secure-redact 1.0.0

package/dist/assets/advanced.worker-DSVrF0gl.js

@@ -0,0 +1,890 @@
+function S(e) {
+  if (e.length === 0) return [];
+  const n = e.map((i) => i.bbox.h).sort((i, s) => i - s), a = (n[Math.floor(n.length / 2)] || 20) * 0.5, r = [...e].sort((i, s) => {
+    const l = i.bbox.y - s.bbox.y;
+    return Math.abs(l) < a ? i.bbox.x - s.bbox.x : l;
+  }), c = [];
+  let o = [r[0]];
+  for (let i = 1; i < r.length; i++) {
+    const s = r[i - 1], l = r[i];
+    Math.abs(l.bbox.y - s.bbox.y) < a ? o.push(l) : (c.push(v(o)), o = [l]);
+  }
+  return o.length > 0 && c.push(v(o)), c;
+}
+function v(e) {
+  const n = e.map((c) => c.bbox.x), t = e.map((c) => c.bbox.y), a = e.map((c) => c.bbox.h), r = e.map((c) => c.bbox.x + c.bbox.w);
+  return {
+    words: e,
+    y: t.reduce((c, o) => c + o, 0) / t.length,
+    x: Math.min(...n),
+    width: Math.max(...r) - Math.min(...n),
+    height: Math.max(...a),
+    text: e.map((c) => c.text).join(" ")
+  };
+}
+function w(e, n) {
+  if (e.length === 0) return [];
+  const t = [...e].sort((s, l) => s.y - l.y), a = [];
+  for (let s = 1; s < t.length; s++)
+    a.push(t[s].y - (t[s - 1].y + t[s - 1].height));
+  const c = (a.reduce((s, l) => s + l, 0) / a.length || 20) * 2.5, o = [];
+  let i = [t[0]];
+  for (let s = 1; s < t.length; s++) {
+    const l = t[s - 1], d = t[s];
+    d.y - (l.y + l.height) > c ? (o.push(k(i, n)), i = [d]) : i.push(d);
+  }
+  return i.length > 0 && o.push(k(i, n)), o;
+}
+function k(e, n) {
+  const t = e.map((o) => o.x), a = e.map((o) => o.y), r = e.map((o) => o.x + o.width), c = e.map((o) => o.y + o.height);
+  return {
+    lines: e,
+    bbox: {
+      x: Math.min(...t),
+      y: Math.min(...a),
+      w: Math.max(...r) - Math.min(...t),
+      h: Math.max(...c) - Math.min(...a),
+      pageIndex: n
+    },
+    text: e.map((o) => o.text).join(`
+`)
+  };
+}
+const W = [
+  // Names
+  { pattern: /^(name|patient name|customer name|account holder|holder name|full name|cardholder|person name):?$/i, type: "NAME", confidence: 0.99 },
+  { pattern: /^(father'?s? name|father name|spouse name):?$/i, type: "NAME", confidence: 0.95 },
+  // Addresses
+  { pattern: /^(address|residence|location|billing address|shipping address|permanent address|correspondence address):?$/i, type: "ADDRESS", confidence: 0.98 },
+  // Phone
+  { pattern: /^(phone|mobile|tel|telephone|contact|ph\.?|mob\.?|contact no\.?|phone no\.?|mobile no\.?):?$/i, type: "PHONE", confidence: 0.97 },
+  // Email
+  { pattern: /^(email|e-mail|email id|email address):?$/i, type: "EMAIL", confidence: 0.98 },
+  // Aadhaar
+  { pattern: /^(aadhaar|aadhar|uid|aadhaar no\.?|aadhar no\.?|aadhaar number|enrollment no\.?):?$/i, type: "AADHAAR", confidence: 0.99 },
+  // PAN
+  { pattern: /^(pan|pan no\.?|pan number|permanent account number):?$/i, type: "PAN", confidence: 0.99 },
+  // Bank Account
+  { pattern: /^(account no\.?|account number|a\/c no\.?|acc no\.?|acct no\.?):?$/i, type: "ACCOUNT_NUMBER", confidence: 0.99 },
+  { pattern: /^(ifsc|ifsc code|branch code|micr|micr code):?$/i, type: "IFSC", confidence: 0.95 },
+  // DOB
+  { pattern: /^(dob|date of birth|birth date|d\.o\.b\.?):?$/i, type: "DOB", confidence: 0.97 },
