mask-privacy 3.0.0 → 3.2.0

package/dist/index.mjs

@@ -231,10 +231,13 @@ var init_exceptions = __esm({
 function looksLikeToken(value) {
   if (typeof value !== "string") return false;
   const v7 = value.trim();
-  if (v7.startsWith("tkn-") && v7.includes("@email.com")) {
-    return true;
+  if (v7.startsWith("tkn-") && v7.includes("@")) {
+    const parts = v7.split("@");
+    if (parts.length === 2 && parts[0].length >= 12 && parts[1].includes(".")) {
+      return true;
+    }
   }
-  if (v7.startsWith("+1-555-") && v7.length === 14) {
+  if (/^\+[1-9]\d{0,3}-555-\d{7}$/.test(v7)) {
     return true;
   }
   if (v7.startsWith("000-00-") && v7.length === 11) {
@@ -246,16 +249,13 @@ function looksLikeToken(value) {
   if (v7.startsWith("000000") && v7.length === 9) {
     return true;
   }
-  if (v7.startsWith("784-0000-") && v7.length === 18) {
-    return true;
-  }
-  if (v7.length === 11 && v7.startsWith("990000") && /^\d+$/.test(v7) && parseInt(v7[v7.length - 1], 10) % 2 === 0) {
+  if (v7.length === 9 && v7.startsWith("000") && /[A-Z]$/.test(v7)) {
     return true;
   }
-  if (v7.length === 10 && v7.startsWith("100000") && /^\d+$/.test(v7)) {
+  if (/^[A-Z]{2}00[A-F0-9]{4,16}$/.test(v7)) {
     return true;
   }
-  if (/^[A-Z]{2}00[A-F0-9]{4,16}$/.test(v7)) {
+  if (/^<(PER|LOC|ORG):[^>]+>$/.test(v7)) {
     return true;
   }
   if (v7.startsWith("[TKN-") && v7.endsWith("]")) {
@@ -267,7 +267,7 @@ var TOKEN_PATTERN;
 var init_fpe_utils = __esm({
   "src/core/fpe_utils.ts"() {
     TOKEN_PATTERN = new RegExp(
-      "tkn-[a-f0-9]{8,64}@email\\.com|\\+1-555-\\d{7}|000-00-\\d{4}|4000-0000-0000-\\d{4}|000000\\d{3}|990000\\d{4}[02468]|100000\\d{4}|784-0000-\\d{7}-\\d|[A-Z]{2}00[A-F0-9]{4,16}|\\[TKN-[a-f0-9]{8,64}\\]",
+      "tkn-[a-f0-9]{8,64}@[A-Za-z0-9.\\-]+\\.[A-Za-z]{2,}|\\+[1-9]\\d{0,3}-555-\\d{7}|000-00-\\d{4}|4000-0000-0000-\\d{4}|000000\\d{3}|000\\d{5}[A-Z]|[A-Z]{2}00[A-F0-9]{4,16}|<(?:PER|LOC|ORG):[^>]+>|\\[TKN-[a-f0-9]{8,64}\\]",
       // Opaque
       "g"
     );
@@ -316,42 +316,87 @@ async function _hmacDigits(plaintext, n6, offset = 0) {
   }
   return result.join("");
 }
-async function generateFPEToken(rawText) {
-  const text = rawText.trim();
-  if (_EMAIL_RE.test(text)) {
-    return `tkn-${await _hmacHex(text)}@email.com`;
-  }
-  if (_PHONE_RE.test(text)) {
-    return `+1-555-${await _hmacDigits(text, 7)}`;
+async function _pickFromArray(plaintext, array) {
+  const digits = await _hmacDigits(plaintext, 8);
+  const num = parseInt(digits, 10);
+  return array[num % array.length];
+}
+function _computeLuhnDigit(partialNum) {
+  const digits = partialNum.split("").map(Number);
+  let sum = 0;
+  let shouldDouble = true;
+  for (let i6 = digits.length - 1; i6 >= 0; i6--) {
+    let digit = digits[i6];
+    if (shouldDouble) {
+      digit *= 2;
+      if (digit > 9) digit -= 9;
+    }
+    sum += digit;
+    shouldDouble = !shouldDouble;
   }
-  if (_SSN_RE.test(text)) {
+  return ((10 - sum % 10) % 10).toString();
+}
+function _computeEsIdCheck(num) {
+  return "TRWAGMYFPDXBNJZSQVHLCKE"[num % 23];
+}
+async function generateFPEToken(rawText, entityType = "UNKNOWN") {
+  const text = rawText.trim();
+  let type = (entityType || "UNKNOWN").toUpperCase();
+  if (type === "UNKNOWN") {
+    if (_EMAIL_RE.test(text)) type = "EMAIL_ADDRESS";
+    else if (_SSN_RE.test(text)) type = "US_SSN";
+    else if (_CC_RE.test(text)) type = "CREDIT_CARD";
+    else if (_ROUTING_RE.test(text)) type = "US_ROUTING_NUMBER";
+    else if (_ES_ID_RE.test(text)) type = "ES_DNI";
+    else if (_IBAN_RE.test(text)) type = "INTL_BANK_IBAN";
+    else if (_PHONE_RE.test(text)) type = "PHONE_NUMBER";
+  }
+  if (type === "EMAIL_ADDRESS" || type === "EMAIL_ADDR") {
+    const parts = text.split("@");
+    const domain = parts.length === 2 ? parts[1] : "email.com";
+    return `tkn-${await _hmacHex(text)}@${domain}`;
+  }
+  if (type === "PHONE_NUMBER" || type === "PHONE_NUM" || type === "PHONE_NUM_INTL") {
+    const m6 = text.match(/^\+([1-9]\d{0,3})/);
+    const cc = m6 ? m6[1] : "1";
+    return `+${cc}-555-${await _hmacDigits(text, 7)}`;
+  }
+  if (type === "US_SSN") {
     return `000-00-${await _hmacDigits(text, 4)}`;
   }
-  if (_CC_RE.test(text)) {
-    return `4000-0000-0000-${await _hmacDigits(text, 4)}`;
+  if (type === "CREDIT_CARD" || type === "CREDIT_CARD_NUMBER") {
