{ "$schema": "https://ui.shadcn.com/schema/registry-item.json", "name": "synthetic-data-generator", "title": "Synthetic Data Generator", "type": "registry:block", "files": [ { "path": "packages/registry-react/blocks/synthetic-data-generator/SyntheticDataGenerator.tsx", "content": "'use client'\n\nimport * as React from 'react'\nimport {\n Check,\n Code2,\n Cpu,\n Database,\n Download,\n FileSpreadsheet,\n Info,\n Layers,\n Lock,\n Plus,\n RefreshCw,\n ShieldCheck,\n Sliders,\n Sparkles,\n Trash2,\n} from 'lucide-react'\nimport { cn } from '@/lib/utils'\nimport { Badge } from '@/components/ui/badge'\nimport { Button } from '@/components/ui/button'\nimport { Card, CardContent, CardDescription, CardHeader, CardTitle } from '@/components/ui/card'\nimport { Input } from '@/components/ui/input'\nimport { Select, SelectContent, SelectItem, SelectTrigger, SelectValue } from '@/components/ui/select'\nimport { Separator } from '@/components/ui/separator'\nimport { Slider } from '@/components/ui/slider'\nimport { Switch } from '@/components/ui/switch'\nimport type {\n FieldDataType,\n FidelityMetrics,\n GeneratedRecord,\n SchemaField,\n SyntheticDataGeneratorProps,\n TemplatePreset,\n} from './synthetic-data-types'\nimport {\n boxMullerRandom as boxMullerRandomHelper,\n CARD_BRANDS,\n FIRST_NAMES,\n LAST_NAMES,\n TEMPLATES,\n} from './synthetic-data-templates'\nimport { SyntheticDataPreview } from './SyntheticDataPreview'\n\nexport type {\n FieldDataType,\n FidelityMetrics,\n GeneratedRecord,\n SchemaField,\n SyntheticDataGeneratorProps,\n TemplatePreset,\n}\n\nexport function SyntheticDataGenerator({\n className,\n initialTemplate = 'customer-transactions',\n initialRowCount = 1000,\n initialEpsilon = 1.2,\n}: SyntheticDataGeneratorProps) {\n // --- STATE ---\n const [selectedTemplateKey, setSelectedTemplateKey] = React.useState(initialTemplate)\n const [outputFormat, setOutputFormat] = React.useState<'csv' | 'jsonl'>('csv')\n const [rowCount, setRowCount] = React.useState(initialRowCount)\n const [epsilon, setEpsilon] = React.useState(initialEpsilon)\n const [kAnonymity, setKAnonymity] = React.useState(true)\n const [saltedHash, setSaltedHash] = React.useState(true)\n const [noiseMechanism, setNoiseMechanism] = React.useState(true)\n\n const [activeFields, setActiveFields] = React.useState(() => {\n return TEMPLATES[initialTemplate]?.fields ? JSON.parse(JSON.stringify(TEMPLATES[initialTemplate].fields)) : []\n })\n\n const [isGenerating, setIsGenerating] = React.useState(false)\n const [copied, setCopied] = React.useState(false)\n const [downloadToast, setDownloadToast] = React.useState(null)\n const [previewTab, setPreviewTab] = React.useState<'table' | 'json' | 'metrics'>('table')\n const [generationSeed, setGenerationSeed] = React.useState(48291)\n\n // Field addition state\n const [showAddField, setShowAddField] = React.useState(false)\n const [newFieldName, setNewFieldName] = React.useState('')\n const [newFieldType, setNewFieldType] = React.useState('gaussian_float')\n\n // Handle template selection switch\n const handleTemplateChange = (key: string) => {\n setSelectedTemplateKey(key)\n const tpl = TEMPLATES[key]\n if (tpl) {\n setActiveFields(JSON.parse(JSON.stringify(tpl.fields)))\n triggerRegeneration()\n }\n }\n\n // Generation trigger\n const triggerRegeneration = () => {\n setIsGenerating(true)\n setGenerationSeed(Math.floor(Math.random() * 90000) + 10000)\n setTimeout(() => {\n setIsGenerating(false)\n }, 450)\n }\n\n // Box-Muller Gaussian random generator with differential privacy noise\n const boxMullerRandom = (mean: number, stdDev: number, seedOffset: number): number => {\n return boxMullerRandomHelper(mean, stdDev, seedOffset, generationSeed, noiseMechanism, epsilon)\n }\n\n // Dynamic preview records computation\n const previewRecords = React.useMemo(() => {\n const rows: GeneratedRecord[] = []\n const count = 6\n\n for (let i = 0; i < count; i++) {\n const fn = FIRST_NAMES[(i * 3 + generationSeed) % FIRST_NAMES.length]\n const ln = LAST_NAMES[(i * 5 + generationSeed + 2) % LAST_NAMES.length]\n const fullName = `${fn} ${ln}`\n const baseSlug = `${fn.charAt(0).toLowerCase()}.