideographic space",
+					"code": "U+3000"
+				},
+				{
+					"char": "​",
+					"name": "zero width space",
+					"code": "U+200B"
+				}
+			]
+		},
+		"quote-double": {
+			"description": "Double quotation marks",
+			"chars": [
+				{
+					"char": "\"",
+					"name": "quotation mark",
+					"code": "U+0022"
+				},
+				{
+					"char": "“",
+					"name": "left double quotation mark",
+					"code": "U+201C"
+				},
+				{
+					"char": "”",
+					"name": "right double quotation mark",
+					"code": "U+201D"
+				},
+				{
+					"char": "„",
+					"name": "double low-9 quotation mark",
+					"code": "U+201E"
+				},
+				{
+					"char": "«",
+					"name": "left guillemet",
+					"code": "U+00AB"
+				},
+				{
+					"char": "»",
+					"name": "right guillemet",
+					"code": "U+00BB"
+				}
+			]
+		},
+		"quote-single": {
+			"description": "Single quotation marks and apostrophes",
+			"chars": [
+				{
+					"char": "'",
+					"name": "apostrophe",
+					"code": "U+0027"
+				},
+				{
+					"char": "‘",
+					"name": "left single quotation mark",
+					"code": "U+2018"
+				},
+				{
+					"char": "’",
+					"name": "right single quotation mark",
+					"code": "U+2019"
+				},
+				{
+					"char": "‚",
+					"name": "single low-9 quotation mark",
+					"code": "U+201A"
+				},
+				{
+					"char": "‹",
+					"name": "left single guillemet",
+					"code": "U+2039"
+				},
+				{
+					"char": "›",
+					"name": "right single guillemet",
+					"code": "U+203A"
+				},
+				{
+					"char": "ʼ",
+					"name": "modifier letter apostrophe",
+					"code": "U+02BC"
+				},
+				{
+					"char": "´",
+					"name": "acute accent",
+					"code": "U+00B4"
+				},
+				{
+					"char": "`",
+					"name": "grave accent",
+					"code": "U+0060"
+				},
+				{
+					"char": "ʽ",
+					"name": "modifier letter reversed comma",
+					"code": "U+02BD"
+				},
+				{
+					"char": "ʾ",
+					"name": "modifier letter right half ring",
+					"code": "U+02BE"
+				},
+				{
+					"char": "ʿ",
+					"name": "modifier letter left half ring",
+					"code": "U+02BF"
+				},
+				{
+					"char": "ˈ",
+					"name": "modifier letter vertical line",
+					"code": "U+02C8"
+				}
+			]
+		},
+		"dot": {
+			"description": "Dot-like characters",
+			"chars": [
+				{
+					"char": ".",
+					"name": "full stop",
+					"code": "U+002E"
+				},
+				{
+					"char": "·",
+					"name": "middle dot",
+					"code": "U+00B7"
+				},
+				{
+					"char": "•",
+					"name": "bullet",
+					"code": "U+2022"
+				},
+				{
+					"char": "․",
+					"name": "one dot leader",
+					"code": "U+2024"
+				},
+				{
+					"char": "。",
+					"name": "CJK full stop",
+					"code": "U+3002"
+				},
+				{
+					"char": "׃",
+					"name": "Hebrew sof pasuq",
+					"code": "U+05C3"
+				}
+			]
+		},
+		"slash": {
+			"description": "Slash-like characters",
+			"chars": [
+				{
+					"char": "/",
+					"name": "solidus",
+					"code": "U+002F"
+				},
+				{
+					"char": "⁄",
+					"name": "fraction slash",
+					"code": "U+2044"
+				},
+				{
+					"char": "∕",
+					"name": "division slash",
+					"code": "U+2215"
+				}
+			]
+		},
+		"colon": {
+			"description": "Colon-like characters",
+			"chars": [
+				{
+					"char": ":",