+    const base = `400000000000${await _hmacDigits(text, 3)}`;
+    const checkDig = _computeLuhnDigit(base);
+    const full = base + checkDig;
+    return `${full.slice(0, 4)}-${full.slice(4, 8)}-${full.slice(8, 12)}-${full.slice(12, 16)}`;
   }
-  if (_ROUTING_RE.test(text)) {
+  if (type === "US_ROUTING_NUMBER" || type === "US_ABA_ROUTING") {
     return `000000${await _hmacDigits(text, 3)}`;
   }
-  if (_TCID_RE.test(text)) {
-    const tail = await _hmacDigits(text, 5);
-    let lastD = parseInt(tail[tail.length - 1], 10);
-    if (lastD % 2 !== 0) lastD = (lastD + 1) % 10;
-    return `990000${tail.slice(0, 4)}${lastD}`;
+  if (type === "INTL_BANK_IBAN" || type === "IBAN_CODE") {
+    const countryCode = text.length >= 2 && /[a-zA-Z]{2}/.test(text.slice(0, 2)) ? text.slice(0, 2).toUpperCase() : "US";
+    return `${countryCode}00${(await _hmacHex(text, 8)).toUpperCase()}`;
+  }
+  if (type === "ES_DNI") {
+    const digits = `000${await _hmacDigits(text, 5)}`;
+    return digits + _computeEsIdCheck(parseInt(digits, 10));
   }
-  if (_SAUDI_NID_RE.test(text)) {
-    return `100000${await _hmacDigits(text, 4)}`;
+  if (type === "PERSON" || type === "PERSON_NAME") {
+    const f6 = await _pickFromArray(text, _FIRST_NAMES);
+    const l6 = await _pickFromArray(text + "last", _LAST_NAMES);
+    return `<PER:${f6}_${l6}>`;
   }
-  if (_UAE_EID_RE.test(text)) {
-    return `784-0000-${await _hmacDigits(text, 7)}-${await _hmacDigits(text, 1, 20)}`;
+  if (type === "LOCATION" || type === "PHYS_ADDRESS") {
+    const c6 = await _pickFromArray(text, _CITIES);
+    return `<LOC:${c6}>`;
   }
-  if (_IBAN_RE.test(text)) {
-    const countryCode = text.slice(0, 2);
-    return `${countryCode}00${(await _hmacHex(text, 8)).toUpperCase()}`;
+  if (type === "ORGANIZATION") {
+    const c6 = await _pickFromArray(text, _LAST_NAMES);
+    return `<ORG:${c6}_Inc>`;
   }
   return `[TKN-${await _hmacHex(text)}]`;
 }
-var _masterKey, _EMAIL_RE, _PHONE_RE, _SSN_RE, _CC_RE, _ROUTING_RE, _TCID_RE, _SAUDI_NID_RE, _UAE_EID_RE, _IBAN_RE;
+var _masterKey, _EMAIL_RE, _PHONE_RE, _SSN_RE, _CC_RE, _ROUTING_RE, _ES_ID_RE, _IBAN_RE, _FIRST_NAMES, _LAST_NAMES, _CITIES;
 var init_fpe = __esm({
   "src/core/fpe.ts"() {
     init_config();
@@ -360,14 +405,15 @@ var init_fpe = __esm({
     init_fpe_utils();
     _masterKey = null;
     _EMAIL_RE = /^[^@\s]+@[^@\s]+\.[^@\s]+$/;
-    _PHONE_RE = /^\+?1?[\s\-.]?\(?\d{3}\)?[\s\-.]?\d{3}[\s\-.]?\d{4}$|^\d{3}[\s\-.]?\d{4}$/;
+    _PHONE_RE = /(?<!\d)(?:\+?1?[\s\-.]?\(?\d{3}\)?[\s\-.]?\d{3}[\s\-.]?\d{4}|\d{3}[\s\-.]?\d{4})(?!\d)/;
     _SSN_RE = /^\d{3}-\d{2}-\d{4}$/;
     _CC_RE = /^(?:\d{4}[ \-]?){3}\d{4}$/;
     _ROUTING_RE = /^\d{9}$/;
-    _TCID_RE = /^[1-9]\d{9}[02468]$/;
-    _SAUDI_NID_RE = /^1\d{9}$/;
-    _UAE_EID_RE = /^784-\d{4}-\d{7}-\d$/;
+    _ES_ID_RE = /^(?:\d{8}[A-Z]|[XYZ]\d{7}[A-Z])$/;
     _IBAN_RE = /^[A-Z]{2}\d{2}[A-Z0-9]{4,30}$/;
+    _FIRST_NAMES = ["Taylor", "Jordan", "Casey", "Morgan", "Riley", "Avery", "Rowan", "Quinn", "Charlie", "Peyton", "Blake", "Dakota", "Reese", "Skyler", "Finley", "Eden", "Harley", "Rory", "Emerson", "Remi"];
+    _LAST_NAMES = ["Smith", "Johnson", "Williams", "Brown", "Jones", "Garcia", "Miller", "Davis", "Rodriguez", "Martinez", "Hernandez", "Lopez", "Gonzalez", "Wilson", "Anderson", "Thomas", "Taylor", "Moore", "Jackson", "Martin"];
+    _CITIES = ["London", "Paris", "Berlin", "Tokyo", "Rome", "Madrid", "Vienna", "Sydney", "Toronto", "Chicago", "Seattle", "Austin", "Boston", "Denver", "Dallas", "Miami", "Seoul", "Dubai", "Mumbai", "Cairo"];
   }
 });
 
@@ -42968,7 +43014,7 @@ async function encode(rawText, options = {}) {
       return existingToken;
     }
   }
-  const token = await generateFPEToken(text);
+  const token = await generateFPEToken(text, options.entityType || "UNKNOWN");
   const ciphertext = cryptoEngine.encrypt(text);
   const ttl = options.ttl || DEFAULT_TTL;
   await vault.store(token, ciphertext, ttl, ptHash);
@@ -43420,19 +43466,8 @@ var SCRIPT_SIGNATURES, LanguageContextResolver;
 var init_assessor = __esm({
   "src/core/dlp/assessor.ts"() {
     SCRIPT_SIGNATURES = [
-      // CJK / East-Asian — checked first because they are unambiguous
-      { tag: "zh", pattern: /[\u4e00-\u9fff\u3400-\u4dbf]/g },
-      { tag: "ja", pattern: /[\u3040-\u309f\u30a0-\u30ff\u31f0-\u31ff]/g },
-      // Arabic script — covers Standard Arabic, Urdu overlap, etc.