${ln.toLowerCase().replace(/[^a-z]/g, '')}`\n\n const record: GeneratedRecord = {\n id: saltedHash\n ? `syn_${((generationSeed * 9301 + i * 49297) % 233280).toString(16).padStart(6, '0')}`\n : `usr_${1000 + i}`,\n }\n\n for (const field of activeFields) {\n if (field.name.includes('id') || field.name.includes('uuid') || field.name.includes('token')) {\n record[field.name] = record.id\n continue\n }\n\n switch (field.type) {\n case 'name':\n record[field.name] = fullName\n break\n\n case 'email': {\n const dom = field.params.domain || 'synthetic-vault.io'\n record[field.name] = `${baseSlug}@${dom}`\n break\n }\n\n case 'credit_card': {\n const last4 = (((generationSeed * 7 + i * 1337 + 1000) % 9000) + 1000).toString()\n const brand = CARD_BRANDS[i % CARD_BRANDS.length]\n record[field.name] = `${brand} •••• ${last4}`\n break\n }\n\n case 'gaussian_float': {\n const m = field.params.mean ?? 100\n const s = field.params.stdDev ?? 20\n const rawVal = boxMullerRandom(m, s, i)\n const val = Math.max(m * 0.1, rawVal)\n\n if (field.params.currency) {\n record[field.name] = `${field.params.currency}${val.toFixed(2)}`\n } else if (field.params.unit) {\n record[field.name] = `${val.toFixed(1)} ${field.params.unit}`\n } else {\n record[field.name] = parseFloat(val.toFixed(2))\n }\n break\n }\n\n case 'category_weights': {\n const cats = field.params.categories || ['Standard', 'Premium']\n const catIndex = (i + Math.floor(generationSeed / 100)) % cats.length\n record[field.name] = cats[catIndex]\n break\n }\n\n case 'date_range': {\n const day = ((10 + i * 4) % 28) + 1\n const month = (i % 8) + 1\n const dayStr = day.toString().padStart(2, '0')\n const monthStr = month.toString().padStart(2, '0')\n const hour = (8 + i * 3) % 24\n const hourStr = hour.toString().padStart(2, '0')\n record[field.name] = `2025-${monthStr}-${dayStr} ${hourStr}:24:18 UTC`\n break\n }\n\n default:\n record[field.name] = `value_${i}`\n }\n }\n\n rows.push(record)\n }\n\n return rows\n }, [activeFields, generationSeed, saltedHash, noiseMechanism, epsilon])\n\n // Statistical fidelity metrics\n const fidelityMetrics = React.useMemo(() => {\n const rawSimilarity = Math.min(99.4, Math.max(82.0, 99.2 - ((5.0 - epsilon) / 4.9) * 9.6))\n const similarity = parseFloat(rawSimilarity.toFixed(1))\n const leakageRisk = kAnonymity && saltedHash && epsilon <= 3.5 ? 0.0 : Math.max(0, (epsilon - 3.5) * 0.04)\n const correlationRetention = parseFloat(Math.min(99.1, 98.6 - ((5.0 - epsilon) / 4.9) * 6.2).toFixed(1))\n const wassersteinDistance = parseFloat((0.032 + ((5.0 - epsilon) / 4.9) * 0.024).toFixed(3))\n const laplaceNoiseStd = (0.12 / Math.max(0.1, epsilon)).toFixed(3)\n\n let privacyTier = 'Balanced Privacy & Utility'\n let privacyBadgeColor = 'bg-info/10 text-info border-info/30'\n\n if (epsilon <= 0.5) {\n privacyTier = 'Maximum Privacy / High DP Noise'\n privacyBadgeColor = 'bg-success/10 text-success border-success/30'\n } else if (epsilon > 2.5) {\n privacyTier = 'High Statistical Utility / Minimal Noise'\n privacyBadgeColor = 'bg-warning/10 text-warning border-warning/30'\n }\n\n return {\n similarity,\n leakageRisk: leakageRisk.toFixed(2),\n correlationRetention,\n wassersteinDistance,\n laplaceNoiseStd,\n privacyTier,\n privacyBadgeColor,\n }\n }, [epsilon, kAnonymity, saltedHash])\n\n // Estimated file size calculation\n const estimatedFileSize = React.useMemo(() => {\n const bytesPerRow = outputFormat === 'csv' ? 