+  // Medical
+  { pattern: /^(diagnosis|condition|disease|medication|prescription|blood group|blood type):?$/i, type: "MEDICAL", confidence: 0.9 },
+  // Invoice/Tax
+  { pattern: /^(invoice no\.?|bill no\.?|invoice number|bill number):?$/i, type: "INVOICE_NO", confidence: 0.85 },
+  { pattern: /^(gst no\.?|gstin|gst number|tax id):?$/i, type: "GST", confidence: 0.85 }
+];
+function $(e, n) {
+  const t = [];
+  for (const a of e)
+    for (let r = 0; r < a.words.length; r++) {
+      const c = a.words[r], o = _(c.text);
+      if (o) {
+        const i = I(a.words, r);
+        if (i.length > 0) {
+          const s = [c], l = i.map((h) => h.text).join(" "), d = [...s, ...i], u = P(d, n);
+          t.push({
+            key: s,
+            value: i,
+            keyText: c.text,
+            valueText: l,
+            confidence: o.confidence,
+            type: o.type,
+            bbox: u
+          }), r += i.length;
+        }
+      }
+    }
+  return t;
+}
+function _(e) {
+  const n = e.trim();
+  for (const { pattern: t, type: a, confidence: r } of W)
+    if (t.test(n))
+      return { type: a, confidence: r };
+  return null;
+}
+function I(e, n) {
+  const t = [];
+  for (let a = n + 1; a < e.length; a++) {
+    const r = e[a];
+    if (_(r.text)) break;
+    if (/^[.,:;!?]+$/.test(r.text.trim())) {
+      if (t.length > 0) break;
+      continue;
+    }
+    if (t.length > 0) {
+      const c = t[t.length - 1], o = r.bbox.x - (c.bbox.x + c.bbox.w), i = c.bbox.w;
+      if (o > i * 2) break;
+    }
+    if (t.push(r), t.length >= 15) break;
+  }
+  return t;
+}
+function P(e, n) {
+  if (e.length === 0)
+    return { x: 0, y: 0, w: 0, h: 0, pageIndex: n };
+  const t = e.map((o) => o.bbox.x), a = e.map((o) => o.bbox.y), r = e.map((o) => o.bbox.x + o.bbox.w), c = e.map((o) => o.bbox.y + o.bbox.h);
+  return {
+    x: Math.min(...t),
+    y: Math.min(...a),
+    w: Math.max(...r) - Math.min(...t),
+    h: Math.max(...c) - Math.min(...a),
+    pageIndex: n
+  };
+}
+function O(e) {
+  return e.map((n) => ({
+    id: "kv_" + crypto.randomUUID().substring(0, 8),
+    type: n.type,
+    value: n.valueText,
+    confidence: n.confidence,
+    bbox: n.bbox,
+    masked: !0,
+    layer: 3
+    // Spatial layer (higher priority than NLP)
+  }));
+}
+function L(e) {
+  const n = /* @__PURE__ */ new Map(), t = 15;
+  for (const r of e)
+    for (const c of r.words) {
+      const o = c.bbox.x;
+      let i = !1;
+      for (const [s, l] of n)
+        if (Math.abs(o - s) < t) {
+          l.push(c), i = !0;
+          break;
+        }
+      i || n.set(o, [c]);
+    }
+  return Array.from(n.entries()).sort((r, c) => r[0] - c[0]).map((r, c) => ({
+    columnIndex: c,
+    words: r[1]
+  }));
+}
+const B = [
+  "invoice",
+  "bill",
+  "bill no",
+  "invoice no",
+  "inv no",
+  "invoice number",
+  "bill to",
+  "ship to",
+  "customer",
+  "vendor",
+  "supplier",
+  "po number",
+  "purchase order",
+  "payment terms",
+  "due date",
+  "subtotal",
+  "tax",
+  "gst",
+  "igst",
+  "cgst",
+  "sgst",
+  "total amount",
+  "amount due",
+  "line item",
+  "qty",
+  "quantity",
+  "price",
+  "item description",