-      { tag: "ar", pattern: /[\u0600-\u06ff\u0750-\u077f\u08a0-\u08ff\ufb50-\ufdff\ufe70-\ufeff]/g },
-      // Turkish — distinguished by dotless-i (ı), soft-g (ğ), ş, and cedilla ç
-      { tag: "tr", pattern: /[ğıİşŞ]/g },
-      // German — umlauts and Eszett
-      { tag: "de", pattern: /[äöüÄÖÜß]/g },
       // Spanish — ñ and inverted punctuation
-      { tag: "es", pattern: /[ñÑ¡¿]/g },
-      // French — cedilla, accented vowels with circumflex / diaeresis
-      { tag: "fr", pattern: /[àâçéèêëïîôùûüÿœæ]/gi }
+      { tag: "es", pattern: /[ñÑ¡¿]/g }
     ];
     LanguageContextResolver = class {
       constructor(charThreshold = 1) {
@@ -43490,34 +43525,12 @@ var init_registry = __esm({
     })(SensitiveCategory || {});
     LOCALE_NAME_RULES = {
       en: [
-        /\b[A-Z][a-z]+ [A-Z][a-z]+(?:\s+[A-Z][a-z]+)?\b/g,
-        /\b(?:Mr|Mrs|Ms|Dr|Prof)\.?\s+[A-Z][a-z]+\b/g
+        /\b[A-Z][a-z\-\']+ [A-Z][a-z\-\']+(?:\s+[A-Z][a-z\-\']+)?\b/g,
+        /\b(?:Mr|Mrs|Ms|Dr|Prof)\.?\s+[A-Z][a-z\-\']+\b/g
       ],
       es: [
-        /\b[A-Z][a-záéíóúñ]+ [A-Z][a-záéíóúñ]+(?:\s+[A-Z][a-záéíóúñ]+)?\b/g,
-        /\b(?:Sr|Sra|Srta)\.?\s+[A-Z][a-záéíóúñ]+\b/g
-      ],
-      fr: [
-        /\b[A-Z][a-zàâçéèêëïîôùûü]+ [A-Z][a-zàâçéèêëïîôùûü]+\b/g,
-        /\b(?:M|Mme|Mlle)\.?\s+[A-Z][a-zàâçéèêëïîôùûü]+\b/g
-      ],
-      de: [
-        /\b[A-Z][a-zäöüß]+ [A-Z][a-zäöüß]+\b/g,
-        /\b(?:Herr|Frau)\.?\s+[A-Z][a-zäöüß]+\b/g
-      ],
-      tr: [
-        /\b[A-ZÇĞİÖŞÜ][a-zçğıöşü]+ [A-ZÇĞİÖŞÜ][a-zçğıöşü]+\b/g,
-        /\b(?:Bay|Bayan|Sayın)\.?\s+[A-ZÇĞİÖŞÜ][a-zçğıöşü]+\b/g
-      ],
-      ar: [
-        /[\u0621-\u064a][\u0600-\u06ff]+ [\u0621-\u064a][\u0600-\u06ff]+/g,
-        /(?:أبو|أم|ابن|بنت)\s+[\u0621-\u064a][\u0600-\u06ff]+/gi
-      ],
-      ja: [
-        /\b[A-Z][a-z]+(?:moto|yama|kawa|mura|ta|da|shi|no)\s+[A-Z][a-z]+\b/g
-      ],
-      zh: [
-        /\b[A-Z][a-z]{1,3}\s+[A-Z][a-z]+\b/g
+        /\b[A-Z][a-záéíóúñ\-\']+ [A-Z][a-záéíóúñ\-\']+(?:\s+[A-Z][a-záéíóúñ\-\']+)?\b/g,
+        /\b(?:Sr|Sra|Srta)\.?\s+[A-Z][a-záéíóúñ\-\']+\b/g
       ]
     };
     LOCALE_ADDRESS_RULES = {
@@ -43525,26 +43538,8 @@ var init_registry = __esm({
         /\b\d{1,5}\s+[A-Z][a-z]+(?:\s+[A-Z][a-z]+)*\s+(?:Street|St|Avenue|Ave|Road|Rd|Boulevard|Blvd|Lane|Ln|Drive|Dr|Court|Ct|Way)\b/g,
         /\b[A-Z][a-z]+(?:\s+[A-Z][a-z]+)*,\s*[A-Z]{2}\s+\d{5}(?:-\d{4})?\b/g
       ],
-      fr: [
-        /\b\d{1,4}\s+(?:rue|avenue|boulevard|place|chemin)\s+[A-ZÀ-ÖØ-Ý][a-zà-öø-ÿ]+\b/gi
-      ],
-      de: [
-        /\b[A-ZÄÖÜa-zäöüß]+(?:straße|strasse|weg|gasse|platz)\s+\d{1,4}\b/g
-      ],
-      tr: [
-        /\b[A-ZÇĞİÖŞÜa-zçğıöşü]+\s+(?:Cad|Sok|Mah)\.?\s+/gi,
-        /\b\d{5}\s+[A-ZÇĞİÖŞÜa-zçğıöşü]+\/[A-ZÇĞİÖŞÜa-zçğıöşü]+\b/g
-      ],
-      ar: [
-        /شارع\s+[\u0600-\u06ff]+/g,
-        /حي\s+[\u0600-\u06ff]+/g,
-        /(?:ص\.ب|P\.?O\.?\s*Box)\s*\d{3,6}/gi
-      ],
-      uk_postcode: [
-        /\b[A-Z]{1,2}\d{1,2}[A-Z]?\s*\d[A-Z]{2}\b/g
-      ],
-      ca_postal: [
-        /\b[A-Z]\d[A-Z]\s*\d[A-Z]\d\b/g
+      es: [
+        /\b(?:Calle|Carrera|Avenida|Paseo|Plaza)\s+[A-ZÀ-ÖØ-Ý][a-zà-öø-ÿ]+\b/gi
       ]
     };
     RAW_PATTERNS = [
@@ -43552,7 +43547,6 @@ var init_registry = __esm({
       [
         "US_SSN",
         "\\b(?!000|666|9\\d{2})\\d{3}-(?!00)\\d{2}-(?!0000)\\d{4}\\b",
-        "g",
         ["ssn", "social security", "tax id", "taxpayer"],
         0.95,
         "FINANCIAL" /* FINANCIAL */,
@@ -43561,7 +43555,6 @@ var init_registry = __esm({
       [
         "CREDIT_CARD_NUMBER",
         "\\b(?:4\\d{3}|5[1-5]\\d{2}|3[47]\\d{2}|6(?:011|5\\d{2}))[- ]?\\d{4}[- ]?\\d{4}[- ]?\\d{4}\\b",
-        "g",
         ["card", "credit", "visa", "mastercard", "amex", "payment"],
         0.97,
         "FINANCIAL" /* FINANCIAL */,
@@ -43570,7 +43563,6 @@ var init_registry = __esm({
       [
         "INTL_BANK_IBAN",
         "\\b[A-Z]{2}\\d{2}[A-Z0-9]{4}\\d{7}[A-Z0-9]{0,16}\\b",
-        "g",
         ["iban", "swift", "sepa", "wire", "bank transfer"],
         0.96,
         "FINANCIAL" /* FINANCIAL */,
@@ -43579,7 +43571,6 @@ var init_registry = __esm({
       [
         "CRYPTO_BTC",
         "\\b(?:[13][a-km-zA-HJ-NP-Z1-9]{25,34}|bc1[a-z0-9]{39,59})\\b",
-        "g",
         ["bitcoin", "btc", "wallet", "crypto"],
         0.94,
         "FINANCIAL" /* FINANCIAL */,
@@ -43588,7 +43579,6 @@ var init_registry = __esm({
       [
         "CRYPTO_ETH",
         "\\b0x[a-fA-F0-9]{40}\\b",
-        "g",
         ["ethereum", "eth", "wallet", "0x"],
         0.93,
         "FINANCIAL" /* FINANCIAL */,
@@ -43596,8 +43586,7 @@ var init_registry = __esm({
       ],
       [
         "US_ABA_ROUTING",
-        "\\b\\d{9}\\b",
-        "g",
+        /(?<!\d)\d{9}(?!\d)/,
         ["routing", "aba", "wire", "bank"],
         0.88,
         "FINANCIAL" /* FINANCIAL */,
@@ -43605,17 +43594,15 @@ var init_registry = __esm({
       ],
       [
         "BANK_ACCT_NUM",
-        "\\b\\d{8,17}\\b",
-        "g",
+        /(?<!\d)\d{8,17}(?!\d)/,
         ["account", "checking", "savings", "deposit", "bank"],
-        0.83,
+        0.5,
         "FINANCIAL" /* FINANCIAL */,
-        null
+        "luhn_soft"
       ],
       [
         "SWIFT_BIC",
         "\\b[A-Z]{4}[A-Z]{2}[A-Z0-9]{2}(?:[A-Z0-9]{3})?\\b",
-        "gi",
         ["swift", "bic", "bank code", "transfer"],
         0.6,
         "FINANCIAL" /* FINANCIAL */,
@@ -43625,7 +43612,6 @@ var init_registry = __esm({
       [
         "EMAIL_ADDR",
         "\\b[A-Za-z0-9._%+\\-]+@[A-Za-z0-9.\\-]+\\.[A-Za-z]{2,}\\b",
-        "g",
         ["email", "mail", "contact", "address"],
         0.99,
         "CONTACT" /* CONTACT */,
@@ -43633,26 +43619,23 @@ var init_registry = __esm({
       ],
       [
         "PHONE_NUM",
-        "(?:\\+?[1-9]\\d{0,3}[-.\\s]?)?\\(?\\d{1,4}\\)?[-.\\s]?\\d{1,4}[-.\\s]?\\d{1,4}[-.\\s]?\\d{1,9}",
-        "g",
+        /(?<!\d)(?:\+?[1-9]\d{0,3}[-.\s]?)?\(?\d{1,4}\)?[-.\s]?\d{1,4}[-.\s]?\d{1,4}[-.\s]?\d{1,9}(?!\d)/,
         ["phone", "call", "mobile", "tel", "whatsapp", "number"],
-        0.92,
+        0.8,