180 : 380\n const totalKb = (rowCount * bytesPerRow) / 1024\n if (totalKb < 1024) {\n return `~${totalKb.toFixed(1)} KB`\n }\n return `~${(totalKb / 1024).toFixed(2)} MB`\n }, [outputFormat, rowCount])\n\n // Add field handler\n const handleAddCustomField = () => {\n if (!newFieldName.trim()) return\n\n const cleanName = newFieldName.trim().toLowerCase().replace(/\\s+/g, '_')\n const newField: SchemaField = {\n id: `f-custom-${Date.now()}`,\n name: cleanName,\n label: newFieldName.trim(),\n type: newFieldType,\n description: `Custom ${newFieldType} schema column`,\n params:\n newFieldType === 'gaussian_float'\n ? { mean: 500, stdDev: 120, currency: '$' }\n : newFieldType === 'category_weights'\n ? { categories: ['Alpha', 'Beta', 'Gamma'], weights: [0.5, 0.3, 0.2] }\n : {},\n }\n\n setActiveFields((prev) => [...prev, newField])\n setNewFieldName('')\n setShowAddField(false)\n triggerRegeneration()\n }\n\n // Remove field handler\n const removeField = (fieldId: string) => {\n if (activeFields.length <= 2) return\n setActiveFields((prev) => prev.filter((f) => f.id !== fieldId))\n triggerRegeneration()\n }\n\n // Copy JSON lines\n const copyJsonLines = async () => {\n try {\n const jsonLinesText = previewRecords.map((r) => JSON.stringify(r)).join('\\n')\n if (navigator?.clipboard?.writeText) {\n await navigator.clipboard.writeText(jsonLinesText)\n }\n setCopied(true)\n setTimeout(() => {\n setCopied(false)\n }, 2000)\n } catch (err) {\n console.error('Failed to copy to clipboard', err)\n }\n }\n\n // Download simulation\n const handleDownloadDataset = () => {\n const filename = `synthetic_${selectedTemplateKey}_${rowCount}_rows.${outputFormat === 'csv' ? 'csv' : 'jsonl'}`\n setDownloadToast(`Downloaded ${rowCount.toLocaleString()} synthetic records (${filename})`)\n setTimeout(() => {\n setDownloadToast(null)\n }, 3500)\n }\n\n return (\n
\n {/* TOP HEADER HERO */}\n \n \n
\n
\n
\n
\n \n
\n \n Synthetic Dataset Generator & Privacy Engine\n \n \n \n Differential Privacy Enabled\n \n
\n \n Mathematical & LLM-guided synthetic tabular dataset synthesizer with configurable (ε, δ)-DP noise\n injection.\n \n
\n\n {/* Header Right Quick Actions */}\n
\n {/* Format Selector Segmented Control */}\n
\n setOutputFormat('csv')}\n >\n \n Tabular CSV\n \n setOutputFormat('jsonl')}\n >\n \n JSON Lines\n \n
\n\n {/* Download Button */}\n \n \n Download Dataset\n \n\n {/* Primary Generate Batch Button */}\n \n {isGenerating ? : }\n {isGenerating ? 'Synthesizing...' : 'Generate Synthetic Batch'}\n \n
\n
\n\n {/* Template Selector Bar */}\n
\n
\n \n Schema Template:\n
\n\n
\n \n
\n\n \n
\n \n \n\n {/* DOWNLOAD SUCCESS NOTIFICATION */}\n {downloadToast && (\n
\n
\n \n {downloadToast}\n
\n \n {estimatedFileSize}\n \n
\n )}\n\n {/* 2-COLUMN STUDIO GENERATOR */}\n
\n {/* LEFT COLUMN: SCHEMA & PRIVACY CONFIGURATION (40% / 5 Cols) */}\n
\n {/* CARD 1: ROW COUNT & BATCH SCALE */}\n \n \n
\n
\n
\n \n
\n Dataset Generation Volume\n
\n \n {rowCount.toLocaleString()} rows\n \n
\n \n\n \n {/* Row Slider */}\n
\n
\n 100 rows\n Estimated: {estimatedFileSize}\n 50,000 rows\n
\n\n setRowCount(val[0])}\n />\n
\n\n {/* Quick Preset Chips */}\n
\n Presets:\n {[100, 1000, 5000, 10000, 50000].map((preset) => (\n setRowCount(preset)}\n >\n {preset >= 1000 ? `${preset / 1000}k` : preset}\n \n ))}\n
\n \n \n\n {/* CARD 2: DIFFERENTIAL PRIVACY EPSILON (ε) CONTROLS */}\n \n \n
\n
\n
\n \n
\n
\n Differential Privacy Budget\n Laplace (ε, δ)-DP noise calibration\n
\n
\n \n ε = {epsilon.toFixed(1)}\n \n
\n \n\n \n {/* Epsilon Slider */}\n
\n
\n ε=0.1 (High Privacy)\n ε=5.0 (High Fidelity)\n
\n\n setEpsilon(val[0])}\n />\n
\n\n {/* Epsilon Interpretation Banner */}\n \n \n
\n
{fidelityMetrics.privacyTier}
\n