+					"name": "colon",
+					"code": "U+003A"
+				},
+				{
+					"char": "꞉",
+					"name": "modifier letter colon",
+					"code": "U+A789"
+				},
+				{
+					"char": "∶",
+					"name": "ratio",
+					"code": "U+2236"
+				},
+				{
+					"char": "：",
+					"name": "full-width colon",
+					"code": "U+FF1A"
+				},
+				{
+					"char": "፥",
+					"name": "Ethiopic colon",
+					"code": "U+1365"
+				}
+			]
+		},
+		"comma": {
+			"description": "Comma-like characters",
+			"chars": [
+				{
+					"char": ",",
+					"name": "comma",
+					"code": "U+002C"
+				},
+				{
+					"char": "‚",
+					"name": "single low-9 quotation mark (used as comma)",
+					"code": "U+201A"
+				},
+				{
+					"char": "،",
+					"name": "Arabic comma",
+					"code": "U+060C"
+				},
+				{
+					"char": "、",
+					"name": "ideographic comma",
+					"code": "U+3001"
+				},
+				{
+					"char": "，",
+					"name": "full-width comma",
+					"code": "U+FF0C"
+				},
+				{
+					"char": "﹐",
+					"name": "small comma",
+					"code": "U+FE50"
+				},
+				{
+					"char": "՝",
+					"name": "Armenian comma",
+					"code": "U+055D"
+				}
+			]
+		},
+		"ellipsis": {
+			"description": "Ellipsis characters",
+			"chars": [{
+				"char": "…",
+				"name": "horizontal ellipsis",
+				"code": "U+2026"
+			}]
+		},
+		"question-mark": {
+			"description": "Question mark variants",
+			"chars": [
+				{
+					"char": "?",
+					"name": "question mark",
+					"code": "U+003F"
+				},
+				{
+					"char": "？",
+					"name": "full-width question mark",
+					"code": "U+FF1F"
+				},
+				{
+					"char": "⁇",
+					"name": "double question mark",
+					"code": "U+2047"
+				},
+				{
+					"char": "⁈",
+					"name": "question exclamation mark",
+					"code": "U+2048"
+				},
+				{
+					"char": "؟",
+					"name": "Arabic question mark",
+					"code": "U+061F"
+				},
+				{
+					"char": "‽",
+					"name": "interrobang",
+					"code": "U+203D"
+				}
+			]
+		},
+		"exclamation-mark": {
+			"description": "Exclamation mark variants",
+			"chars": [
+				{
+					"char": "!",
+					"name": "exclamation mark",
+					"code": "U+0021"
+				},
+				{
+					"char": "！",
+					"name": "full-width exclamation mark",
+					"code": "U+FF01"
+				},
+				{
+					"char": "‼",
+					"name": "double exclamation mark",
+					"code": "U+203C"
+				},
+				{
+					"char": "¡",
+					"name": "inverted exclamation mark",
+					"code": "U+00A1"
+				}
+			]
+		},
+		"semicolon": {
+			"description": "Semicolon variants",
+			"chars": [
+				{
+					"char": ";",
+					"name": "semicolon",
+					"code": "U+003B"
+				},
+				{
+					"char": "؛",
+					"name": "Arabic semicolon",
+					"code": "U+061B"
+				},
+				{
+					"char": "；",
+					"name": "full-width semicolon",
+					"code": "U+FF1B"
+				},
+				{
+					"char": "፤",
+					"name": "Ethiopic semicolon",
+					"code": "U+1364"
+				}
+			]
+		},
+		"underscore": {
+			"description": "Underscore variants",
+			"chars": [{
+				"char": "_",
+				"name": "low line",
+				"code": "U+005F"
+			}, {
+				"char": "＿",
+				"name": "full-width low line",
+				"code": "U+FF3F"
+			}]
+		}
+	}
+};
+//#endregion
 //#region src/util/char-groups.ts
 /**
 * Centralized Unicode character equivalence groups.