         "CONTACT" /* CONTACT */,
         null
       ],
       [
         "PHONE_NUM_INTL",
-        "\\+(?:44|33|49|90|966|971)[-.\\s]?\\(?\\d{1,5}\\)?(?:[-.\\s]?\\d{2,4}){2,4}",
-        "g",
+        /(?<!\d)\+(?:[1-9]\d{0,3})[-.\s]?\(?\d{1,5}\)?(?:[-.\s]?\d{2,4}){2,4}(?!\d)/,
         ["phone", "call", "mobile", "tel"],
-        0.93,
+        0.8,
         "CONTACT" /* CONTACT */,
         null
       ],
       [
         "IPV4_ADDR",
         "\\b(?:(?:25[0-5]|2[0-4]\\d|[01]?\\d\\d?)\\.){3}(?:25[0-5]|2[0-4]\\d|[01]?\\d\\d?)\\b",
-        "g",
         ["ip", "server", "host", "network", "address"],
         0.94,
         "CONTACT" /* CONTACT */,
@@ -43661,7 +43644,6 @@ var init_registry = __esm({
       [
         "IPV6_ADDR",
         "\\b(?:[A-Fa-f0-9]{1,4}:){7}[A-Fa-f0-9]{1,4}\\b",
-        "g",
         ["ipv6", "ip", "network", "server"],
         0.93,
         "CONTACT" /* CONTACT */,
@@ -43670,7 +43652,6 @@ var init_registry = __esm({
       [
         "HW_MAC_ADDR",
         "\\b(?:[0-9A-Fa-f]{2}[:-]){5}[0-9A-Fa-f]{2}\\b",
-        "g",
         ["mac", "hardware", "network", "device"],
         0.91,
         "CONTACT" /* CONTACT */,
@@ -43680,7 +43661,6 @@ var init_registry = __esm({
       [
         "BIRTH_DATE",
         "\\b(?:0[1-9]|1[0-2])[/-](?:0[1-9]|[12]\\d|3[01])[/-](?:19|20)\\d{2}\\b",
-        "g",
         ["birth", "dob", "born", "birthday", "date of birth"],
         0.88,
         "PERSONAL" /* PERSONAL */,
@@ -43689,16 +43669,14 @@ var init_registry = __esm({
       [
         "US_DRIVERS_LIC",
         "\\b(?:[A-Z]\\d{7,12}|\\d{7,12}[A-Z]?)\\b",
-        "g",
         ["driver", "license", "licence", "dl", "dmv"],
-        0.85,
+        0.55,
         "PERSONAL" /* PERSONAL */,
         null
       ],
       [
         "US_PASSPORT_NUM",
         "\\b[A-Z]\\d{8}\\b",
-        "g",
         ["passport", "travel", "visa", "immigration"],
         0.87,
         "PERSONAL" /* PERSONAL */,
@@ -43708,7 +43686,6 @@ var init_registry = __esm({
       [
         "VEHICLE_VIN",
         "\\b[A-HJ-NPR-Z0-9]{17}\\b",
-        "g",
         ["vin", "vehicle", "chassis", "automobile"],
         0.92,
         "VEHICLE" /* VEHICLE */,
@@ -43717,7 +43694,6 @@ var init_registry = __esm({
       [
         "VEHICLE_PLATE",
         "\\b[A-Z0-9]{1,3}[\\-\\s][A-Z0-9]{1,4}[\\-\\s][A-Z0-9]{1,4}\\b",
-        "g",
         ["plate", "registration", "vehicle", "plaka"],
         0.45,
         "VEHICLE" /* VEHICLE */,
@@ -43727,7 +43703,6 @@ var init_registry = __esm({
       [
         "MED_RECORD_ID",
         "\\b(?:MRN|Patient ID|Medical Record)[:\\s]*[A-Z0-9]{6,10}\\b",
-        "g",
         ["patient", "medical", "record", "mrn", "hospital"],
         0.96,
         "HEALTHCARE" /* HEALTHCARE */,
@@ -43736,7 +43711,6 @@ var init_registry = __esm({
       [
         "US_MEDICARE_ID",
         "\\b\\d{3}-\\d{2}-\\d{4}[A-Z]\\b",
-        "g",
         ["medicare", "cms", "beneficiary", "health insurance"],
         0.91,
         "HEALTHCARE" /* HEALTHCARE */,
@@ -43745,7 +43719,6 @@ var init_registry = __esm({
       [
         "US_DEA_NUM",
         "\\b[A-Z]{2}\\d{7}\\b",
-        "g",
         ["dea", "prescriber", "drug", "enforcement"],
         0.89,
         "HEALTHCARE" /* HEALTHCARE */,
@@ -43754,7 +43727,6 @@ var init_registry = __esm({
       [
         "US_NPI_NUM",
         "\\b\\d{10}\\b",
-        "g",
         ["npi", "provider", "national provider", "healthcare"],
         0.87,
         "HEALTHCARE" /* HEALTHCARE */,
@@ -43764,7 +43736,6 @@ var init_registry = __esm({
       [
         "US_EIN_TAX",
         "\\b\\d{2}-\\d{7}\\b",
-        "g",
         ["ein", "federal", "employer", "tax id"],
         0.89,
         "IDENTITY_US" /* IDENTITY_US */,
@@ -43774,71 +43745,33 @@ var init_registry = __esm({
       [
         "UK_NATL_INS",
         "\\b[A-Z]{2}\\d{6}[A-Z]\\b",
-        "g",
         ["nino", "national insurance", "ni number", "uk"],
         0.9,
         "IDENTITY_INTL" /* IDENTITY_INTL */,
-        null
+        "uk_nino"
       ],
       [
         "CA_SOCIAL_INS",
         "\\b\\d{3}[-\\s]?\\d{3}[-\\s]?\\d{3}\\b",
-        "g",
         ["sin", "social insurance", "canada", "canadian"],
         0.89,
         "IDENTITY_INTL" /* IDENTITY_INTL */,
-        null
+        "ca_sin"
       ],
       [
-        "FR_INSEE_NUM",
-        "\\b[12]\\d{2}[01]\\d\\d{8}\\d{2}\\b",
-        "g",
-        ["insee", "s\xE9curit\xE9 sociale", "france", "num\xE9ro"],
-        0.88,
-        "IDENTITY_INTL" /* IDENTITY_INTL */,
-        null
-      ],
-      [
-        "DE_STEUER_ID",
-        "\\b\\d{2}\\s?\\d{3}\\s?\\d{3}\\s?\\d{3}\\b",
-        "g",
-        ["steuer", "steuernummer", "finanzamt", "deutschland"],
-        0.87,
-        "IDENTITY_INTL" /* IDENTITY_INTL */,
-        null
-      ],
-      [
-        "TR_TCID",
-        "\\b[1-9]\\d{9}[02468]\\b",
-        "g",
-        ["tc", "kimlik", "vatanda\u015Fl\u0131k", "n\xFCfus", "t\xFCrkiye"],
-        0.92,
-        "IDENTITY_INTL" /* IDENTITY_INTL */,
-        "tcid"
-      ],
-      [
-        "SA_NATIONAL_ID",
-        "\\b1\\d{9}\\b",
-        "g",
-        ["\u0647\u0648\u064A\u0629", "\u0631\u0642\u0645 \u0627\u0644\u0647\u0648\u064A\u0629", "saudi", "\u0648\u0637\u0646\u064A\u0629", "identity"],
-        0.91,
-        "IDENTITY_INTL" /* IDENTITY_INTL */,
-        "saudi_nid"
-      ],
-      [
-        "UAE_EMIRATES_ID",
-        "\\b784-\\d{4}-\\d{7}-\\d\\b",
-        "g",
-        ["emirates", "\u0647\u0648\u064A\u0629", "uae", "emirati", "identity"],
-        0.93,
+        "ES_DNI",
+        "(?:\\d{8}[A-Z]|[XYZ]\\d{7}[A-Z])",
+        ["dni", "nie", "identidad", "nif", "spain"],
+        0.94,
         "IDENTITY_INTL" /* IDENTITY_INTL */,
-        "luhn"
+        "es_id",
+        true,
+        ["*", "es"]
       ],
       // ── CORPORATE ──────────────────────────────────────────────────────
       [
         "CORP_EMPLOYEE_ID",
-        "\\b(?:EMP|EMPLOYEE|ID)[:\\s]?[A-Z0-9]{5,10}\\b",
-        "gi",
+        "(?:EMP|EMPLOYEE|ID)[:\\s]?[A-Z0-9]{5,10}",
         ["employee", "staff", "personnel", "worker"],