\n Laplace noise scale:{' '}\n σ=±{fidelityMetrics.laplaceNoiseStd}{' '}\n • Delta: δ=10⁻⁵ • Zero memorization\n guarantee.\n

\n
\n
\n\n {/* Quick Epsilon Presets */}\n
\n {[\n { eps: 0.5, label: 'Healthcare' },\n { eps: 1.2, label: 'Standard' },\n { eps: 3.0, label: 'Analytics' },\n ].map(({ eps, label }) => (\n setEpsilon(eps)}\n >\n
ε = {eps.toFixed(1)}
\n
{label}
\n \n ))}\n
\n \n \n\n {/* CARD 3: FIELD SCHEMAS GENERATOR */}\n \n \n
\n
\n
\n \n
\n
\n Schema Fields Generator\n \n {activeFields.length} active synthesizer distributions\n \n
\n
\n setShowAddField(!showAddField)}\n >\n \n {showAddField ? 'Cancel' : 'Add Field'}\n \n
\n \n\n \n {/* Add Field Form */}\n {showAddField && (\n
\n
Configure New Synthetic Field
\n
\n
\n \n setNewFieldName(e.target.value)}\n placeholder=\"e.g. account_balance\"\n className=\"h-8 text-xs\"\n />\n
\n
\n \n \n
\n
\n\n
\n \n \n Add to Schema\n \n
\n
\n )}\n\n {/* Fields List */}\n
\n {activeFields.map((field) => (\n \n
\n
\n {field.name}\n \n {field.type.replace('_', ' ')}\n \n
\n