 *
-* Loads char-groups.json from @stll/anonymize-data and
-* provides helpers to build regex character classes that
+* Centralized Unicode character equivalence groups.
+*
+* Provides helpers to build regex character classes that
 * match all typographic variants of a character type
 * (dashes, spaces, quotes, etc.).
 *
@@ -1102,7 +1577,7 @@ const POST_NOMINALS = [
 /** Chars that need escaping inside a regex char class. */
 const REGEX_CLASS_SPECIAL = /[\\\]^-]/;
 const escapeForCharClass = (ch) => REGEX_CLASS_SPECIAL.test(ch) ? `\\${ch}` : ch;
-const config = charGroupsJson;
+const config = char_groups_default;
 /**
 * Get the raw characters for a named group.
 * Throws if the group does not exist.
@@ -1441,7 +1916,7 @@ const buildDatePatternsFromMonths = (alt) => {
 /** Cached promise for date patterns. Loaded once. */
 let datePatternPromise = null;
 const loadDatePatterns = async () => {
-	const mod = await import("@stll/anonymize-data/config/date-months.json");
+	const mod = await import("./date-months.mjs");
 	return buildDatePatternsFromMonths(buildMonthAlternation(mod.default ?? mod));
 };
 /**
@@ -1508,7 +1983,7 @@ const buildCurrencyPatterns = (data) => {
 /** Cached promise for currency patterns. Loaded once. */
 let currencyPatternPromise = null;
 const loadCurrencyPatterns = async () => {
-	const mod = await import("@stll/anonymize-data/config/currencies.json");
+	const mod = await import("./currencies.mjs");
 	return buildCurrencyPatterns(mod.default ?? mod);
 };
 /**
@@ -1590,7 +2065,7 @@ const SIGNING_CLAUSE_META = {
 };
 let signingPatternPromise = null;
 const loadSigningPatterns = async () => {
-	const mod = await import("@stll/anonymize-data/config/signing-clauses.json");
+	const mod = await import("./signing-clauses.mjs");
 	return buildSigningClausePatterns(mod.default ?? mod);
 };
 const getSigningClausePatterns = () => {
@@ -1624,7 +2099,7 @@ const buildPatternString = (forms) => {
 const buildLegalFormPatterns = async () => {
 	let data = {};
 	try {
-		data = (await import("@stll/anonymize-data/config/legal-forms.json")).default;
+		data = (await import("./legal-forms.mjs")).default;
 	} catch {
 		return [];
 	}
@@ -1942,14 +2417,14 @@ const buildTriggerPatterns = async () => {
 		rules.push(...expandTriggerGroups(groups));
 	}
 	try {
-		const globalMod = await import("@stll/anonymize-data/config/triggers.global.json");
+		const globalMod = await import("./triggers.global.mjs");
 		const globalGroups = globalMod.default ?? globalMod;
 		if (Array.isArray(globalGroups)) rules.push(...expandTriggerGroups(globalGroups));
 	} catch (err) {
 		if (!(err instanceof Error) || !err.message.includes("Cannot find module")) throw err;
 	}
 	try {
-		const yearMod = await import("@stll/anonymize-data/config/year-words.json");
+		const yearMod = await import("./year-words.mjs");
 		const data = yearMod.default ?? yearMod;
 		const seen = /* @__PURE__ */ new Set();
 		const yearValidation = compileValidations([{
@@ -2566,7 +3041,7 @@ const loadGenericRoles = (ctx = defaultContext) => {
 	if (ctx.genericRolesPromise) return ctx.genericRolesPromise;
 	ctx.genericRolesPromise = (async () => {
 		try {
-			const mod = await import("@stll/anonymize-data/config/generic-roles.json");
+			const mod = await import("./generic-roles.mjs");
 			const set = new Set(mod.default?.roles ?? []);
 			ctx.genericRoles = set;
 			return set;
@@ -2654,22 +3129,11 @@ const normalizeForSearch = (text) => {
 };
 //#endregion
 //#region src/detectors/deny-list.ts
-/**
-* Try to load the optional @stll/anonymize-data package.