         0.55,
         "CORPORATE" /* CORPORATE */,
@@ -43848,7 +43781,11 @@ var init_registry = __esm({
     DLPPatternRegistry = class {
       constructor(loadGroups) {
         this.catalogue = /* @__PURE__ */ new Map();
+        this.localeCategoryRegexMap = /* @__PURE__ */ new Map();
         this.buildCatalogue(loadGroups ?? null);
+        for (const loc of ["*", "en", "es"]) {
+          this.compileForLocale(loc);
+        }
       }
       get typeNames() {
         return [...this.catalogue.keys()];
@@ -43860,23 +43797,74 @@ var init_registry = __esm({
       descriptorFor(typeName) {
         return this.catalogue.get(typeName);
       }
-      /** Return locale-tuned name regexes, falling back to English. */
       namePatternsFor(lang) {
         return LOCALE_NAME_RULES[lang] ?? LOCALE_NAME_RULES["en"];
       }
-      /** Return locale-tuned address regexes, falling back to English. */
       addressPatternsFor(lang) {
         return LOCALE_ADDRESS_RULES[lang] ?? LOCALE_ADDRESS_RULES["en"];
       }
+      getCategoryRegexesMap(locale = "en") {
+        if (!this.localeCategoryRegexMap.has(locale)) {
+          this.compileForLocale(locale);
+        }
+        return this.localeCategoryRegexMap.get(locale);
+      }
+      getCategoryTypeMap(categoryName, locale = "en") {
+        return this.localeCategoryRegexMap.get(locale)?.get(categoryName)?.typeOrder ?? [];
+      }
+      compileForLocale(locale) {
+        const localePool = /* @__PURE__ */ new Map();
+        for (const [typeName, desc] of this.catalogue.entries()) {
+          if (desc.supportedLocales.includes("*") || desc.supportedLocales.includes(locale)) {
+            const catKey = desc.category;
+            if (!localePool.has(catKey)) localePool.set(catKey, []);
+            localePool.get(catKey).push([typeName, desc]);
+          }
+        }
+        const categoryMap = /* @__PURE__ */ new Map();
+        for (const [catKey, entries] of localePool.entries()) {
+          entries.sort(([, a6], [, b6]) => {
+            const aVal = a6.validatorTag ? 0 : 1;
+            const bVal = b6.validatorTag ? 0 : 1;
+            if (aVal !== bVal) return aVal - bVal;
+            return b6.compiledRe.source.length - a6.compiledRe.source.length;
+          });
+          const parts = [];
+          const typeOrder = [];
+          for (const [typeName, desc] of entries) {
+            parts.push(`(?<${typeName}>${desc.compiledRe.source})`);
+            typeOrder.push(typeName);
+          }
+          const combinedSource = parts.join("|");
+          const needsI = entries.some(([, d6]) => d6.compiledRe.flags.includes("i"));
+          const flags = needsI ? "gi" : "g";
+          try {
+            const re = new RegExp(combinedSource, flags);
+            categoryMap.set(catKey, { re, typeOrder });
+          } catch (err2) {
+            console.error(`[DLPPatternRegistry] Locale [${locale}] category [${catKey}] failed:`, err2);
+          }
+        }
+        this.localeCategoryRegexMap.set(locale, categoryMap);
+      }
       buildCatalogue(restrict) {
-        for (const [typeName, regexStr, flags, terms, risk, cat, vtag] of RAW_PATTERNS) {
+        for (const entry of RAW_PATTERNS) {
+          const [typeName, regexSource, terms, risk, cat, vtag, isHighEntropy, supportedLocales] = entry;
           if (restrict !== null && !restrict.has(cat)) continue;
+          let re;
+          if (regexSource instanceof RegExp) {
+            re = regexSource;
+          } else {
+            re = new RegExp(regexSource, "g");
+          }
           this.catalogue.set(typeName, {
-            compiledRe: new RegExp(regexStr, flags),
+            compiledRe: re,
             proximityTerms: new Set(terms),
             baseRisk: risk,
             category: cat,
-            validatorTag: vtag
+            validatorTag: vtag,
+            isHighEntropy: isHighEntropy ?? vtag !== null,
+            supportedLocales: supportedLocales ?? ["*"]
           });
         }
       }
@@ -43973,29 +43961,13 @@ function checkIpv4Octets(raw) {
   }
   return true;
 }
-function checkTcidNumber(raw) {
-  const digitsStr = raw.replace(/\D/g, "");
-  if (digitsStr.length !== 11) return false;
-  const d6 = digitsStr.split("").map(Number);
-  if (d6[0] === 0) return false;
-  if (d6[10] % 2 !== 0) return false;
-  const oddSum = d6[0] + d6[2] + d6[4] + d6[6] + d6[8];
-  const evenSum = d6[1] + d6[3] + d6[5] + d6[7];
-  const computedD10 = ((oddSum * 7 - evenSum) % 10 + 10) % 10;
-  if (computedD10 !== d6[9]) return false;
-  const firstTenSum = d6.slice(0, 10).reduce((a6, b6) => a6 + b6, 0);
-  if (firstTenSum % 10 !== d6[10]) return false;
-  return true;
-}
-function checkSaudiNid(raw) {
-  const digitsStr = raw.replace(/\D/g, "");
-  if (digitsStr.length !== 10) return false;
-  const d6 = digitsStr.split("").map(Number);
-  if (d6[0] !== 1) return false;
+function checkCaSin(raw) {
+  const digits = raw.replace(/\D/g, "");
+  if (digits.length !== 9) return false;
   let total = 0;
-  for (let idx = 0; idx < 10; idx++) {
-    let val = d6[idx];
-    if (idx % 2 === 0) {
+  for (let idx = 0; idx < digits.length; idx++) {
+    let val = parseInt(digits[idx], 10);
+    if (idx % 2 === 1) {
       val *= 2;
       if (val > 9) val -= 9;
     }
@@ -44003,7 +43975,30 @@ function checkSaudiNid(raw) {
   }
   return total % 10 === 0;
 }
-var IBAN_COUNTRY_LENGTHS, VIN_TRANSLITERATION, VIN_WEIGHTS, VALIDATOR_DISPATCH, DLPValidationEngine;
+function checkUkNino(raw) {
+  const cleaned = raw.replace(/ /g, "").toUpperCase();
+  if (cleaned.length !== 9) return false;
+  return UK_NINO_REGEX.test(cleaned);
+}
+function checkEsId(raw) {
+  const cleaned = raw.replace(/[\s-]/g, "").toUpperCase();
+  if (cleaned.length !== 9) return false;
+  const mapping = { X: "0", Y: "1", Z: "2" };
+  const firstChar = cleaned[0];
+  let numStr;
+  if (firstChar in mapping) {
+    numStr = mapping[firstChar] + cleaned.slice(1, 8);
+  } else if (/^\d$/.test(firstChar)) {
+    numStr = cleaned.slice(0, 8);
+  } else {
+    return false;
+  }
