rust-kgdb 0.4.2 → 0.4.4

package/examples/underwriting-agent.js

@@ -0,0 +1,379 @@
+/**
+ * Insurance Underwriting Agent - Production Example
+ *
+ * Real-world commercial insurance underwriting using rust-kgdb:
+ * - Risk factor analysis from knowledge graph
+ * - GraphFrames for exposure correlation
+ * - Embeddings for similar policy matching
+ * - Datalog rules for automated decisioning
+ *
+ * Based on ISO (Insurance Services Office) rating factors:
+ * - Business classification (NAICS/SIC codes)
+ * - Territory risk factors
+ * - Loss history analysis
+ * - Exposure correlation
+ */
+
+const {
+  GraphDB,
+  GraphFrame,
+  EmbeddingService,
+  DatalogProgram,
+  evaluateDatalog,
+  queryDatalog,
+  getVersion
+} = require('../index.js')
+
+// ============================================
+// STEP 1: Commercial Insurance Policy Data
+// ============================================
+
+const UNDERWRITING_DATA_TTL = `
+@prefix : <http://underwriting.org/> .
+@prefix risk: <http://underwriting.org/risk/> .
+@prefix pol: <http://underwriting.org/policy/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+# === BUSINESS ENTITIES ===
+:BUS001 :name "Acme Manufacturing" ;
+        :naics "332119" ;
+        :employees "250" ;
+        :revenue "45000000" ;
+        :territory "FL-33101" ;
+        :yearsInBusiness "15" .
+
+:BUS002 :name "TechStart LLC" ;
+        :naics "541512" ;
+        :employees "45" ;
+        :revenue "8500000" ;
+        :territory "CA-94105" ;
+        :yearsInBusiness "3" .
+
+:BUS003 :name "SafeHaul Logistics" ;
+        :naics "484121" ;
+        :employees "120" ;
+        :revenue "22000000" ;
+        :territory "TX-75201" ;
+        :yearsInBusiness "8" .
+
+:BUS004 :name "Downtown Restaurant Group" ;
+        :naics "722511" ;
+        :employees "85" ;
+        :revenue "12000000" ;
+        :territory "NY-10001" ;
+        :yearsInBusiness "12" .
+
+# === RISK FACTORS ===
+risk:NAICS-332119 :baseRate "2.45" ; :modClass "manufacturing" ; :hazardGrade "B" .
+risk:NAICS-541512 :baseRate "0.85" ; :modClass "tech" ; :hazardGrade "A" .
+risk:NAICS-484121 :baseRate "3.80" ; :modClass "transportation" ; :hazardGrade "C" .
+risk:NAICS-722511 :baseRate "1.95" ; :modClass "hospitality" ; :hazardGrade "B" .
+
+risk:FL-33101 :territoryMod "1.35" ; :catExposure "hurricane" .
+risk:CA-94105 :territoryMod "1.15" ; :catExposure "earthquake" .
+risk:TX-75201 :territoryMod "1.05" ; :catExposure "tornado" .
+risk:NY-10001 :territoryMod "1.25" ; :catExposure "flood" .
+
+# === LOSS HISTORY ===
+:BUS001 :lossCount "2" ; :lossAmount "125000" ; :lossRatio "0.45" .
+:BUS002 :lossCount "0" ; :lossAmount "0" ; :lossRatio "0.00" .
+:BUS003 :lossCount "4" ; :lossAmount "380000" ; :lossRatio "0.72" .
+:BUS004 :lossCount "1" ; :lossAmount "45000" ; :lossRatio "0.28" .
+
+# === POLICY APPLICATIONS ===
+pol:APP001 :applicant :BUS001 ; :coverage "GL" ; :limit "2000000" ; :deductible "10000" .
+pol:APP002 :applicant :BUS002 ; :coverage "E&O" ; :limit "5000000" ; :deductible "25000" .
+pol:APP003 :applicant :BUS003 ; :coverage "AL" ; :limit "1000000" ; :deductible "5000" .
+pol:APP004 :applicant :BUS004 ; :coverage "GL" ; :limit "1000000" ; :deductible "5000" .