+  "hsn",
+  "sac code"
+], j = [
+  "bank statement",
+  "account statement",
+  "statement of account",
+  "account number",
+  "account no",
+  "ifsc",
+  "ifsc code",
+  "branch",
+  "transaction",
+  "transaction date",
+  "credit",
+  "debit",
+  "balance",
+  "opening balance",
+  "closing balance",
+  "withdrawal",
+  "deposit",
+  "cheque",
+  "check",
+  "rtgs",
+  "neft",
+  "imps",
+  "upi",
+  "iban",
+  "swift",
+  "account holder",
+  "from date",
+  "to date",
+  "statement period"
+], K = [
+  "patient",
+  "patient name",
+  "doctor",
+  "physician",
+  "hospital",
+  "clinic",
+  "medical report",
+  "lab report",
+  "pathology",
+  "radiology",
+  "diagnosis",
+  "prescription",
+  "medication",
+  "blood test",
+  "urine test",
+  "mri",
+  "ct scan",
+  "x-ray",
+  "ultrasound",
+  "ecg",
+  "ekg",
+  "test results",
+  "normal range",
+  "abnormal",
+  "positive",
+  "negative",
+  "hemoglobin",
+  "glucose",
+  "cholesterol",
+  "blood pressure",
+  "heart rate",
+  "pulse",
+  "temperature",
+  "weight",
+  "bmi"
+], z = [
+  "income tax",
+  "tax return",
+  "itr",
+  "assessment year",
+  "financial year",
+  "pan",
+  "permanent account number",
+  "tax deducted",
+  "tds",
+  "form 16",
+  "form 26as",
+  "gross income",
+  "taxable income",
+  "deductions",
+  "exemptions",
+  "tax payable",
+  "tax refund",
+  "acknowledgement",
+  "ack no",
+  "return filed",
+  "tan",
+  "employer",
+  "salary",
+  "wages",
+  "capital gains"
+], H = [
+  "aadhaar",
+  "aadhar",
+  "uid",
+  "unique identification",
+  "uidai",
+  "government of india",
+  "bharatiya prachnya patr",
+  "enrollment no",
+  "vid",
+  "virtual id",
+  "date of birth",
+  "dob",
+  "gender",
+  "male",
+  "female",
+  "address",
+  "yob",
+  "year of birth"
+], F = [
+  "pan",
+  "permanent account number",
+  "income tax department",
+  "father name",
+  "fathers name",
+  "date of birth",
+  "dob",
+  "signature",
+  "photograph"
+], Y = [
+  "health report",
+  "medical certificate",
+  "fitness certificate",
+  "blood group",
+  "allergies",
+  "past medical history",
+  "current medications",
+  "vital signs",
+  "examination",
+  "clinical findings"
+];
+function g(e, n) {
+  const t = e.toLowerCase();
+  let a = 0;
+  const r = [];
+  for (const c of n)
+    if (t.includes(c)) {
+      const o = c.split(" ").length;
+      a += o, r.push(c);
+    }
+  return { score: a, matched: r };
+}
+function U(e) {
+  const n = g(e, B), t = g(e, j), a = g(e, K), r = g(e, z), c = g(e, H), o = g(e, F), i = g(e, Y), s = [
+    { type: "invoice", score: n.score, matched: n.matched },
+    { type: "bank_statement", score: t.score, matched: t.matched },
+    { type: "medical_report", score: a.score, matched: a.matched },
+    { type: "tax_return", score: r.score, matched: r.matched },
+    { type: "id_card_aadhaar", score: c.score, matched: c.matched },
+    { type: "id_card_pan", score: o.score, matched: o.matched },
+    { type: "health_report", score: i.score, matched: i.matched }
+  ];
+  s.sort((p, m) => m.score - p.score);
+  const l = s[0], d = s.reduce((p, m) => p + m.score, 0);
+  if (l.score === 0)
+    return {
+      primaryType: "generic",
