euredact 0.1.0 → 0.3.0

package/dist/cjs/rules/bicRegistry.js

@@ -0,0 +1,239 @@
+"use strict";
+/**
+ * BIC registry — tier-1 membership testing for ISO 9362 bank identifier codes.
+ *
+ * BIC is the only bank identifier in the supported set with **no check digit**
+ * (IBAN has mod-97, VAT has country-specific checksums). Structure alone cannot
+ * separate a genuine BIC from an ordinary 8- or 11-character uppercase word,
+ * because characters 5-6 of everyday words are frequently valid ISO 3166
+ * country codes (`DRINGEND` -> `GE`, `HOSPITAL` -> `IT`). Registry membership
+ * is therefore the strongest precision signal available.
+ *
+ * Two sources are consulted, in order:
+ *
+ * 1. A deployment-supplied registry, installed via {@link setBicRegistry}.
+ *    This is the path for anyone holding a licensed directory (SWIFTRef).
+ * 2. {@link SEED_BIC6_PREFIXES} — a small bundled set of institution+country
+ *    prefixes for major European banks, compiled from publicly published bank
+ *    data. No licensed directory is bundled.
+ *
+ * Matching is on the **BIC6 prefix** (4-letter institution + 2-letter country),
+ * so any location and branch code on a known institution is accepted. The
+ * prefix is the stable part of a BIC; branch codes churn as branches open and
+ * close.
+ *
+ * The registry is used as an *accept* signal, never as a filter: a candidate
+ * that misses it falls through to the context gate in `suppressors.ts` and is
+ * still detected when banking context is present. A stale list therefore costs
+ * a little recall on bare, contextless BICs — it never causes a leak. Annual
+ * review is sufficient.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.SEED_BIC6_PREFIXES = void 0;
+exports.setBicRegistry = setBicRegistry;
+exports.getBicRegistry = getBicRegistry;
+exports.isRegisteredBic = isRegisteredBic;
+// BIC6 prefixes (institution + country) of major European institutions,
+// compiled from publicly published bank data — not from a licensed directory.
+// Kept deliberately small: this covers the banks that appear bare and
+// contextless in real documents. Anything outside this list is still detected
+// via the context gate.
+exports.SEED_BIC6_PREFIXES = new Set([
+    // --- Netherlands ---
+    "ABNANL", // ABN AMRO
+    "INGBNL", // ING
+    "RABONL", // Rabobank
+    "SNSBNL", // SNS Bank
+    "TRIONL", // Triodos
+    "BUNQNL", // bunq
+    "KNABNL", // Knab
+    "ASNBNL", // ASN Bank
+    "RBRBNL", // RegioBank
+    "FVLBNL", // Van Lanschot
+    "BITSNL", // BinckBank
+    "DEUTNL", // Deutsche Bank NL
+    // --- Belgium ---
+    "BBRUBE", // BNP Paribas Fortis
+    "GEBABE", // BNP Paribas Fortis (legacy)
+    "KREDBE", // KBC
+    "GKCCBE", // Belfius
+    "BPOTBE", // bpost bank
+    "ARSPBE", // Argenta
+    "AXABBE", // AXA Bank Belgium
+    "CREGBE", // Belfius (legacy)
+    "NICABE", // Crelan
+    "TRIOBE", // Triodos Belgium
+    "CTBKBE", // Banca Monte Paschi Belgio
+    "JVBABE", // Van Breda
+    // --- Germany ---
+    "DEUTDE", // Deutsche Bank
+    "COBADE", // Commerzbank
+    "PBNKDE", // Postbank
+    "GENODE", // Volksbanken / Raiffeisen
+    "MALADE", // Sparkassen
+    "HELADE", // Helaba / Sparkassen
+    "BYLADE", // BayernLB / Sparkassen
+    "SOLADE", // Sparkassen (Baden-Wurttemberg)
+    "WELADE", // Sparkassen (Westfalen)
+    "NOLADE", // Norddeutsche Landesbank
+    "SPKKDE", // Sparkasse
+    "HYVEDE", // UniCredit / HypoVereinsbank
+    "DRESDE", // Dresdner (legacy Commerzbank)
+    "INGDDE", // ING-DiBa
+    "NTSBDE", // N26
+    "SFRTDE", // Sutor / fintech BaaS
+    "TEDADE", // Telda / BaaS
+    // --- France ---
+    "BNPAFR", // BNP Paribas
+    "SOGEFR", // Societe Generale
+    "AGRIFR", // Credit Agricole
+    "CCBPFR", // Banque Populaire
+    "CEPAFR", // Caisse d'Epargne
+    "CMCIFR", // Credit Mutuel / CIC
+    "PSSTFR", // La Banque Postale
+    "BREDFR", // BRED Banque Populaire
+    "CRLYFR", // LCL
+    "HSBCFR", // HSBC France
+    "TRZOFR", // Treezor
+    "QNTOFR", // Qonto
+    // --- Austria ---
+    "GIBAAT", // Erste Bank / Sparkassen
+    "BKAUAT", // UniCredit Bank Austria
+    "RZBAAT", // Raiffeisen Bank International
+    "RLNWAT", // Raiffeisen NO-Wien
+    "OBKLAT", // Oberbank
+    "BAWAAT", // BAWAG P.S.K.