+  if (!/^\d+$/.test(numStr)) return false;
+  const num = parseInt(numStr, 10);
+  const validLetters = "TRWAGMYFPDXBNJZSQVHLCKE";
+  return cleaned[8] === validLetters[num % 23];
+}
+var IBAN_COUNTRY_LENGTHS, VIN_TRANSLITERATION, VIN_WEIGHTS, UK_NINO_REGEX, VALIDATOR_DISPATCH, DLPValidationEngine;
 var init_handlers = __esm({
   "src/core/dlp/handlers.ts"() {
     IBAN_COUNTRY_LENGTHS = {
@@ -44106,6 +44101,7 @@ var init_handlers = __esm({
       Z: 9
     };
     VIN_WEIGHTS = [8, 7, 6, 5, 4, 3, 2, 10, 0, 9, 8, 7, 6, 5, 4, 3, 2];
+    UK_NINO_REGEX = /^(?!BG|GB|NK|KN|TN|NT|ZZ)[A-CEGHJ-PR-TW-Z]{2}[0-9]{6}[A-D]$/;
     VALIDATOR_DISPATCH = {
       luhn: checkLuhn,
       ssn_area: checkSsnArea,
@@ -44114,8 +44110,9 @@ var init_handlers = __esm({
       vin_format: checkVinFormat,
       btc_format: checkBtcFormat,
       ipv4: checkIpv4Octets,
-      tcid: checkTcidNumber,
-      saudi_nid: checkSaudiNid
+      ca_sin: checkCaSin,
+      uk_nino: checkUkNino,
+      es_id: checkEsId
     };
     DLPValidationEngine = class {
       /**
@@ -44152,7 +44149,8 @@ var init_scorer = __esm({
       keywordBoost: 0.1,
       validatorOverride: 0.99,
       maxConfidence: 0.99,
-      penaltyFactor: 0.65
+      penaltyFactor: 0.99
+      // Renamed functionally to validator failure penalty subtraction
     };
     DLPConfidenceScorer = class {
       constructor(overrides = {}) {
@@ -44171,7 +44169,7 @@ var init_scorer = __esm({
       score(input) {
         if (input.validatorPassed === true) return this.valOverride;
         if (input.validatorPassed === false) {
-          return Math.min(this.ceil, input.baseRisk * this.penalty);
+          return Math.max(0, input.baseRisk - this.penalty);
         }
         const windowLo = Math.max(0, input.matchStart - this.window);
         const windowHi = Math.min(input.fullText.length, input.matchEnd + this.window);
@@ -44197,6 +44195,47 @@ var init_scorer = __esm({
   }
 });
 
+// src/core/span.ts
+function resolveOverlaps(spans) {
+  if (spans.length === 0) return [];
+  const sorted = [...spans].sort((a6, b6) => {
+    if (a6.start !== b6.start) return a6.start - b6.start;
+    const lenDiff = b6.end - b6.start - (a6.end - a6.start);
+    if (lenDiff !== 0) return lenDiff;
+    return b6.confidence - a6.confidence;
+  });
+  const resolved = [];
+  let occupiedEnd = -1;
+  for (const span of sorted) {
+    if (span.start >= occupiedEnd) {
+      resolved.push(span);
+      occupiedEnd = span.end;
+    } else if (span.end <= occupiedEnd) {
+      continue;
+    } else {
+      const last = resolved[resolved.length - 1];
+      if (span.confidence > last.confidence) {
+        resolved.pop();
+        resolved.push(span);
+        occupiedEnd = span.end;
+      }
+    }
+  }
+  return resolved.sort((a6, b6) => b6.start - a6.start);
+}
+function reconstruct(text, resolvedSpans) {
+  let result = text;
+  for (const span of resolvedSpans) {
+    if (span.maskedValue == null) continue;
+    result = result.slice(0, span.start) + span.maskedValue + result.slice(span.end);
+  }
+  return result;
+}
+var init_span = __esm({
+  "src/core/span.ts"() {
+  }
+});
+
 // node_modules/delayed-stream/lib/delayed_stream.js
 var require_delayed_stream = __commonJS({
   "node_modules/delayed-stream/lib/delayed_stream.js"(exports2, module) {
@@ -58867,7 +58906,7 @@ var init_transformers_scanner = __esm({
               confidence = Math.min(1, confidence + 0.2);
             }
             if (confidence >= confidenceThreshold && !looksLikeToken(val) && val.length > 1) {
-              const token = await encodeFn(val);
+              const token = await encodeFn(val, { entityType });
               entities.push({
                 type: entityType,
                 value: val,
@@ -58974,17 +59013,18 @@ var init_scanner = __esm({
     init_registry();
     init_handlers();
     init_scorer();
+    init_span();
     _dlpLanguageResolver = new LanguageContextResolver();
     _dlpPatternRegistry = new DLPPatternRegistry();
     _dlpValidationEngine = new DLPValidationEngine();
     _dlpConfidenceScorer = new DLPConfidenceScorer();
     REGEX_PATTERNS = {
       "EMAIL_ADDRESS": /[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+/g,
-      "PHONE_NUMBER": /\+?1?[\s\-.]?\(?\d{3}\)?[\s\-.]?\d{3}[\s\-.]?\d{4}|\d{3}[\s\-.]?\d{4}/g,
-      "PHONE_NUMBER_INTL": /\+(?:44|33|49)[\s\-.]?\(?\d{1,5}\)?(?:[\s\-.]?\d{2,4}){2,4}/g,
-      "US_SSN": /\d{3}-\d{2}-\d{4}/g,
-      "CREDIT_CARD": /(?:\d{4}[ \-]?){3}\d{4}/g,
-      "US_ROUTING_NUMBER": /\b\d{9}\b/g,
+      "PHONE_NUMBER": /(?<!\d)(?:\+?1?[\s\-.]?\(?\d{3}\)?[\s\-.]?\d{3}[\s\-.]?\d{4}|\d{3}[\s\-.]?\d{4})(?!\d)/g,
+      "PHONE_NUMBER_INTL": /(?<!\d)\+(?:[1-9]\d{0,3})[-.\s]?\(?\d{1,5}\)?(?:[-.\s]?\d{2,4}){2,4}(?!\d)/g,
+      "US_SSN": /(?<!\d)\d{3}-\d{2}-\d{4}(?!\d)/g,
+      "CREDIT_CARD": /(?<!\d)(?:\d{4}[ \-]?){3}\d{4}(?!\d)/g,
+      "US_ROUTING_NUMBER": /(?<!\d)\d{9}(?!\d)/g,
       "US_PASSPORT": /\b[A-Z]\d{8}\b/g,
       "DATE_OF_BIRTH": /\b(?:0[1-9]|1[0-2])\/(?:0[1-9]|[12]\d|3[01])\/(?:19|20)\d{2}\b|\b(?:19|20)\d{2}-(?:0[1-9]|1[0-2])-(?:0[1-9]|[12]\d|3[01])\b/g
     };
@@ -59039,26 +59079,55 @@ var init_scanner = __esm({
         const checksum = 3 * (d6[0] + d6[3] + d6[6]) + 7 * (d6[1] + d6[4] + d6[7]) + (d6[2] + d6[5] + d6[8]);
         return checksum % 10 === 0;
       }
-      async _tier0Dlp(text, encodeFn, confidenceThreshold) {
+      async _tier0CollectSpans(text, confidenceThreshold) {
         const detectedLanguage = _dlpLanguageResolver.resolve(text);
-        const rawHits = [];
-        for (const [typeTag, descriptor] of _dlpPatternRegistry.iterDescriptors()) {
-          const re = new RegExp(descriptor.compiledRe.source, descriptor.compiledRe.flags);
+        const spans = [];
+        const categoryMap = _dlpPatternRegistry.getCategoryRegexesMap();
+        for (const [catKey, { re, typeOrder }] of categoryMap.entries()) {
+          const megaRe = new RegExp(re.source, re.flags);
           let m6;
-          while ((m6 = re.exec(text)) !== null) {
+          while ((m6 = megaRe.exec(text)) !== null) {