+      confidence: 0,
+      secondaryTypes: [],
+      detectedKeywords: []
+    };
+  const u = d > 0 ? l.score / d : 0, h = s.slice(1).filter((p) => p.score > 0).map((p) => ({
+    type: p.type,
+    confidence: d > 0 ? p.score / d : 0
+  }));
+  return {
+    primaryType: l.type,
+    confidence: Math.min(u, 0.95),
+    // Cap at 95%
+    secondaryTypes: h,
+    detectedKeywords: l.matched
+  };
+}
+function G(e) {
+  switch (e) {
+    case "invoice":
+      return {
+        prioritizeSpatialContext: !0,
+        enableTableDetection: !0,
+        aggressiveRegex: !0,
+        skipParagraphNER: !0,
+        confidenceBoost: 1.2
+      };
+    case "bank_statement":
+      return {
+        prioritizeSpatialContext: !0,
+        enableTableDetection: !0,
+        aggressiveRegex: !0,
+        skipParagraphNER: !0,
+        confidenceBoost: 1.3
+      };
+    case "medical_report":
+    case "health_report":
+      return {
+        prioritizeSpatialContext: !1,
+        enableTableDetection: !1,
+        aggressiveRegex: !1,
+        skipParagraphNER: !1,
+        confidenceBoost: 1
+      };
+    case "tax_return":
+      return {
+        prioritizeSpatialContext: !0,
+        enableTableDetection: !0,
+        aggressiveRegex: !0,
+        skipParagraphNER: !0,
+        confidenceBoost: 1.15
+      };
+    case "id_card_aadhaar":
+    case "id_card_pan":
+      return {
+        prioritizeSpatialContext: !0,
+        enableTableDetection: !1,
+        aggressiveRegex: !0,
+        skipParagraphNER: !0,
+        confidenceBoost: 1.25
+      };
+    default:
+      return {
+        prioritizeSpatialContext: !1,
+        enableTableDetection: !1,
+        aggressiveRegex: !1,
+        skipParagraphNER: !1,
+        confidenceBoost: 1
+      };
+  }
+}
+function b(e) {
+  let n = e.confidence;
+  switch (e.layer) {
+    case 0:
+      n = 1;
+      break;
+    case 1:
+      n = Math.max(n, 0.95);
+      break;
+    case 2:
+      n = Math.min(Math.max(n, 0.6), 0.9);
+      break;
+    case 3:
+      n = Math.min(Math.max(n, 0.95), 0.99);
+      break;
+    case 4:
+      n = Math.min(Math.max(n, 0.5), 0.95);
+      break;
+    default:
+      n = Math.max(n, 0.5);
+  }
+  return {
+    ...e,
+    confidence: n
+  };
+}
+function V(e, n) {
+  if (e.pageIndex !== n.pageIndex) return 0;
+  const t = Math.max(0, Math.min(e.x + e.w, n.x + n.w) - Math.max(e.x, n.x)), a = Math.max(0, Math.min(e.y + e.h, n.y + n.h) - Math.max(e.y, n.y)), r = t * a, c = e.w * e.h, o = n.w * n.h, i = c + o - r;
+  return i > 0 ? r / i : 0;
+}
+function X(e, n = 0.5) {
+  const t = [...e].sort((c, o) => c.layer !== o.layer ? o.layer - c.layer : Math.abs(c.confidence - o.confidence) > 0.01 ? o.confidence - c.confidence : o.value.length - c.value.length), a = [], r = [];
+  for (const c of t) {
+    let o = !0;
+    for (const i of a)
+      if (V(c.bbox, i.bbox) > n) {
+        o = !1, r.push(c);
+        break;
+      }
+    o && a.push(c);
+  }
+  return {
+    deduplicated: a,
+    removedCount: r.length
+  };
+}
+class Z {
+  config;
+  constructor(n) {
+    this.config = n;
+  }
+  /**
+   * Main fusion method: combines detections from all sources.