+    "VBOEAT", // Volksbank
+    "HYPTAT", // Hypo Tirol
+    "SPIHAT", // Sparkasse
+    // --- Luxembourg ---
+    "BCEELU", // Banque et Caisse d'Epargne de l'Etat
+    "BILLLU", // Banque Internationale a Luxembourg
+    "BGLLLU", // BGL BNP Paribas
+    "CCRALU", // Banque Raiffeisen
+    "BLUXLU", // Banque de Luxembourg
+    // --- Ireland / UK ---
+    "AIBKIE", // Allied Irish Banks
+    "BOFIIE", // Bank of Ireland
+    "ULSBIE", // Ulster Bank
+    "CITIIE", // Citibank Ireland
+    "BARCGB", // Barclays
+    "HBUKGB", // HSBC UK
+    "MIDLGB", // HSBC (legacy Midland)
+    "LOYDGB", // Lloyds
+    "NWBKGB", // NatWest
+    "RBOSGB", // Royal Bank of Scotland
+    "TSBSGB", // TSB
+    "REVOGB", // Revolut
+    "MONZGB", // Monzo
+    "STARGB", // Starling
+    "CHASGB", // JP Morgan Chase UK
+    // --- Nordics ---
+    "NDEADK", // Nordea Denmark
+    "DABADK", // Danske Bank
+    "JYBADK", // Jyske Bank
+    "SYBKDK", // Sydbank
+    "NDEASE", // Nordea Sweden
+    "ESSESE", // SEB
+    "HANDSE", // Handelsbanken
+    "SWEDSE", // Swedbank
+    "DNBANO", // DNB
+    "NDEANO", // Nordea Norway
+    "NDEAFI", // Nordea Finland
+    "OKOYFI", // OP Corporate Bank
+    "HELSFI", // Aktia
+    "NBIIIS", // Landsbankinn (NBIIISRE)
+    "ESJAIS", // Arion Bank (ESJAISRE)
+    "GLITIS", // Islandsbanki (GLITISRE)
+    "MPBAIS", // Kvika banki (MPBAISRE)
+    "KBNONO", // BNbank ASA (KBNONO22)
+    "PRBONO", // Boligbanken ASA (PRBONO22)
+    "AARBDE", // Aareal Bank AG (AARBDE5W)
+    // --- Southern / Central Europe ---
+    "UNCRIT", // UniCredit Italy
+    "BCITIT", // Intesa Sanpaolo
+    "PASCIT", // Monte dei Paschi
+    "BPMOIT", // Banco BPM
+    "CRPPIT", // Cassa di Risparmio
+    "BSCHES", // Banco Santander
+    "BBVAES", // BBVA
+    "CAIXES", // CaixaBank
+    "SABBES", // Banco Sabadell
+    "BKBKES", // Bankinter
+    "BCOMPT", // Millennium BCP
+    "CGDIPT", // Caixa Geral de Depositos
+    "BPIPPT", // Banco BPI
+    "NOVBPT", // Novo Banco
+    "BPKOPL", // PKO Bank Polski
+    "BREXPL", // mBank
+    "INGBPL", // ING Bank Slaski
+    "WBKPPL", // Santander Bank Polska
+    "KOMBCZ", // Komercni banka
+    "CEKOCZ", // CSOB
+    "GIBACZ", // Ceska sporitelna
+    "TATRSK", // Tatra banka
+    "SUBASK", // Vseobecna uverova banka
+    "GIBASK", // Slovenska sporitelna
+    "OTPVHU", // OTP Bank
+    "BUDAHU", // Budapest Bank
+    "MKKBHU", // MKB Bank
+    // --- Switzerland ---
+    "UBSWCH", // UBS
+    "CRESCH", // Credit Suisse
+    "POFICH", // PostFinance
+    "ZKBKCH", // Zurcher Kantonalbank
+    "RAIFCH", // Raiffeisen Switzerland
+    "MIGRCH", // Migros Bank
+]);
+let provider = null;
+/**
+ * Install a deployment-supplied BIC registry, consulted ahead of the seed list.