+            const groups = m6.groups ?? {};
+            let typeTag;
+            for (const name of typeOrder) {
+              if (groups[name] !== void 0) {
+                typeTag = name;
+                break;
+              }
+            }
+            if (!typeTag) continue;
             const matchedStr = m6[0];
             if (looksLikeToken(matchedStr)) continue;
+            const descriptor = _dlpPatternRegistry.descriptorFor(typeTag);
+            if (!descriptor) continue;
             const validatorResult = _dlpValidationEngine.run(descriptor.validatorTag, matchedStr);
-            const conf = _dlpConfidenceScorer.score({
-              baseRisk: descriptor.baseRisk,
-              matchStart: m6.index,
-              matchEnd: m6.index + matchedStr.length,
-              fullText: text,
-              proximityTerms: descriptor.proximityTerms,
-              validatorPassed: validatorResult
-            });
+            let conf;
+            if (validatorResult === false) {
+              if (descriptor.isHighEntropy) {
+                conf = 0.85;
+              } else {
+                continue;
+              }
+            } else {
+              conf = _dlpConfidenceScorer.score({
+                baseRisk: descriptor.baseRisk,
+                matchStart: m6.index,
+                matchEnd: m6.index + matchedStr.length,
+                fullText: text,
+                proximityTerms: descriptor.proximityTerms,
+                validatorPassed: validatorResult
+              });
+            }
             if (conf >= confidenceThreshold) {
-              rawHits.push({ start: m6.index, end: m6.index + matchedStr.length, tag: typeTag, val: matchedStr, conf });
+              spans.push({
+                start: m6.index,
+                end: m6.index + matchedStr.length,
+                entityType: typeTag,
+                originalValue: matchedStr,
+                confidence: conf,
+                method: "dlp_heuristic",
+                language: detectedLanguage
+              });
             }
           }
         }
@@ -59077,7 +59146,15 @@ var init_scanner = __esm({
               validatorPassed: null
             });
             if (conf >= confidenceThreshold) {
-              rawHits.push({ start: m6.index, end: m6.index + m6[0].length, tag: "PERSON_NAME", val: m6[0], conf });
+              spans.push({
+                start: m6.index,
+                end: m6.index + m6[0].length,
+                entityType: "PERSON_NAME",
+                originalValue: m6[0],
+                confidence: conf,
+                method: "dlp_heuristic",
+                language: detectedLanguage
+              });
             }
           }
         }
@@ -59086,85 +59163,78 @@ var init_scanner = __esm({
           let m6;
           while ((m6 = re.exec(text)) !== null) {
             if (looksLikeToken(m6[0])) continue;
-            rawHits.push({ start: m6.index, end: m6.index + m6[0].length, tag: "PHYS_ADDRESS", val: m6[0], conf: 0.55 });
-          }
-        }
-        rawHits.sort((a6, b6) => a6.start - b6.start || b6.end - b6.start - (a6.end - a6.start) || b6.conf - a6.conf);
-        const deduped = [];
-        let occupiedEnd = -1;
-        for (const hit of rawHits) {
-          if (hit.start >= occupiedEnd) {
-            deduped.push(hit);
-            occupiedEnd = hit.end;
+            spans.push({
+              start: m6.index,
+              end: m6.index + m6[0].length,
+              entityType: "PHYS_ADDRESS",
+              originalValue: m6[0],
+              confidence: 0.55,
+              method: "dlp_heuristic",
+              language: detectedLanguage
+            });
           }
         }
+        return spans;
+      }
+      /** Backward-compat wrapper — collects spans then single-pass encodes. */
+      async _tier0Dlp(text, encodeFn, confidenceThreshold) {
+        const spans = await this._tier0CollectSpans(text, confidenceThreshold);
+        const resolved = resolveOverlaps(spans);
         const entities = [];
-        let excised = text;
-        for (const hit of [...deduped].reverse()) {
-          const token = await encodeFn(hit.val);
-          excised = excised.slice(0, hit.start) + token + excised.slice(hit.end);
+        await Promise.all(resolved.map(async (span) => {
+          span.maskedValue = await encodeFn(span.originalValue, { entityType: span.entityType });
           entities.push({
-            type: hit.tag,
-            value: hit.val,
-            method: "dlp_heuristic",
-            confidence: hit.conf,
-            masked_value: token,
-            language: detectedLanguage
+            type: span.entityType,
+            value: span.originalValue,
+            method: span.method,
+            confidence: span.confidence,
+            masked_value: span.maskedValue,
+            language: span.language
           });
-        }
-        return [excised, entities];
+        }));
+        return [reconstruct(text, resolved), entities];
       }
-      async _tier1Regex(text, encodeFn, boostEntities, aggressive, confidenceThreshold) {
-        let entities = [];
-        let excised = text;
-        let allMatches = [];
+      async _tier1CollectSpans(text, boostEntities, aggressive, confidenceThreshold) {
+        const spans = [];
         for (const [entityType, pattern] of Object.entries(REGEX_PATTERNS)) {
           const re = new RegExp(pattern.source, pattern.flags);
           let match;
           while ((match = re.exec(text)) !== null) {
-            let confidence = 0.95;
-            if (aggressive || boostEntities.has(entityType.toLowerCase().replace(/_/g, " "))) {
-              confidence = 1;
-            }
-            if (entityType === "CREDIT_CARD" && _BaseScanner._luhnChecksum(match[0])) {
-              confidence = Math.max(confidence, 0.99);
-            }
-            if (entityType === "US_ROUTING_NUMBER" && !_BaseScanner._abaChecksum(match[0])) {
-              continue;
+            const val = match[0];
+            if (looksLikeToken(val)) continue;