+   */
+  fuse(n, t, a, r) {
+    const c = n.map(b), o = t.map(b), i = a.map(b), s = r.map(b);
+    let l = [
+      ...c,
+      ...o,
+      ...i,
+      ...s
+    ];
+    console.log(`[FusionEngine] Combined ${l.length} entities from all sources`), console.log(`[FusionEngine] - Regex: ${c.length}`), console.log(`[FusionEngine] - Spatial: ${o.length}`), console.log(`[FusionEngine] - NLP: ${i.length}`), console.log(`[FusionEngine] - ML/Gemini: ${s.length}`);
+    const { deduplicated: d, removedCount: u } = X(
+      l,
+      this.config.deduplicationOverlapThreshold
+    );
+    console.log(`[FusionEngine] Deduplicated: removed ${u}, kept ${d.length}`);
+    const h = d.filter(
+      (m) => m.confidence >= this.config.confidenceThreshold
+    );
+    console.log(`[FusionEngine] After confidence filter (>=${this.config.confidenceThreshold}): ${h.length}`);
+    const p = this.generateStats(h, u);
+    return {
+      entities: h,
+      stats: p
+    };
+  }
+  /**
+   * Generates statistics about the fused entities.
+   */
+  generateStats(n, t) {
+    const a = {};
+    let r = 0, c = 0, o = 0;
+    for (const i of n)
+      a[i.layer] = (a[i.layer] || 0) + 1, i.confidence < 0.7 ? r++ : i.confidence < 0.9 ? c++ : o++;
+    return {
+      total: n.length,
+      byLayer: a,
+      byConfidenceRange: { low: r, medium: c, high: o },
+      deduplicated: t
+    };
+  }
+  /**
+   * Applies document-specific confidence boost.
+   */
+  applyDocumentTypeBoost(n) {
+    const t = this.getConfidenceBoost();
+    return t === 1 ? n : n.map((a) => ({
+      ...a,
+      confidence: Math.min(a.confidence * t, 1)
+    }));
+  }
+  getConfidenceBoost() {
+    switch (this.config.documentType) {
+      case "invoice":
+        return 1.2;
+      case "bank_statement":
+        return 1.3;
+      case "id_card_aadhaar":
+      case "id_card_pan":
+        return 1.25;
+      case "tax_return":
+        return 1.15;
+      default:
+        return 1;
+    }
+  }
+}
+const E = /\b(\d{4}[\s\-]?\d{4}[\s\-]?\d{4})\b/g;
+function q(e) {
+  const n = [];
+  let t;
+  for (E.lastIndex = 0; (t = E.exec(e)) !== null; ) {
+    const a = t[1], r = a.replace(/\D/g, "");
+    r.length === 12 && !/^0/.test(r) && !/^1/.test(r) && n.push({
+      type: "AADHAAR",
+      value: r,
+      startIndex: t.index,
+      endIndex: t.index + t[0].length,
+      raw: a
+    });
+  }
+  return n;
+}
+const A = /\b([A-Z]{5}[0-9]{4}[A-Z])\b/g;
+function J(e) {
+  const n = [];
+  let t;
+  for (A.lastIndex = 0; (t = A.exec(e)) !== null; )
+    n.push({
+      type: "PAN",
+      value: t[1],
+      startIndex: t.index,
+      endIndex: t.index + t[0].length,
+      raw: t[1]
+    });
+  return n;
+}
+const D = /\b(\d{4}[\s\-]?\d{4}[\s\-]?\d{4}[\s\-]?\d{1,7})\b/g;
+function Q(e) {
+  const n = [];
+  let t;
+  for (D.lastIndex = 0; (t = D.exec(e)) !== null; ) {
+    const a = t[1], r = a.replace(/\D/g, "");
+    r.length >= 13 && r.length <= 19 && n.push({