+ *
+ * Accepts a membership callable, any iterable of BIC strings, or `null` to
+ * remove a previously installed registry. Entries may be full BIC8/BIC11 codes
+ * or bare BIC6 prefixes; both are matched, case-insensitively and ignoring
+ * spaces.
+ */
+function setBicRegistry(registry) {
+    if (registry === null) {
+        provider = null;
+        return;
+    }
+    if (typeof registry === "function") {
+        provider = registry;
+        return;
+    }
+    const known = new Set();
+    for (const entry of registry) {
+        const e = String(entry).trim();
+        if (e && !e.startsWith("#"))
+            known.add(e.replace(/ /g, "").toUpperCase());
+    }
+    provider = (bic) => known.has(bic);
+}
+/** Return the installed registry provider, or null if none is installed. */
+function getBicRegistry() {
+    return provider;
+}
+/**
+ * Tier 1: is `candidate` a BIC of a known institution?
+ *
+ * Checks the deployment-supplied registry first (full code, then BIC6 prefix),
+ * then the bundled seed prefixes.
+ */
+function isRegisteredBic(candidate) {
+    const c = candidate.replace(/ /g, "").toUpperCase();
+    if (c.length !== 8 && c.length !== 11)
+        return false;
+    const prefix = c.slice(0, 6);
+    if (provider !== null) {
+        try {
+            if (provider(c) || provider(prefix))
+                return true;
+        }
+        catch {
+            // a broken provider must not break detection
+        }
+    }
+    return exports.SEED_BIC6_PREFIXES.has(prefix);
+}

package/dist/cjs/rules/countries/index.js

@@ -2,6 +2,26 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.COUNTRY_CONFIGS = void 0;
 const types_js_1 = require("../../types.js");
+const validators_js_1 = require("../validators.js");
+/**
+ * Build a country-independent IBAN pattern from the ISO 13616 length table.
+ *
+ * An IBAN carries its own country code and a mod-97 checksum, so it is
+ * self-identifying and must never depend on the caller's `countries` setting —
+ * a Belgian account in an Austrian file is still an account.
+ *
+ * Each alternative pins the *exact* character count for its country, which
+ * stops the match absorbing a following word: without it, greedy matching
+ * turns `BE68 5390 0754 7034 KBC` into one over-long candidate that fails
+ * mod-97 and is lost entirely.
+ */
+function buildIbanPattern() {
+    const alternatives = Object.entries(validators_js_1.IBAN_LENGTHS)
+        .sort(([a], [b]) => a.localeCompare(b))
+        .map(([code, length]) => `${code}\\d{2}(?:[ \\-]?[A-Z0-9]){${length - 4}}`);
+    return String.raw `\b(?:` + alternatives.join("|") + String.raw `)\b`;
+}
+const GENERIC_IBAN = buildIbanPattern();
 /**
  * Helper to create a PatternDef with sensible defaults.