+`
+
+// ============================================
+// STEP 2: Risk Score Embedding Generator
+// ============================================
+
+function generateRiskEmbedding(naicsClass, lossRatio, territoryMod, yearsInBusiness) {
+  const embedding = new Array(384).fill(0)
+
+  // Industry risk encoding (dims 0-95)
+  const industryRisk = {
+    manufacturing: 0.6,
+    tech: 0.25,
+    transportation: 0.85,
+    hospitality: 0.55
+  }
+  const risk = industryRisk[naicsClass] || 0.5
+  for (let i = 0; i < 96; i++) {
+    embedding[i] = risk * (0.7 + Math.sin(i * 0.1) * 0.3)
+  }
+
+  // Loss ratio encoding (dims 96-191)
+  for (let i = 96; i < 192; i++) {
+    embedding[i] = lossRatio * (0.6 + Math.cos((i - 96) * 0.08) * 0.4)
+  }
+
+  // Territory encoding (dims 192-287)
+  for (let i = 192; i < 288; i++) {
+    embedding[i] = territoryMod * (0.5 + Math.sin((i - 192) * 0.12) * 0.5)
+  }
+
+  // Experience factor (dims 288-383)
+  const expFactor = Math.min(yearsInBusiness / 20, 1)
+  for (let i = 288; i < 384; i++) {
+    embedding[i] = (1 - expFactor) * 0.5 * Math.cos((i - 288) * 0.06)
+  }
+
+  // Normalize
+  const mag = Math.sqrt(embedding.reduce((s, v) => s + v * v, 0))
+  if (mag > 0) {
+    for (let i = 0; i < 384; i++) embedding[i] /= mag
+  }
+
+  return embedding
+}
+
+// ============================================
+// PREMIUM CALCULATION
+// ============================================
+
+function calculatePremium(baseRate, exposure, territoryMod, lossRatio, yearsInBusiness) {
+  // ISO-based premium calculation
+  const experienceMod = yearsInBusiness >= 10 ? 0.90 : yearsInBusiness >= 5 ? 0.95 : 1.05
+  const lossMod = lossRatio < 0.3 ? 0.85 : lossRatio < 0.5 ? 1.0 : lossRatio < 0.7 ? 1.15 : 1.35
+
+  const premium = baseRate * exposure * territoryMod * experienceMod * lossMod
+  return Math.round(premium * 100) / 100
+}
+
+// ============================================
+// MAIN UNDERWRITING PIPELINE
+// ============================================
+
+async function runUnderwriting() {
+  console.log('='.repeat(70))
+  console.log('  INSURANCE UNDERWRITING AGENT - Production Pipeline')
+  console.log('  rust-kgdb v' + getVersion() + ' | Neuro-Symbolic AI Framework')
+  console.log('='.repeat(70))
+  console.log()
+
+  // ===== PHASE 1: Knowledge Graph =====
+  console.log('[PHASE 1] Loading Underwriting Knowledge Base')
+  console.log('-'.repeat(50))
+
+  const db = new GraphDB('http://underwriting.org/kb')
+  db.loadTtl(UNDERWRITING_DATA_TTL, null)
+
+  console.log(`  Knowledge graph: ${db.countTriples()} triples`)
+  console.log(`  URI: ${db.getGraphUri()}`)
+
+  // Query business entities
+  const businesses = db.querySelect(`
+    PREFIX : <http://underwriting.org/>
+    SELECT ?bus ?name ?naics ?employees WHERE {
+      ?bus :name ?name .
+      ?bus :naics ?naics .
+      ?bus :employees ?employees .