\n {field.description || 'Configured synthesizer parameter'}\n

\n
\n\n removeField(field.id)}\n >\n \n \n
\n ))}\n
\n \n \n\n {/* CARD 4: PII ANONYMIZATION POLICY SWITCHES */}\n \n \n
\n
\n \n
\n
\n PII Anonymization Policy\n Zero linkage & re-identification guarantees\n
\n
\n \n\n \n {/* K-Anonymity Switch */}\n
\n
\n
K-Anonymity (k = 5 Cohort Enforcement)
\n

\n Generalizes quasi-identifiers so each demographic bucket contains ≥ 5 records.\n

\n
\n \n
\n\n \n\n {/* Salted Hash Switch */}\n
\n
\n
Salted HMAC-SHA256 for Primary Keys
\n

\n Deterministically pseudonymizes primary keys with ephemeral cryptographic salt.\n

\n
\n \n
\n\n \n\n {/* Noise Injection Switch */}\n
\n
\n
Differential Privacy Gaussian Noise
\n

\n Injects calibrated noise calibrated to global sensitivity (Δf / ε).\n

\n
\n \n
\n \n \n
\n\n {/* RIGHT COLUMN: LIVE GENERATED SYNTHETIC DATA PREVIEW (60% / 7 Cols) */}\n
\n \n
\n \n \n )\n}\n\nexport default SyntheticDataGenerator\n", "type": "registry:block", "target": "~/components/blocks/SyntheticDataGenerator.tsx" }, { "path": "packages/registry-react/blocks/synthetic-data-generator/SyntheticDataPreview.tsx", "content": "'use client'\n\nimport * as React from 'react'\nimport { BarChart3, Binary, Check, Code2, Copy, Shield, ShieldCheck, Table as TableIcon, Zap } from 'lucide-react'\nimport { Badge } from '@/components/ui/badge'\nimport { Button } from '@/components/ui/button'\nimport { Card, CardContent, CardDescription, CardFooter, CardHeader, CardTitle } from '@/components/ui/card'\nimport { Progress } from '@/components/ui/progress'\nimport { Table, TableBody, TableCell, TableHead, TableHeader, TableRow } from '@/components/ui/table'\nimport { Tabs, TabsList, TabsTrigger } from '@/components/ui/tabs'\nimport type { FidelityMetrics, GeneratedRecord, SchemaField } from './synthetic-data-types'\n\ninterface SyntheticDataPreviewProps {\n previewTab: 'table' | 'json' | 'metrics'\n onPreviewTabChange: (tab: 'table' | 'json' | 'metrics') => void\n previewRecords: GeneratedRecord[]\n activeFields: SchemaField[]\n fidelityMetrics: FidelityMetrics\n epsilon: number\n rowCount: number\n outputFormat: 'csv' | 'jsonl'\n generationSeed: number\n copied: boolean\n onCopyJsonLines: () => void\n}\n\nexport function SyntheticDataPreview({\n previewTab,\n onPreviewTabChange,\n previewRecords,\n activeFields,\n fidelityMetrics,\n epsilon,\n rowCount,\n outputFormat,\n generationSeed,\n copied,\n onCopyJsonLines,\n}: SyntheticDataPreviewProps) {\n return (\n