  */
@@ -9,6 +29,45 @@ function p(entityType, pattern, validator = null, description = "", contextKeywo
     return { entityType, pattern, validator, description, contextKeywords, requiresContext };
 }
 // ---------------------------------------------------------------------------
+// Secret context keywords (consolidated, reused across SECRET patterns)
+// ---------------------------------------------------------------------------
+const SECRET_CONTEXT = [
+    // English
+    "key", "token", "secret", "password", "credential",
+    "api_key", "apikey", "api-key", "auth", "bearer",
+    "aws_secret_access_key", "access_key", "private_key",
+    "AES_KEY", "ENCRYPTION_KEY", "SIGNING_KEY",
+    "ACCOUNT_SID", "AUTH_TOKEN", "TWILIO",
+    "RAILS_SECRET", "SECRET_KEY_BASE", "DATABASE_URL",
+    "NPM_TOKEN", "DIGITALOCEAN", "GOOGLE_API",
+    "DB_PASS", "DB_PASSWORD", "DATASOURCE_PASSWORD",
+    "JWT_SIGNING", "passwd", "DISCORD_TOKEN",
+    // Dutch
+    "wachtwoord", "sleutel", "geheim",
+    // French
+    "mot de passe", "mot_de_passe", "clé", "jeton",
+    // German
+    "Passwort", "Kennwort", "Schlüssel", "Geheimnis",
+    // Italian
+    "chiave", "segreto",
+    // Spanish
+    "contraseña", "clave",
+    // Portuguese
+    "senha", "chave",
+    // Polish
+    "hasło", "klucz",
+    // Czech/Slovak
+    "heslo", "kľúč",
+    // Hungarian
+    "jelszó", "kulcs",
+    // Romanian
+    "parola",
+    // Nordic
+    "lösenord", "nøkkel", "nyckel", "avain", "lykill",
+    // Greek
+    "κωδικός",
+];
+// ---------------------------------------------------------------------------
 // DOB context keywords (reused across SHARED patterns)
 // ---------------------------------------------------------------------------
 const DOB_CONTEXT = [
@@ -36,8 +95,28 @@ const SHARED = {
     code: "SHARED",
     name: "Shared EU Patterns",
     patterns: [
-        p(types_js_1.EntityType.EMAIL, String.raw `\b[\w._%+\-]+@[\w.\-]+\.[a-zA-Z]{2,}\b`),
+        // JavaScript's \w and \b are ASCII-only, so the Python engine's \w (which
+        // is Unicode-aware) matched local parts like "petras_ž@..." while this one
+        // did not. \p{L}\p{N} with the `u` flag restores parity; the leading guard
+        // is a Unicode-safe word boundary.
+        p(types_js_1.EntityType.EMAIL, String.raw `(?<![\p{L}\p{N}._%+\-])[\p{L}\p{N}._%+\-]+@(?:[\p{L}\p{N}\-]+\.)+\p{L}{2,}\b`),
+        // IBAN (any country) — structure + mod-97 only. Country-specific IBAN
+        // patterns still exist and win on identical spans; this one guarantees an
+        // IBAN is never missed just because its country was not requested.
+        p(types_js_1.EntityType.BANK_ACCOUNT, GENERIC_IBAN, "iban", "IBAN — any country (ISO 13616, mod-97 validated)"),
+        // International phone (E.164, any country). Runs alongside the per-country
+        // patterns, which each hard-code a single grouping and miss the rest:
+        // "+43 664 8213 907" was invisible to the Austrian pattern, which expects
+        // an unbroken subscriber number. A leading "+" is self-identifying, so
+        // this is deliberately not country-gated.
+        p(types_js_1.EntityType.PHONE, String.raw `\+\d{1,3}(?:[\s\-/]?\(?\d{1,4}\)?){2,7}`, "e164", "International phone number (E.164) — any country"),
         p(types_js_1.EntityType.BIC, String.raw `\b[A-Z]{4}[A-Z]{2}[A-Z0-9]{2}(?:[A-Z0-9]{3})?\b`, "bic"),
+        // Space-separated form: "RZBA AT WW", "NICA BE BB". Written as a separate
+        // pattern requiring a space between *every* group, rather than making the
+        // spaces optional above — optional spaces would let "GEBABEBB KBC" match
+        // as one 11-character code and claim the following word. Only a literal
+        // space is allowed, so a match cannot span a line break.