-* Returns null if not installed.
-*/
-const loadDataModule = async () => {
-	try {
-		return await import("@stll/anonymize-data");
-	} catch {
-		return null;
-	}
-};
 const loadAllowList = (ctx) => {
 	if (ctx.allowListPromise) return ctx.allowListPromise;
 	ctx.allowListPromise = (async () => {
 		try {
-			const mod = await import("@stll/anonymize-data/config/allow-list.json");
+			const mod = await import("./allow-list.mjs");
 			const set = new Set(mod.default?.words ?? []);
 			ctx.allowList = set;
 			return set;
@@ -2744,7 +3208,7 @@ const loadStopwords = (ctx) => {
 	if (ctx.stopwordsPromise) return ctx.stopwordsPromise;
 	ctx.stopwordsPromise = (async () => {
 		try {
-			const list = ((await import("@stll/anonymize-data/config/stopwords.json")).default ?? []).filter((w) => !getFirstNameExclusions(ctx).has(w));
+			const list = ((await import("./stopwords.mjs")).default ?? []).filter((w) => !getFirstNameExclusions(ctx).has(w));
 			const set = new Set(list);
 			ctx.stopwords = set;
 			return set;
@@ -2764,7 +3228,7 @@ const loadPersonStopwords = (ctx) => {
 	if (ctx.personStopwordsPromise) return ctx.personStopwordsPromise;
 	ctx.personStopwordsPromise = (async () => {
 		try {
-			const mod = await import("@stll/anonymize-data/config/person-stopwords.json");
+			const mod = await import("./person-stopwords.mjs");
 			const set = new Set(mod.default?.words ?? []);
 			ctx.personStopwords = set;
 			return set;
@@ -2783,46 +3247,33 @@ const PERSON_CHAIN_BREAK_RE = /[!?;:]|,/u;
 const isInitialContinuationGap = (text, gap) => /^\p{Lu}$/u.test(text) && /^\.[^\S\n]{1,2}$/u.test(gap) || /^[^\S\n]{1,2}(?:\p{Lu}\.[^\S\n]{1,2})+$/u.test(gap);
 /**
 * Resolve which dictionaries to load based on country
-* and category filters, load them, and build the deny
-* list data. The returned data provides PatternEntry[]
-* for the unified builder and parallel arrays for
+* and category filters, then build the deny list data.
+* The returned data provides PatternEntry[] for the
+* unified builder and parallel arrays for
 * post-processing.
 *
-* Requires `@stll/anonymize-data` to be installed.
-* Returns null if the data package is not available.
+* Dictionary data is injected via `config.dictionaries`.
+* Returns null if no dictionaries are provided.