\n {/* TOP GAUGE BANNER: STATISTICAL FIDELITY VS PRIVACY */}\n \n
\n
\n
\n \n \n {fidelityMetrics.similarity}% Distribution Similarity • {fidelityMetrics.leakageRisk}% PII Leakage\n Risk\n \n
\n \n ε-DP Bound: ≤ {epsilon.toFixed(1)}\n \n
\n
\n\n \n {/* Metric 1: Fidelity Score */}\n
\n
\n Fidelity Score\n \n
\n
\n {fidelityMetrics.similarity}%\n
\n \n
\n\n {/* Metric 2: Correlation Preservation */}\n
\n
\n Covariance Retention\n \n
\n
\n {fidelityMetrics.correlationRetention}%\n
\n \n
\n\n {/* Metric 3: Wasserstein Distance */}\n
\n
\n Wasserstein Distance\n \n
\n
\n {fidelityMetrics.wassersteinDistance}\n
\n
Low error bound (< 0.05)
\n
\n \n \n\n {/* PREVIEW TABS CARD */}\n \n \n
\n
\n
\n \n
\n
\n Live Generated Synthetic Data Preview\n \n Real-time sampler showing first {previewRecords.length} synthesized records\n \n
\n
\n\n {/* View Switcher */}\n onPreviewTabChange(val as 'table' | 'json' | 'metrics')}\n className=\"w-auto\"\n >\n \n \n \n Table\n \n \n \n JSONL\n \n \n \n Marginals\n \n \n \n
\n \n\n \n {/* TAB 1: TABULAR PREVIEW */}\n {previewTab === 'table' && (\n
\n
\n
\n \n \n \n {activeFields.map((f) => (\n \n {f.name}\n \n ))}\n \n \n \n {previewRecords.map((row, idx) => (\n \n {activeFields.map((f) => (\n \n {f.name === 'id' ||\n f.name.includes('id') ||\n f.name.includes('uuid') ||\n f.name.includes('token') ? (\n {row[f.name]}\n ) : f.type === 'category_weights' ? (\n \n {row[f.name]}\n \n ) : f.type === 'gaussian_float' ? (\n \n {row[f.name]}\n \n ) : f.type === 'credit_card' ? (\n {row[f.name]}\n ) : f.type === 'email' ? (\n {row[f.name]}\n ) : (\n {row[f.name]}\n )}\n \n ))}\n \n ))}\n \n
\n
\n
\n\n
\n \n \n Showing {previewRecords.length} sample records of {rowCount.toLocaleString()} generated rows\n \n Seed: #{generationSeed}\n
\n
\n )}\n\n {/* TAB 2: JSON LINES PREVIEW */}\n {previewTab === 'json' && (\n
\n
\n Streaming JSON Lines formatted output:\n \n {copied ? : }\n {copied ? 'Copied JSONL' : 'Copy JSON Lines'}\n \n
\n\n
\n {previewRecords.map((rec, idx) => (\n
\n {idx + 1} \n {'{ '}\n {Object.entries(rec).map(([key, val], kIdx, arr) => (\n \n \"{key}\"\n : \n \n {typeof val === 'number' ? val : `\"${val}\"`}\n \n {kIdx < arr.length - 1 && , }\n \n ))}\n {' }'}\n
\n ))}\n
\n
\n )}\n\n {/* TAB 3: MARGINALS & FIDELITY DENSITY */}\n {previewTab === 'metrics' && (\n
\n
\n {/* Density Gauge 1 */}\n
\n
\n Gaussian Distribution Overlap\n \n 96.2%\n \n
\n