+        p(types_js_1.EntityType.BIC, String.raw `\b[A-Z]{4} [A-Z]{2} [A-Z0-9]{2}(?: [A-Z0-9]{3})?\b`, "bic", "BIC/SWIFT code (space-separated groups)"),
         p(types_js_1.EntityType.CREDIT_CARD, String.raw `\b(?:4[0-9]{3}[\s\-]?[0-9]{4}[\s\-]?[0-9]{4}[\s\-]?[0-9]{4}|5[1-5][0-9]{2}[\s\-]?[0-9]{4}[\s\-]?[0-9]{4}[\s\-]?[0-9]{4}|3[47][0-9]{2}[\s\-]?[0-9]{6}[\s\-]?[0-9]{5})\b`, "luhn"),
         p(types_js_1.EntityType.VIN, String.raw `\b[A-HJ-NPR-Z0-9]{17}\b`, "vin"),
         p(types_js_1.EntityType.IP_ADDRESS, String.raw `\b(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\b`),
@@ -48,7 +127,40 @@ const SHARED = {
         p(types_js_1.EntityType.IMEI, String.raw `\b\d{2}[\-\s]\d{6}[\-\s]\d{6}[\-\s]\d\b`, "imei", "formatted"),
         p(types_js_1.EntityType.GPS_COORDINATES, String.raw `-?(?:[1-8]?\d(?:\.\d{4,})|90(?:\.0{4,}))\s*[,;/]\s*-?(?:1[0-7]\d(?:\.\d{4,})|180(?:\.0{4,})|\d{1,2}(?:\.\d{4,}))`),
         p(types_js_1.EntityType.UUID, String.raw `\b[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[1-5][0-9a-fA-F]{3}-[89abAB][0-9a-fA-F]{3}-[0-9a-fA-F]{12}\b`),
-        p(types_js_1.EntityType.SOCIAL_HANDLE, String.raw `(?<!\w)@[a-zA-Z][a-zA-Z0-9_.]{1,29}\b`),
+        p(types_js_1.EntityType.SOCIAL_HANDLE, String.raw `(?<![\p{L}\p{N}_])@\p{L}[\p{L}\p{N}_.]{1,29}\b`),
+        // --- Secret / API Key (known prefixes — always active) ---
+        p(types_js_1.EntityType.SECRET, String.raw `\bAKIA[A-Z0-9]{16}\b`, null, "AWS Access Key ID"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bgh[poas]_[a-zA-Z0-9]{36,}\b`, null, "GitHub token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bgithub_pat_[a-zA-Z0-9_]{22,}\b`, null, "GitHub fine-grained PAT"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bgl[a-z]t-[a-zA-Z0-9_\-]{20,}\b`, null, "GitLab token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bglpat-[a-zA-Z0-9_\-]{20,}\b`, null, "GitLab PAT"),
+        p(types_js_1.EntityType.SECRET, String.raw `\b[sp]k_(?:live|test)_[a-zA-Z0-9]{24,}\b`, null, "Stripe key"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bsk-(?:ant-)?[a-zA-Z0-9\-_]{20,}\b`, null, "OpenAI/Anthropic key"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bxox[bpas]-\d[\d\-]{10,}`, null, "Slack numeric token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bxox[bpas]-[a-zA-Z0-9\-]{10,}`, null, "Slack token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\beyJ[a-zA-Z0-9_\-]+\.[a-zA-Z0-9_\-]+\.[a-zA-Z0-9_\-]+\b`, null, "JWT token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bSG\.[a-zA-Z0-9_\-]{22}\.[a-zA-Z0-9_\-]{43}\b`, null, "SendGrid key"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bhvs\.[a-zA-Z0-9_\-]{20,}\b`, null, "Vault token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bkey-[a-f0-9]{20,}\b`, null, "Mailgun key"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bnpm_[a-zA-Z0-9]{36,}\b`, null, "npm token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bAIza[a-zA-Z0-9_\-]{35}\b`, null, "Google API key"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bdop_v1_[a-f0-9]{64}\b`, null, "DigitalOcean token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bAC[a-f0-9]{32}\b`, null, "Twilio Account SID"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bxapp-\d[\d\-]{20,}`, null, "Slack app numeric token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bxapp-[a-zA-Z0-9\-]{20,}`, null, "Slack app token"),
+        p(types_js_1.EntityType.SECRET, String.raw `\bdapi[a-f0-9]{32,}\b`, null, "Databricks PAT"),
+        // Discord bot token (3-segment base64 with numeric first segment)
+        p(types_js_1.EntityType.SECRET, String.raw `\b[MN][A-Za-z0-9]{23,}\.[A-Za-z0-9_\-]{6}\.[A-Za-z0-9_\-]{27,}\b`, null, "Discord bot token"),
+        // PEM private key headers
+        p(types_js_1.EntityType.SECRET, String.raw `-----BEGIN\s(?:RSA\s)?PRIVATE\sKEY-----[\s\S]{0,16384}?-----END\s(?:RSA\s)?PRIVATE\sKEY-----`, null, "PEM private key"),
+        // AWS Secret Access Key (40 chars, base64 with +/ after an AKIA line)
+        p(types_js_1.EntityType.SECRET, String.raw `(?<=aws_secret_access_key\s{0,4}[:=]\s{0,4})[A-Za-z0-9+/]{40}`, null, "AWS Secret Access Key"),