 */
 const buildDenyList = async (config, ctx = defaultContext) => {
-	await initNameCorpus(ctx);
+	await initNameCorpus(ctx, config.dictionaries);
 	await Promise.all([
 		loadStopwords(ctx),
 		loadAllowList(ctx),
 		loadPersonStopwords(ctx),
 		loadGenericRoles(ctx)
 	]);
-	const dataModule = await loadDataModule();
-	if (!dataModule) return null;
+	const dictionaries = config.dictionaries;
+	const hasDenyList = dictionaries?.denyList && dictionaries?.denyListMeta;
+	const hasCities = dictionaries?.cities && dictionaries.cities.length > 0;
+	if (!hasDenyList && !hasCities) return buildNameCorpusOnly(config, ctx);
 	const allowedCountries = resolveCountries(config.denyListRegions, config.denyListCountries);
 	const excluded = config.denyListExcludeCategories;
 	const excludeCategories = excluded ? new Set(excluded) : /* @__PURE__ */ new Set();
-	const ids = [...dataModule.ALL_DICTIONARY_IDS].filter((id) => {
-		const meta = dataModule.DICTIONARY_META[id];
-		if (!meta) return false;
-		if (!config.enableNameCorpus && meta.category === "Names") return false;
-		if (excludeCategories.has(meta.category)) return false;
-		if (allowedCountries === null) return true;
-		if (meta.country === null) return true;
-		return allowedCountries.has(meta.country);
-	});
 	const patternList = [];
 	const labelList = [];
 	const sourceList = [];
 	const patternIndex = /* @__PURE__ */ new Map();
-	const results = await Promise.all(ids.map(async (id) => {
-		return {
-			id,
-			entries: await dataModule.loadDictionary(id)
-		};
-	}));
 	const addDenyListEntry = (entry, label) => {
 		const normalized = normalizeForSearch(entry).replace(/[|\\]/g, "");
 		if (normalized.length === 0) return;
@@ -2838,47 +3289,59 @@ const buildDenyList = async (config, ctx = defaultContext) => {
 			sourceList.push(["deny-list"]);
 		}
 	};
-	for (const { id, entries } of results) {
-		const meta = dataModule.DICTIONARY_META[id];
-		if (!meta) continue;
-		for (const entry of entries) addDenyListEntry(entry, meta.label);
-	}
-	if (!excludeCategories.has("Places")) {
-		const cityCountries = allowedCountries !== null ? [...allowedCountries] : [
-			"AT",
-			"AU",
-			"BE",
-			"BG",
-			"BR",
-			"CA",
-			"CH",
-			"CZ",
-			"DE",
-			"DK",
-			"ES",
-			"FI",
-			"FR",
-			"GB",
-			"GR",
-			"HR",
-			"HU",
-			"IE",
-			"IT",
-			"LU",
-			"NL",
-			"NO",
-			"NZ",
-			"PL",
-			"PT",
-			"RO",
-			"SE",
-			"SI",
-			"SK",
-			"US"
-		];
-		const cityEntries = await dataModule.loadCityDictionaries(cityCountries);
-		for (const entry of cityEntries) addDenyListEntry(entry, "address");
+	if (hasDenyList) {
+		const denyListData = dictionaries.denyList;
+		const metaData = dictionaries.denyListMeta;
+		for (const [id, entries] of Object.entries(denyListData)) {
+			const meta = metaData[id];
+			if (!meta) continue;
+			if (!config.enableNameCorpus && meta.category === "Names") continue;
+			if (excludeCategories.has(meta.category)) continue;
+			if (allowedCountries !== null && meta.country !== null) {
+				if (!allowedCountries.has(meta.country)) continue;
+			}
+			for (const entry of entries) addDenyListEntry(entry, meta.label);
+		}
 	}
+	if (hasCities && !excludeCategories.has("Places")) for (const entry of dictionaries.cities) addDenyListEntry(entry, "address");
+	appendNameCorpusEntries(config, ctx, patternList, labelList, sourceList, patternIndex);
+	if (patternList.length === 0) return null;
+	return {
+		labels: labelList,
+		originals: patternList,
+		sources: sourceList
+	};
+};
+/**
+* Build deny-list data containing only name corpus
+* entries (no external dictionaries). Used when no
+* dictionary data is injected but name corpus is
+* enabled.
+*/
+const buildNameCorpusOnly = (config, ctx) => {
+	if (!config.enableNameCorpus) return null;
+	const excluded = config.denyListExcludeCategories;
+	if ((excluded ? new Set(excluded) : /* @__PURE__ */ new Set()).has("Names")) return null;
+	const patternList = [];
+	const labelList = [];
+	const sourceList = [];
+	appendNameCorpusEntries(config, ctx, patternList, labelList, sourceList, /* @__PURE__ */ new Map());
+	if (patternList.length === 0) return null;
+	return {
+		labels: labelList,
+		originals: patternList,
+		sources: sourceList
+	};
+};
+/**
+* Append name corpus entries (first names, surnames,
+* titles) to the pattern arrays. Shared between
+* buildDenyList and buildNameCorpusOnly.