+            let confidence = aggressive || boostEntities.has(entityType.toLowerCase().replace(/_/g, " ")) ? 1 : 0.95;
+            if (entityType === "CREDIT_CARD" && _BaseScanner._luhnChecksum(val)) confidence = Math.max(confidence, 0.99);
+            if (entityType === "US_ROUTING_NUMBER" && !_BaseScanner._abaChecksum(val)) continue;
+            if (confidence >= confidenceThreshold) {
+              spans.push({
+                start: match.index,
+                end: match.index + val.length,
+                entityType,
+                originalValue: val,
+                confidence,
+                method: "regex"
+              });
             }
-            allMatches.push({
-              start: match.index,
-              end: match.index + match[0].length,
-              type: entityType,
-              value: match[0],
-              confidence
-            });
-          }
-        }
-        allMatches.sort((a6, b6) => a6.start - b6.start || b6.end - b6.start - (a6.end - a6.start));
-        let filtered = [];
-        let lastEnd = -1;
-        for (const m6 of allMatches) {
-          if (m6.start >= lastEnd) {
-            filtered.push(m6);
-            lastEnd = m6.end;
           }
         }
-        const sortedFiltered = [...filtered].sort((a6, b6) => b6.start - a6.start);
-        for (const m6 of sortedFiltered) {
-          if (m6.confidence >= confidenceThreshold && !looksLikeToken(m6.value)) {
-            const token = await encodeFn(m6.value);
-            excised = excised.slice(0, m6.start) + token + excised.slice(m6.end);
-            entities.push({
-              type: m6.type,
-              value: m6.value,
-              method: "regex",
-              confidence: m6.confidence,
-              masked_value: token
-            });
-          }
-        }
-        return [excised, entities];
+        return spans;
+      }
+      /** Backward-compat wrapper. */
+      async _tier1Regex(text, encodeFn, boostEntities, aggressive, confidenceThreshold) {
+        const spans = await this._tier1CollectSpans(text, boostEntities, aggressive, confidenceThreshold);
+        const resolved = resolveOverlaps(spans);
+        const entities = [];
+        await Promise.all(resolved.map(async (span) => {
+          span.maskedValue = await encodeFn(span.originalValue, { entityType: span.entityType });
+          entities.push({
+            type: span.entityType,
+            value: span.originalValue,
+            method: span.method,
+            confidence: span.confidence,
+            masked_value: span.maskedValue
+          });
+        }));
+        return [reconstruct(text, resolved), entities];
       }
       async _tier2Nlp(text, encodeFn, boostEntities, aggressive, confidenceThreshold) {
         return [text, []];
@@ -59184,13 +59254,18 @@ var init_scanner = __esm({
         const _encode = options.encodeFn || encode;
         const confidenceThreshold = options.confidenceThreshold ?? 0.7;
         const boost = this._resolveBoost(options.context);
-        let currentText = text;
+        const allSpans = [];
         if (pipeline.includes("dlp")) {
-          [currentText] = await this._tier0Dlp(currentText, _encode, confidenceThreshold);
+          allSpans.push(...await this._tier0CollectSpans(text, confidenceThreshold));
         }
         if (pipeline.includes("regex") || pipeline.includes("checksum")) {
-          [currentText] = await this._tier1Regex(currentText, _encode, boost, !!options.aggressive, confidenceThreshold);
+          allSpans.push(...await this._tier1CollectSpans(text, boost, !!options.aggressive, confidenceThreshold));
         }
+        const resolved = resolveOverlaps(allSpans);
+        await Promise.all(resolved.map(async (span) => {
+          span.maskedValue = await _encode(span.originalValue, { entityType: span.entityType });
+        }));
+        let currentText = reconstruct(text, resolved);
         if (pipeline.includes("nlp")) {
           [currentText] = await this._tier2Nlp(currentText, _encode, boost, !!options.aggressive, confidenceThreshold);
         }
@@ -59202,20 +59277,29 @@ var init_scanner = __esm({
         const _encode = options.encodeFn || encode;
         const confidenceThreshold = options.confidenceThreshold ?? 0.7;
         const boost = this._resolveBoost(options.context);
-        let allEntities = [];
-        let remaining = text;
+        const allEntities = [];
+        const allSpans = [];
         if (pipeline.includes("dlp")) {
-          const [newText, tier0] = await this._tier0Dlp(remaining, _encode, confidenceThreshold);
-          remaining = newText;
-          allEntities.push(...tier0);
+          allSpans.push(...await this._tier0CollectSpans(text, confidenceThreshold));
         }
         if (pipeline.includes("regex") || pipeline.includes("checksum")) {
-          const [newText, tier1] = await this._tier1Regex(remaining, _encode, boost, !!options.aggressive, confidenceThreshold);
-          remaining = newText;
-          allEntities.push(...tier1);
-        }
+          allSpans.push(...await this._tier1CollectSpans(text, boost, !!options.aggressive, confidenceThreshold));
+        }
+        const resolved = resolveOverlaps(allSpans);
+        await Promise.all(resolved.map(async (span) => {
+          span.maskedValue = await _encode(span.originalValue, { entityType: span.entityType });
+          allEntities.push({
+            type: span.entityType,
+            value: span.originalValue,
+            method: span.method,
+            confidence: span.confidence,
+            masked_value: span.maskedValue,
+            language: span.language
+          });
+        }));
+        const remaining = reconstruct(text, resolved);
         if (pipeline.includes("nlp")) {
-          const [_newText, tier2] = await this._tier2Nlp(remaining, _encode, boost, !!options.aggressive, confidenceThreshold);
+          const [, tier2] = await this._tier2Nlp(remaining, _encode, boost, !!options.aggressive, confidenceThreshold);
           allEntities.push(...tier2);
         }
         return allEntities;