+        // --- Secret / API Key (connection strings with embedded credentials) ---
+        p(types_js_1.EntityType.SECRET, String.raw `(?:mongodb|mysql|postgres(?:ql)?|redis|amqp|rabbitmq):\/\/[^\s:]{1,256}:[^\s@]{1,256}@[^\s]+`, null, "Connection string with credentials"),
+        // --- Secret / API Key (assignment-based: KEY=value or KEY: value) ---
+        p(types_js_1.EntityType.SECRET, String.raw `(?<=[:=] )\S{8,}|(?<=[:=])\S{8,}`, "high_entropy", "Assigned secret value", SECRET_CONTEXT, true),
+        // --- Secret / API Key (entropy-based fallback for longer tokens) ---
+        p(types_js_1.EntityType.SECRET, String.raw `\b[A-Za-z0-9_\-+/]{24,}[A-Za-z0-9_\-+/=]*\b`, "high_entropy", "High-entropy token", SECRET_CONTEXT, true),
         p(types_js_1.EntityType.DOB, String.raw `\b(?:0[1-9]|[12][0-9]|3[01])[/.\-](?:0[1-9]|1[0-2])[/.\-](?:19|20)\d{2}\b`, null, "DD/MM/YYYY", DOB_CONTEXT, true),
         p(types_js_1.EntityType.DATE_OF_DEATH, String.raw `\b(?:0[1-9]|[12][0-9]|3[01])[/.\-](?:0[1-9]|1[0-2])[/.\-](?:19|20)\d{2}\b`, null, "", DATE_OF_DEATH_CONTEXT, true),
         p(types_js_1.EntityType.DOB, String.raw `\b(?:19|20)\d{2}[/.\-](?:0[1-9]|1[0-2])[/.\-](?:0[1-9]|[12][0-9]|3[01])\b`, null, "ISO YYYY-MM-DD", DOB_ISO_CONTEXT, true),
@@ -122,6 +234,11 @@ const AT = {
         p(types_js_1.EntityType.IBAN, String.raw `\bAT\d{2}\s?\d{4}\s?\d{4}\s?\d{4}\s?\d{4}\b`, "iban"),
         p(types_js_1.EntityType.VAT, String.raw `\bATU\d{8}\b`),
         p(types_js_1.EntityType.PHONE, String.raw `\b0\d{3,4}[\s\-]?\d{5,8}\b`),
+        // Short area codes (Vienna is "01") split across three groups:
+        // "01 53460 2215". The pattern above requires 3-4 digits after the trunk
+        // 0, so it never matched these. Both separators are mandatory here, which
+        // keeps "01 2025" (a date fragment) out.
+        p(types_js_1.EntityType.PHONE, String.raw `\b0\d{1,4}[\s\-/]\d{3,7}[\s\-/]\d{2,6}\b`, null, "Austrian national phone — short area code, grouped"),
         p(types_js_1.EntityType.PHONE, String.raw `\+43\s?\d{3,4}[\s\-]?\d{5,8}`),
         p(types_js_1.EntityType.LICENSE_PLATE, String.raw `\b[A-ZÄÖÜ]{1,2}\s?\d{1,5}\s?[A-Z]{1,2}\b`, null, "", ["Kennzeichen", "Nummernschild", "Kfz-Kennzeichen"], true),
         p(types_js_1.EntityType.POSTAL_CODE, String.raw `\b[1-9]\d{3}\b`, null, "", ["PLZ", "Postleitzahl", "Adresse", "Anschrift", "Straße", "Str.", "Gasse", "Weg", "Platz", "Postal:", "Wohnort"], true),

package/dist/cjs/rules/engine.js

@@ -1,14 +1,73 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.RuleEngine = void 0;
+exports.RuleEngine = exports.COUNTRY_CODE_ALIASES = void 0;
+exports.resolveCountryCode = resolveCountryCode;
 const types_js_1 = require("../types.js");
 const matchers_js_1 = require("./matchers.js");
 const suppressors_js_1 = require("./suppressors.js");
 const structural_js_1 = require("./structural.js");
 const index_js_1 = require("./countries/index.js");
+/**
+ * Heuristic to detect nested quantifiers that cause catastrophic backtracking.
+ * After stripping escaped chars and character classes, look for an unbounded
+ * quantifier (+/*) closing a group that is itself quantified (+/*).
+ */
+const NESTED_QUANTIFIER_RE = /[+*]\)[+*]/;
+function validateCustomPattern(pattern) {
+    try {
+        new RegExp(pattern, "gu");
+    }
+    catch (e) {
+        throw new Error(`Invalid regex pattern: ${e instanceof Error ? e.message : e}`);
+    }
+    const stripped = pattern
+        .replace(/\\./g, "X")
+        .replace(/\[(?:[^\]\\]|\\.)*\]/g, "X");
+    if (NESTED_QUANTIFIER_RE.test(stripped)) {
+        throw new Error("Pattern contains nested quantifiers (e.g. (a+)+) which can cause " +
+            "catastrophic backtracking (ReDoS). Restructure the pattern to " +
+            "avoid quantifiers inside quantified groups.");
+    }
+}
+/**
+ * The country modules use the EU/VAT convention (UK, EL) while the public
+ * contract is ISO 3166-1 alpha-2 (GB, GR). Accept both spellings so a caller
+ * passing correct ISO codes is not punished for it.