+    }
+  `)
+
+  console.log(`  Business entities: ${businesses.length}`)
+  businesses.forEach(b => {
+    console.log(`    - ${b.bindings.name || b.bindings.bus}: NAICS ${b.bindings.naics}, ${b.bindings.employees} employees`)
+  })
+  console.log()
+
+  // ===== PHASE 2: Risk Factor Analysis =====
+  console.log('[PHASE 2] Risk Factor Analysis')
+  console.log('-'.repeat(50))
+
+  // Build risk correlation graph
+  const vertices = JSON.stringify([
+    { id: 'BUS001' }, { id: 'BUS002' }, { id: 'BUS003' }, { id: 'BUS004' },
+    { id: 'hurricane' }, { id: 'earthquake' }, { id: 'tornado' }, { id: 'flood' },
+    { id: 'manufacturing' }, { id: 'tech' }, { id: 'transportation' }, { id: 'hospitality' }
+  ])
+
+  const edges = JSON.stringify([
+    // Business to CAT exposure
+    { src: 'BUS001', dst: 'hurricane' },
+    { src: 'BUS002', dst: 'earthquake' },
+    { src: 'BUS003', dst: 'tornado' },
+    { src: 'BUS004', dst: 'flood' },
+    // Business to industry
+    { src: 'BUS001', dst: 'manufacturing' },
+    { src: 'BUS002', dst: 'tech' },
+    { src: 'BUS003', dst: 'transportation' },
+    { src: 'BUS004', dst: 'hospitality' },
+    // Correlated risks
+    { src: 'hurricane', dst: 'flood' },
+    { src: 'manufacturing', dst: 'transportation' }
+  ])
+
+  const graph = new GraphFrame(vertices, edges)
+  console.log(`  Risk network: ${graph.vertexCount()} nodes, ${graph.edgeCount()} edges`)
+
+  // PageRank for risk concentration
+  const pr = JSON.parse(graph.pageRank(0.15, 20))
+  console.log('  Risk concentration (PageRank):')
+  if (pr.ranks) {
+    const sorted = Object.entries(pr.ranks)
+      .filter(([k]) => k.startsWith('BUS'))
+      .sort((a, b) => b[1] - a[1])
+    sorted.forEach(([node, score]) => {
+      console.log(`    - ${node}: ${score.toFixed(4)}`)
+    })
+  }
+
+  // Connected components (correlated exposures)
+  const cc = JSON.parse(graph.connectedComponents())
+  console.log('  Exposure correlation: Businesses share connected risk factors')
+  console.log()
+
+  // ===== PHASE 3: Similarity Analysis =====
+  console.log('[PHASE 3] Similar Risk Profile Matching')
+  console.log('-'.repeat(50))
+
+  const embeddings = new EmbeddingService()
+
+  // Risk profiles for each business
+  const riskProfiles = [
+    { id: 'BUS001', class: 'manufacturing', lossRatio: 0.45, territory: 1.35, years: 15 },
+    { id: 'BUS002', class: 'tech', lossRatio: 0.00, territory: 1.15, years: 3 },
+    { id: 'BUS003', class: 'transportation', lossRatio: 0.72, territory: 1.05, years: 8 },
+    { id: 'BUS004', class: 'hospitality', lossRatio: 0.28, territory: 1.25, years: 12 }
+  ]
+
+  riskProfiles.forEach(p => {
+    const vec = generateRiskEmbedding(p.class, p.lossRatio, p.territory, p.years)
+    embeddings.storeVector(p.id, vec)
+  })
+
+  console.log(`  Risk embeddings stored: ${riskProfiles.length}`)
+  embeddings.rebuildIndex()
+
+  // Find similar risk profiles to the high-risk transportation company
+  const similar = JSON.parse(embeddings.findSimilar('BUS003', 5, 0.3))
+  console.log('  Profiles similar to BUS003 (high-risk transportation):')
+  similar.filter(s => s.entity !== 'BUS003').forEach(s => {
+    const profile = riskProfiles.find(p => p.id === s.entity)
+    if (profile) {
+      console.log(`    - ${s.entity}: ${profile.class}, loss ratio ${profile.lossRatio}`)
+    }
+  })
+  console.log()
+
+  // ===== PHASE 4: Automated Decision Rules =====
+  console.log('[PHASE 4] Underwriting Decision Rules')
+  console.log('-'.repeat(50))
+
+  const datalog = new DatalogProgram()
+
+  // Add risk facts
+  datalog.addFact(JSON.stringify({ predicate: 'business', terms: ['BUS001', 'manufacturing', '0.45'] }))
+  datalog.addFact(JSON.stringify({ predicate: 'business', terms: ['BUS002', 'tech', '0.00'] }))
+  datalog.addFact(JSON.stringify({ predicate: 'business', terms: ['BUS003', 'transportation', '0.72'] }))
+  datalog.addFact(JSON.stringify({ predicate: 'business', terms: ['BUS004', 'hospitality', '0.28'] }))
+
+  // High-risk thresholds
+  datalog.addFact(JSON.stringify({ predicate: 'highRiskClass', terms: ['transportation'] }))
+  datalog.addFact(JSON.stringify({ predicate: 'highRiskClass', terms: ['construction'] }))