+*/
+const appendNameCorpusEntries = (config, ctx, patternList, labelList, sourceList, patternIndex) => {
+	const excluded = config.denyListExcludeCategories;
+	const excludeCategories = excluded ? new Set(excluded) : /* @__PURE__ */ new Set();
+	if (!config.enableNameCorpus || excludeCategories.has("Names")) return;
 	const addNameEntry = (name, source) => {
 		const normalized = normalizeForSearch(name).replace(/[|\\]/g, "");
 		if (normalized.length === 0) return;
@@ -2894,30 +3357,22 @@ const buildDenyList = async (config, ctx = defaultContext) => {
 			sourceList.push([source]);
 		}
 	};
-	if (config.enableNameCorpus && !excludeCategories.has("Names")) {
-		for (const name of getNameCorpusFirstNames(ctx)) addNameEntry(name, "first-name");
-		for (const name of getNameCorpusSurnames(ctx)) addNameEntry(name, "surname");
-		for (const title of getNameCorpusTitles(ctx)) {
-			const norm = normalizeForSearch(title).replace(/[|\\]/g, "");
-			if (norm.length === 0) continue;
-			const lower = norm.toLowerCase();
-			const existing = patternIndex.get(lower);
-			if (existing !== void 0) {
-				if (!sourceList[existing].includes("title")) sourceList[existing].push("title");
-			} else {
-				patternIndex.set(lower, patternList.length);
-				patternList.push(norm);
-				labelList.push(["person"]);
-				sourceList.push(["title"]);
-			}
+	for (const name of getNameCorpusFirstNames(ctx)) addNameEntry(name, "first-name");
+	for (const name of getNameCorpusSurnames(ctx)) addNameEntry(name, "surname");
+	for (const title of getNameCorpusTitles(ctx)) {
+		const norm = normalizeForSearch(title).replace(/[|\\]/g, "");
+		if (norm.length === 0) continue;
+		const lower = norm.toLowerCase();
+		const existing = patternIndex.get(lower);
+		if (existing !== void 0) {
+			if (!sourceList[existing].includes("title")) sourceList[existing].push("title");
+		} else {
+			patternIndex.set(lower, patternList.length);
+			patternList.push(norm);
+			labelList.push(["person"]);
+			sourceList.push(["title"]);
 		}
 	}
-	if (patternList.length === 0) return null;
-	return {
-		labels: labelList,
-		originals: patternList,
-		sources: sourceList
-	};
 };
 /**
 * Ensure all deny-list support data (stopwords, allow
@@ -2927,8 +3382,8 @@ const buildDenyList = async (config, ctx = defaultContext) => {
 * the search instance was built on a different context
 * (e.g. cachedSearch).