+      type: "CREDIT_CARD",
+      value: r,
+      startIndex: t.index,
+      endIndex: t.index + t[0].length,
+      raw: a
+    });
+  }
+  return n;
+}
+const M = /(?:\+91[\s\-]?|0)?([6-9]\d{4}[\s\-]?\d{5})\b/g;
+function ee(e) {
+  const n = [];
+  let t;
+  for (M.lastIndex = 0; (t = M.exec(e)) !== null; ) {
+    const a = t[0], r = a.replace(/\D/g, "");
+    r.length >= 10 && r.length <= 12 && n.push({
+      type: "PHONE",
+      value: r,
+      startIndex: t.index,
+      endIndex: t.index + t[0].length,
+      raw: a
+    });
+  }
+  return n;
+}
+function te(e) {
+  return [
+    ...q(e),
+    ...J(e),
+    ...Q(e),
+    ...ee(e)
+  ];
+}
+const ne = [
+  [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
+  [1, 2, 3, 4, 0, 6, 7, 8, 9, 5],
+  [2, 3, 4, 0, 1, 7, 8, 9, 5, 6],
+  [3, 4, 0, 1, 2, 8, 9, 5, 6, 7],
+  [4, 0, 1, 2, 3, 9, 5, 6, 7, 8],
+  [5, 9, 8, 7, 6, 0, 4, 3, 2, 1],
+  [6, 5, 9, 8, 7, 1, 0, 4, 3, 2],
+  [7, 6, 5, 9, 8, 2, 1, 0, 4, 3],
+  [8, 7, 6, 5, 9, 3, 2, 1, 0, 4],
+  [9, 8, 7, 6, 5, 4, 3, 2, 1, 0]
+], oe = [
+  [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
+  [1, 5, 7, 6, 2, 8, 3, 0, 9, 4],
+  [5, 8, 0, 3, 7, 9, 6, 1, 4, 2],
+  [8, 9, 1, 6, 0, 4, 3, 5, 2, 7],
+  [9, 4, 5, 3, 1, 2, 6, 8, 7, 0],
+  [4, 2, 8, 6, 5, 7, 3, 9, 0, 1],
+  [2, 7, 9, 3, 8, 0, 6, 4, 1, 5],
+  [7, 0, 4, 6, 9, 1, 3, 2, 5, 8]
+];
+function ae(e) {
+  const n = e.replace(/\D/g, "");
+  if (n.length !== 12) return !1;
+  let t = 0;
+  const a = n.length;
+  for (let r = 0; r < a; r++) {
+    const c = parseInt(n[a - r - 1]), o = oe[r % 8][c];
+    t = ne[t][o];
+  }
+  return t === 0;
+}
+function ce(e) {
+  const n = e.replace(/\D/g, "");
+  if (n.length < 13 || n.length > 19) return !1;
+  let t = 0, a = !1;
+  for (let r = n.length - 1; r >= 0; r--) {
+    let c = parseInt(n[r]);
+    a && (c *= 2, c > 9 && (c -= 9)), t += c, a = !a;
+  }
+  return t % 10 === 0;
+}
+function re(e) {
+  if (!/^[A-Z]{5}[0-9]{4}[A-Z]$/.test(e)) return !1;
+  const t = e[3];
+  return "ABCFGHLJPT".includes(t);
+}
+function ie(e, n, t = 0) {
+  const a = te(e), r = [];
+  for (const c of a) {
+    let o = 0.85, i = !0;
+    switch (c.type) {
+      case "AADHAAR":
+        i = ae(c.value), o = i ? 1 : 0.6;
+        break;
+      case "CREDIT_CARD":
+        i = ce(c.value), o = i ? 1 : 0.5;
+        break;
+      case "PAN":
+        i = re(c.value), o = i ? 1 : 0.7;
+        break;
+      case "PHONE":
+        o = 0.9;
+        break;
+    }
+    if (!i && o < 0.5) continue;
+    const s = se(c.startIndex, c.endIndex, e, n, t);
+    r.push({
+      id: "l1_" + crypto.randomUUID().substring(0, 8),
+      type: c.type,
+      value: c.raw,
+      confidence: o,
+      bbox: s,
+      masked: !0,
+      layer: 1
+    });
+  }
+  return r;
+}
+function x(e, n) {
+  const t = Math.min(...e.map((o) => o.bbox.x)), a = Math.min(...e.map((o) => o.bbox.y)), r = Math.max(...e.map((o) => o.bbox.x + o.bbox.w)), c = Math.max(...e.map((o) => o.bbox.y + o.bbox.h));