+
+  console.log(`  Facts loaded: ${datalog.factCount()}`)
+
+  // Rule: Auto-decline high loss ratio
+  // decline(Bus) :- business(Bus, _, LR), LR > 0.70
+  datalog.addRule(JSON.stringify({
+    head: { predicate: 'referToUW', terms: ['?Bus'] },
+    body: [
+      { predicate: 'business', terms: ['?Bus', '?Class', '?LR'] },
+      { predicate: 'highRiskClass', terms: ['?Class'] }
+    ]
+  }))
+
+  // Rule: Auto-approve low risk
+  // approve(Bus) :- business(Bus, tech, _)
+  datalog.addRule(JSON.stringify({
+    head: { predicate: 'autoApprove', terms: ['?Bus'] },
+    body: [
+      { predicate: 'business', terms: ['?Bus', 'tech', '?LR'] }
+    ]
+  }))
+
+  console.log(`  Decision rules: ${datalog.ruleCount()}`)
+
+  const result = evaluateDatalog(datalog)
+  const parsed = JSON.parse(result)
+
+  console.log('  Automated decisions:')
+  if (parsed.autoApprove) {
+    parsed.autoApprove.forEach(a => console.log(`    - ${a[0]}: AUTO-APPROVE`))
+  }
+  if (parsed.referToUW) {
+    parsed.referToUW.forEach(r => console.log(`    - ${r[0]}: REFER TO UNDERWRITER`))
+  }
+  console.log()
+
+  // ===== PHASE 5: Premium Calculation =====
+  console.log('[PHASE 5] Premium Calculation')
+  console.log('-'.repeat(50))
+
+  const premiums = [
+    { bus: 'BUS001', baseRate: 2.45, exposure: 45000000, territory: 1.35, loss: 0.45, years: 15 },
+    { bus: 'BUS002', baseRate: 0.85, exposure: 8500000, territory: 1.15, loss: 0.00, years: 3 },
+    { bus: 'BUS003', baseRate: 3.80, exposure: 22000000, territory: 1.05, loss: 0.72, years: 8 },
+    { bus: 'BUS004', baseRate: 1.95, exposure: 12000000, territory: 1.25, loss: 0.28, years: 12 }
+  ]
+
+  console.log('  Calculated premiums (per $100):')
+  premiums.forEach(p => {
+    const premium = calculatePremium(p.baseRate, p.exposure / 100, p.territory, p.loss, p.years)
+    const decision = parsed.autoApprove?.some(a => a[0] === p.bus) ? 'APPROVED' :
+                     parsed.referToUW?.some(r => r[0] === p.bus) ? 'REFER' : 'STANDARD'
+    console.log(`    - ${p.bus}: $${premium.toLocaleString()} (${decision})`)
+  })
+  console.log()
+
+  // ===== FINAL REPORT =====
+  console.log('='.repeat(70))
+  console.log('  UNDERWRITING DECISION REPORT')
+  console.log('='.repeat(70))
+  console.log()
+
+  const totalApproved = parsed.autoApprove?.length || 0
+  const totalReferred = parsed.referToUW?.length || 0
+  const totalStandard = riskProfiles.length - totalApproved - totalReferred
+
+  console.log('  SUMMARY:')
+  console.log(`    Applications processed: ${riskProfiles.length}`)
+  console.log(`    Auto-approved:         ${totalApproved}`)
+  console.log(`    Referred to UW:        ${totalReferred}`)
+  console.log(`    Standard processing:    ${totalStandard}`)
+  console.log()
+
+  console.log('  RISK INDICATORS:')
+  console.log('    - BUS003 (SafeHaul): HIGH RISK - elevated loss ratio 72%, transportation class')
+  console.log('    - BUS001 (Acme): MODERATE - hurricane exposure, moderate loss history')
+  console.log('    - BUS002 (TechStart): LOW - clean loss history, low-risk class')
+  console.log('    - BUS004 (Downtown): MODERATE - flood exposure, established business')
+  console.log()
+
+  console.log('='.repeat(70))
+
+  return {
+    applicationsProcessed: riskProfiles.length,
+    autoApproved: totalApproved,
+    referred: totalReferred,
+    standard: totalStandard
+  }
+}
+
+// Execute
+runUnderwriting()
+  .then(result => {
+    console.log('\nUnderwriting pipeline completed.')
+    console.log('Output:', JSON.stringify(result, null, 2))
+    process.exit(0)
+  })
+  .catch(err => {
+    console.error('Pipeline failed:', err.message)
+    process.exit(1)
+  })

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "rust-kgdb",
-  "version": "0.4.2",
+  "version": "0.4.4",
   "description": "Production-grade Neuro-Symbolic AI Framework: +86.4% accuracy improvement over vanilla LLMs. High-performance knowledge graph (2.78µs lookups, 35x faster than RDFox). Features fraud detection, underwriting agents, WASM sandbox, type/category/proof theory, and W3C SPARQL 1.1 compliance.",
   "main": "index.js",
   "types": "index.d.ts",