+ */
+exports.COUNTRY_CODE_ALIASES = {
+    GB: "UK",
+    GR: "EL",
+    UK: "UK",
+    EL: "EL",
+};
+/** Map a caller-supplied code to a registered one, or null if unknown. */
+function resolveCountryCode(countryCode) {
+    let code = countryCode.trim().toUpperCase();
+    code = exports.COUNTRY_CODE_ALIASES[code] ?? code;
+    return code in index_js_1.COUNTRY_CONFIGS && code !== "SHARED" ? code : null;
+}
+const warnedCountries = new Set();
+function warnUnknownCountry(code) {
+    if (warnedCountries.has(code))
+        return;
+    warnedCountries.add(code);
+    const available = Object.keys(index_js_1.COUNTRY_CONFIGS).filter(c => c !== "SHARED").sort();
+    console.warn(`[euredact] Unknown country code: ${JSON.stringify(code)}. Continuing with ` +
+        `shared country-independent patterns only (email, IBAN, international ` +
+        `phone, credit card, ...). Available: ${available.join(", ")}`);
+}
 class RuleEngine {
     matcher = new matchers_js_1.MultiPatternMatcher();
     loadedCountries = new Set();
+    addCustomPattern(name, pattern) {
+        validateCustomPattern(pattern);
+        const pdef = {
+            entityType: name, pattern, validator: null,
+            description: "", contextKeywords: [], requiresContext: false,
+        };
+        this.matcher.addPattern(pdef, "CUSTOM");
+        this.matcher.compile();
+    }
     loadCountries(countryCodes) {
         let configs;
         if (countryCodes === null) {
@@ -17,11 +76,18 @@ class RuleEngine {
         else {
             configs = [];
             for (const c of countryCodes) {
-                const code = c.toUpperCase();
+                const code = resolveCountryCode(c);
+                if (code === null) {
+                    // Detection continues with the shared, country-independent
+                    // patterns. Throwing on an unknown locale invites callers to wrap
+                    // the call in try/catch and skip redaction — failing open, with
+                    // unredacted PII in the output.
+                    warnUnknownCountry(c);
+                    continue;
+                }
                 const config = index_js_1.COUNTRY_CONFIGS[code];
                 if (config)
                     configs.push(config);
-                // Silently skip unknown country codes (e.g. "INTL")
             }
             const shared = index_js_1.COUNTRY_CONFIGS["SHARED"];
             if (shared)
@@ -40,44 +106,131 @@ class RuleEngine {
     }
     detect(text, countryCodes) {
         this.loadCountries(countryCodes);
-        let rawMatches = this.matcher.scan(text);
+        // Scope the scan to requested countries for efficiency
+        // Resolve through the alias table so `countries: ["GB"]` actually scans
+        // the patterns registered under "UK".
+        let scanCodes = null;
         if (countryCodes !== null) {
-            const codesUpper = new Set(countryCodes.map(c => c.toUpperCase()));
-            rawMatches = rawMatches.filter(m => codesUpper.has(m.countryCode) || m.countryCode === "SHARED");
+            scanCodes = new Set();
+            for (const c of countryCodes) {
+                const code = resolveCountryCode(c);
+                if (code !== null)
+                    scanCodes.add(code);
+            }
+            scanCodes.add("SHARED");
+            scanCodes.add("CUSTOM");
         }
-        const validated = rawMatches.filter(m => this.matcher.validate(m));
-        const detections = [];
-        for (const match of validated) {
+        let rawMatches = this.matcher.scan(text, scanCodes);
+        // Matches that have a validator but fail it create suppression zones:
+        // the span is recognisably a specific entity (e.g. IBAN-shaped) so
+        // overlapping regex-only matches (license plate, phone) are false
+        // positives and must be suppressed.