 */
-const ensureDenyListData = async (ctx = defaultContext) => {
-	await initNameCorpus(ctx);
+const ensureDenyListData = async (ctx = defaultContext, dictionaries) => {
+	await initNameCorpus(ctx, dictionaries);
 	await Promise.all([
 		loadStopwords(ctx),
 		loadAllowList(ctx),
@@ -3137,7 +3592,7 @@ const extendPersonName = (text, start, end, ctx) => {
 let cachedBoundaryRe = null;
 const loadBoundaryWords = async () => {
 	try {
-		return (await import("@stll/anonymize-data/config/address-boundaries.json")).default;
+		return (await import("./address-boundaries.mjs")).default;
 	} catch {
 		return {};
 	}
@@ -3166,7 +3621,7 @@ const getBoundaryRe = async () => {
 const buildStreetTypePatterns = async () => {
 	let config = {};
 	try {
-		config = (await import("@stll/anonymize-data/config/address-street-types.json")).default;
+		config = (await import("./address-street-types.mjs")).default;
 	} catch {
 		return [];
 	}
@@ -3494,7 +3949,7 @@ let _temporalPreps = null;
 let _prepsPromise = null;
 const loadPrepositions = async () => {
 	try {
-		const mod = await import("@stll/anonymize-data/config/address-prepositions.json");
+		const mod = await import("./address-prepositions.mjs");
 		const data = mod.default ?? mod;
 		const addr = /* @__PURE__ */ new Set();
 		const temp = /* @__PURE__ */ new Set();
@@ -3518,7 +3973,7 @@ let _streetAbbrevs = null;
 let _streetAbbrevsPromise = null;
 const loadStreetAbbrevs = async () => {
 	try {
-		const mod = await import("@stll/anonymize-data/config/address-street-types.json");
+		const mod = await import("./address-street-types.mjs");
 		const data = mod.default ?? mod;
 		const abbrevs = /* @__PURE__ */ new Set();
 		for (const [key, words] of Object.entries(data)) {
@@ -3687,11 +4142,11 @@ const ZONE_SCORE_ADJUSTMENTS = {
 	table: .05
 };
 const loadSectionHeadings = async () => {
-	const mod = await import("@stll/anonymize-data/config/section-headings.json");
+	const mod = await import("./section-headings.mjs");
 	return (mod.default ?? mod).patterns.map((p) => new RegExp(p.re, p.flags));
 };
 const loadSigningClauses = async () => {
-	const mod = await import("@stll/anonymize-data/config/signing-clauses.json");
+	const mod = await import("./signing-clauses.mjs");
 	return (mod.default ?? mod).patterns.map((p) => {
 		const prefix = p.prefix || "";
 		const suffix = p.suffix || "";
@@ -3875,7 +4330,7 @@ let search = null;
 let patternToRule = null;
 let initPromise = null;
 const loadRules = async () => {
-	const mod = await import("@stll/anonymize-data/config/hotword-rules.json");
+	const mod = await import("./hotword-rules.mjs");
 	const loaded = (mod.default ?? mod).rules;
 	const patterns = [];
 	const mapping = [];
@@ -4575,7 +5030,7 @@ let amountWordsLoaded = false;
 const getAmountWordsRe = async () => {
 	if (amountWordsLoaded && amountWordsRe) return amountWordsRe;
 	try {
-		const alt = (await import("@stll/anonymize-data/config/amount-words.json")).default.patterns.flatMap((p) => p.keywords).map(escapeRegex).join("|");
+		const alt = (await import("./amount-words.mjs")).default.patterns.flatMap((p) => p.keywords).map(escapeRegex).join("|");
 		amountWordsRe = new RegExp(`^[,;]?[^\\S\\n]*(\\((?:${alt})[:\\s][^)\\n]{1,120}\\))`, "i");
 	} catch {
 		amountWordsRe = /^[,;]?[^\S\n]*(\((?:slovy|slovně)[:\s][^)\n]{1,120}\))/i;
@@ -4684,7 +5139,7 @@ const runPipeline = async (options) => {
 		initStreetAbbrevs(),
 		hotwordInit
 	]);
-	if (cachedSearch && config.enableDenyList) await ensureDenyListData(ctx);
+	if (cachedSearch && config.enableDenyList) await ensureDenyListData(ctx, config.dictionaries);
 	let zones = [];
 	if (config.enableZoneClassification && zoneInitOk) {
 		zones = classifyZones(fullText, ctx);
@@ -4710,7 +5165,7 @@ const runPipeline = async (options) => {
 	let rawNameCorpusEntities = [];
 	let nameCorpusEntities = [];
 	if (config.enableNameCorpus && !config.enableDenyList) {
-		await initNameCorpus(ctx);
+		await initNameCorpus(ctx, config.dictionaries);
 		checkAbort(signal);
 		rawNameCorpusEntities = detectNameCorpus(fullText, ctx);
 		nameCorpusEntities = filterAllowedLabels(rawNameCorpusEntities, preHotwordAllowedLabels);