+  return { x: t, y: a, w: r - t, h: c - a, pageIndex: n };
+}
+function se(e, n, t, a, r) {
+  const o = t.substring(e, n).trim().split(/\s+/).filter((l) => l.length > 0);
+  if (o.length > 0 && a.length > 0) {
+    const l = (u) => u.replace(/[^a-zA-Z0-9]/g, ""), d = o.map(l).filter((u) => u.length > 0);
+    if (d.length > 0) {
+      for (let h = 0; h <= a.length - d.length; h++) {
+        let p = !0;
+        for (let m = 0; m < d.length; m++)
+          if (l(a[h + m].text) !== d[m]) {
+            p = !1;
+            break;
+          }
+        if (p)
+          return x(
+            a.slice(h, h + d.length),
+            r
+          );
+      }
+      const u = d[0];
+      for (let h = 0; h < a.length; h++) {
+        if (l(a[h].text) !== u) continue;
+        const p = [a[h]];
+        let m = 1;
+        for (let f = h + 1; f < a.length && m < d.length; f++)
+          l(a[f].text) === d[m] && (p.push(a[f]), m++);
+        if (p.length >= Math.ceil(d.length / 2))
+          return x(p, r);
+      }
+    }
+  }
+  let i = 0;
+  const s = [];
+  for (const l of a) {
+    const d = t.indexOf(l.text, i);
+    if (d === -1) continue;
+    const u = d + l.text.length;
+    if (u > e && d < n && s.push(l), i = u, d > n) break;
+  }
+  return s.length > 0 ? x(s, r) : { x: 0, y: 0, w: 100, h: 20, pageIndex: r };
+}
+const R = /* @__PURE__ */ new Set([
+  "aarav",
+  "aditi",
+  "aditya",
+  "akash",
+  "amit",
+  "amita",
+  "ananya",
+  "anil",
+  "anita",
+  "anjali",
+  "ankita",
+  "arjun",
+  "arun",
+  "aruna",
+  "ashok",
+  "bhavna",
+  "chandra",
+  "deepak",
+  "deepika",
+  "rahul",
+  "rajesh",
+  "priya",
+  "neha",
+  "vikram",
+  "sneha",
+  "pooja",
+  "rohit",
+  "john",
+  "james",
+  "mary",
+  "patricia",
+  "jennifer",
+  "michael",
+  "william",
+  "david",
+  "sarah",
+  "karen"
+]), T = /* @__PURE__ */ new Set([
+  "sharma",
+  "verma",
+  "gupta",
+  "singh",
+  "kumar",
+  "patel",
+  "joshi",
+  "mishra",
+  "agarwal",
+  "mehta",
+  "reddy",
+  "rao",
+  "nair",
+  "menon",
+  "iyer",
+  "mukherjee",
+  "chatterjee",
+  "das",
+  "roy",
+  "shah",
+  "smith",
+  "johnson",
+  "williams",
+  "brown",
+  "jones",
+  "davis",
+  "miller",
+  "wilson",
+  "moore"
+]);
+function le(e) {
+  const n = [];
+  for (let t = 0; t < e.length; t++) {
+    const a = e[t].text.replace(/[^a-zA-Z]/g, "").toLowerCase();
+    if (!(a.length < 2) && (R.has(a) || T.has(a))) {
+      let r = e[t].text, c = { ...e[t].bbox }, o = 0.65;
+      if (t + 1 < e.length) {
+        const i = e[t + 1].text.replace(/[^a-zA-Z]/g, "").toLowerCase();
+        (R.has(i) || T.has(i)) && (r += " " + e[t + 1].text, c.w = e[t + 1].bbox.x + e[t + 1].bbox.w - c.x, o = 0.78, t++);
+      }
+      n.push({
+        id: "nlp_" + crypto.randomUUID().substring(0, 8),
+        type: "NAME",
+        value: r,
+        confidence: o,
+        bbox: c,
+        masked: !0,
+        layer: 2