+        const validated = [];
+        const rawZones = [];
+        for (const m of rawMatches) {
+            const isValid = this.matcher.validate(m);
+            if (isValid) {
+                validated.push({ match: m, hasValidValidator: m.patternDef.validator !== null });
+            }
+            else if (m.patternDef.validator !== null && !m.patternDef.requiresContext) {
+                rawZones.push([m.start, m.end]);
+            }
+        }
+        // Merge overlapping suppression zones for O(log n) containment checks
+        let zoneStarts = [];
+        let zoneEnds = [];
+        if (rawZones.length > 0) {
+            rawZones.sort((a, b) => a[0] - b[0] || a[1] - b[1]);
+            let [ms, me] = rawZones[0];
+            for (let i = 1; i < rawZones.length; i++) {
+                const [zs, ze] = rawZones[i];
+                if (zs <= me) {
+                    me = Math.max(me, ze);
+                }
+                else {
+                    zoneStarts.push(ms);
+                    zoneEnds.push(me);
+                    ms = zs;
+                    me = ze;
+                }
+            }
+            zoneStarts.push(ms);
+            zoneEnds.push(me);
+        }
+        // Build candidates with priority: validated (3) > custom (2) > regex-only (1)
+        const candidates = [];
+        for (const { match, hasValidValidator } of validated) {
             if ((0, suppressors_js_1.shouldSuppress)(text, match))
                 continue;
-            detections.push({
-                entityType: match.patternDef.entityType,
-                start: match.start,
-                end: match.end,
-                text: match.text,
-                source: types_js_1.DetectionSource.RULES,
-                country: match.countryCode !== "SHARED" ? match.countryCode : null,
-                confidence: "high",
+            // O(log n) binary-search containment check on merged zones
+            if (match.patternDef.validator === null && zoneStarts.length > 0) {
+                let lo = 0, hi = zoneStarts.length;
+                while (lo < hi) {
+                    const mid = (lo + hi) >>> 1;
+                    if (zoneStarts[mid] <= match.start)
+                        lo = mid + 1;
+                    else
+                        hi = mid;
+                }
+                const idx = lo - 1;
+                if (idx >= 0 && zoneEnds[idx] >= match.end)
+                    continue;
+            }
+            // A bare digit run is the weakest evidence in the engine and must never
+            // re-cut a span a structured detector claims (PHONE, SSN, NATIONAL_ID,
+            // BANK_ACCOUNT, VAT...). Postal codes take unclaimed spans only,
+            // whatever the span lengths.
+            const priority = hasValidValidator
+                ? 3
+                : match.countryCode === "CUSTOM"
+                    ? 2
+                    : match.patternDef.entityType === types_js_1.EntityType.POSTAL_CODE
+                        ? 0
+                        : 1;
+            const spanLength = match.end - match.start;
+            const score = priority * 1_000_000 + spanLength;
+            candidates.push({
+                det: {
+                    entityType: match.patternDef.entityType,
+                    start: match.start,
+                    end: match.end,
+                    text: match.text,
+                    source: types_js_1.DetectionSource.RULES,
+                    country: (match.countryCode !== "SHARED" && match.countryCode !== "CUSTOM") ? match.countryCode : null,
+                    confidence: "high",
+                },
+                score,
             });
         }
-        detections.push(...(0, structural_js_1.detectStructuralDob)(text));
-        return this.deduplicate(detections);
+        for (const d of (0, structural_js_1.detectStructuralDob)(text)) {
+            candidates.push({ det: d, score: 1_000_000 + (d.end - d.start) });
+        }
+        return this.deduplicate(candidates);
     }
-    deduplicate(detections) {
-        if (detections.length === 0)
-            return detections;
-        const sorted = [...detections].sort((a, b) => a.start - b.start || (b.end - b.start) - (a.end - a.start));
+    /**
+     * Remove overlapping detections with priority-aware resolution.
+     * Priority: validated (3) > custom (2) > regex-only (1).
+     * Within the same tier, longer span wins.
+     */
+    deduplicate(candidates) {
+        if (candidates.length === 0)
+            return [];
+        const sorted = [...candidates].sort((a, b) => b.score - a.score);
         const result = [];
-        for (const det of sorted) {
+        const occupied = new Set();
+        for (const { det } of sorted) {
             let overlaps = false;
-            for (const kept of result) {
-                if (det.start < kept.end && det.end > kept.start) {
+            for (let p = det.start; p < det.end; p++) {
+                if (occupied.has(p)) {
                     overlaps = true;
                     break;
                 }
             }
-            if (!overlaps)
+            if (!overlaps) {
+                for (let p = det.start; p < det.end; p++)
+                    occupied.add(p);
                 result.push(det);
+            }
         }
         return result;
     }