+        // NLP heuristic layer
+      });
+    }
+  }
+  return n;
+}
+self.onmessage = async (e) => {
+  const { type: n, fullText: t, words: a, pageIndex: r, confidenceThreshold: c } = e.data;
+  if (n === "ADVANCED_DETECT")
+    try {
+      console.log("[Advanced Detection Worker] Starting multi-stage detection..."), console.log(`[Advanced Detection Worker] Input: ${a.length} words, ${t.length} chars`);
+      const o = U(t), i = G(o.primaryType);
+      console.log(`[Advanced Detection Worker] Document Type: ${o.primaryType} (${Math.round(o.confidence * 100)}%)`), console.log("[Advanced Detection Worker] Ruleset:", i);
+      let s = [];
+      if (i.prioritizeSpatialContext) {
+        console.log("[Advanced Detection Worker] Running spatial analysis...");
+        const f = S(a);
+        console.log(`[Advanced Detection Worker] Grouped into ${f.length} lines`);
+        const C = w(f, r);
+        console.log(`[Advanced Detection Worker] Grouped into ${C.length} blocks`);
+        const y = $(f, r);
+        if (console.log(`[Advanced Detection Worker] Detected ${y.length} key-value pairs`), s = O(y), i.enableTableDetection) {
+          const N = L(f);
+          console.log(`[Advanced Detection Worker] Detected ${N.length} table columns`);
+        }
+      }
+      console.log("[Advanced Detection Worker] Running regex detection...");
+      const l = ie(t, a, r);
+      console.log(`[Advanced Detection Worker] Regex detected: ${l.length}`);
+      let d = [];
+      i.skipParagraphNER || (console.log("[Advanced Detection Worker] Running NLP heuristics..."), d = le(a), console.log(`[Advanced Detection Worker] NLP detected: ${d.length}`)), console.log("[Advanced Detection Worker] Fusing detections...");
+      const u = new Z({
+        documentType: o.primaryType,
+        confidenceThreshold: c,
+        preferSpatialOverNLP: i.prioritizeSpatialContext,
+        deduplicationOverlapThreshold: 0.5
+      }), h = u.fuse(
+        l,
+        s,
+        d,
+        []
+        // ML/Gemini entities come from main pipeline
+      );
+      console.log(`[Advanced Detection Worker] Fusion complete: ${h.entities.length} entities`), console.log("[Advanced Detection Worker] Stats:", h.stats);
+      const m = {
+        type: "DETECTION_RESULT",
+        entities: u.applyDocumentTypeBoost(h.entities),
+        documentType: o.primaryType,
+        stats: {
+          ...h.stats,
+          documentConfidence: Math.round(o.confidence * 100),
+          detectedKeywords: o.detectedKeywords.slice(0, 5)
+        }
+      };
+      self.postMessage(m), l.length = 0, s.length = 0, d.length = 0;
+    } catch (o) {
+      console.error("[Advanced Detection Worker] Error:", o), self.postMessage({
+        type: "DETECTION_ERROR",
+        error: o instanceof Error ? o.message : "Detection failed"
+      });
+    }
+};
+self.postMessage({ type: "WORKER_READY" });
+//# sourceMappingURL=advanced.worker-DSVrF0gl.js.map