\n Kolmogorov-Smirnov distance between synthetic batch and continuous Gaussian ground truth.\n

\n \n
\n\n {/* Density Gauge 2 */}\n
\n
\n Categorical Frequency Prior\n \n 98.4%\n \n
\n

\n Total variation distance across categorical weight assignments.\n

\n \n
\n
\n\n {/* Mathematical Guarantees Breakdown */}\n
\n
\n \n Mathematical Privacy Guarantees • Formal Specs\n
\n
\n
\n \n Differential Privacy (ε = {epsilon.toFixed(1)}, δ = 10⁻⁵)\n
\n
\n \n K-Anonymity (k = 5) Cohort Clamping\n
\n
\n \n Zero Direct Row Memorization\n
\n
\n \n Ephemeral In-Memory Compute\n
\n
\n
\n
\n )}\n \n\n \n \n \n Powered by LLM Semantic Synthesis & Laplace Noise Engine\n \n \n Output: {outputFormat.toUpperCase()} • {activeFields.length} fields • {rowCount.toLocaleString()}{' '}\n records\n \n \n \n
\n )\n}\n", "type": "registry:block", "target": "~/components/blocks/SyntheticDataPreview.tsx" }, { "path": "packages/registry-react/blocks/synthetic-data-generator/synthetic-data-types.ts", "content": "export type FieldDataType = 'name' | 'email' | 'credit_card' | 'gaussian_float' | 'date_range' | 'category_weights'\n\nexport interface SchemaField {\n id: string\n name: string\n label: string\n type: FieldDataType\n description?: string\n params: {\n mean?: number\n stdDev?: number\n currency?: string\n unit?: string\n categories?: string[]\n weights?: number[]\n domain?: string\n dateStart?: string\n dateEnd?: string\n }\n}\n\nexport interface TemplatePreset {\n id: string\n name: string\n description: string\n category: string\n fields: SchemaField[]\n}\n\nexport interface GeneratedRecord {\n id: string\n [key: string]: string | number\n}\n\nexport interface FidelityMetrics {\n similarity: number\n leakageRisk: string\n correlationRetention: number\n wassersteinDistance: number\n laplaceNoiseStd: string\n privacyTier: string\n privacyBadgeColor: string\n}\n\nexport interface SyntheticDataGeneratorProps {\n className?: string\n initialTemplate?: string\n initialRowCount?: number\n initialEpsilon?: number\n}\n", "type": "registry:block", "target": "~/components/blocks/synthetic-data-types.ts" }, { "path": "packages/registry-react/blocks/synthetic-data-generator/synthetic-data-templates.ts", "content": "import type { TemplatePreset } from './synthetic-data-types'\n\nexport const TEMPLATES: Record = {\n 'customer-transactions': {\n id: 'customer-transactions',\n name: 'Customer Transactions & Demographics Dataset',\n description:\n 'High-dimensional e-commerce customer cohort, synthetic PII, spending distribution, and country codes.',\n category: 'E-Commerce & Retail',\n fields: [\n {\n id: 'f-id',\n name: 'cust_id',\n label: 'Customer ID',\n type: 'name',\n description: 'Salted pseudonymized primary key',\n params: {},\n },\n {\n id: 'f-name',\n name: 'full_name',\n label: 'Full Name',\n type: 'name',\n description: 'Synthetic person name generator',\n params: {},\n },\n {\n id: 'f-email',\n name: 'email_address',\n label: 'Synthetic Email',\n type: 'email',\n description: 'Faker email with domain isolation',\n params: { domain: 'syn-corp.io' },\n },\n {\n id: 'f-card',\n name: 'masked_card',\n label: 'Card PAN',\n type: 'credit_card',\n description: 'Luhn-compliant masked payment token',\n params: {},\n },\n {\n id: 'f-spend',\n name: 'spending_amount',\n label: 'Spending Amount',\n type: 'gaussian_float',\n description: 'Gaussian spending curve (μ=$342.50, σ=$85.20)',\n params: { mean: 342.5, stdDev: 85.2, currency: '$' },\n },\n {\n id: 'f-country',\n name: 'country_code',\n label: 'Country Region',\n type: 'category_weights',\n description: 'Categorical distribution [US, DE, GB, FR, JP]',\n params: {\n categories: ['US', 'DE', 'GB', 'FR', 'JP'],\n weights: [0.45, 0.2, 0.15, 0.1, 0.1],\n },\n },\n {\n id: 'f-created',\n name: 'created_at',\n label: 'Registration Date',\n type: 'date_range',\n description: 'ISO-8601 timestamp range (2024-2025)',\n params: { dateStart: '2024-01-01', dateEnd: '2025-08-20' },\n },\n ],\n },\n 'healthcare-ehr': {\n id: 'healthcare-ehr',\n name: 'Healthcare Patient Clinical Records (EHR)',\n description: 'HIPAA-compliant patient cohort with biomarker telemetry, blood pressure, and triage severity tiers.',\n category: 'Healthcare & Life Sciences',\n fields: [\n {\n id: 'f-h-id',\n name: 'patient_id',\n label: 'Patient ID',\n type: 'name',\n description: 'Deterministic HIPAA salted identifier',\n params: {},\n },\n {\n id: 'f-h-name',\n name: 'patient_name',\n label: 'Patient Name',\n type: 'name',\n description: 'Synthetic medical record moniker',\n params: {},\n },\n {\n id: 'f-h-email',\n name: 'contact_email',\n label: 'Contact Email',\n type: 'email',\n description: 'Synthetic provider contact',\n params: { domain: 'med-synthetic.org' },\n },\n {\n id: 'f-h-glucose',\n name: 'blood_glucose',\n label: 'Glucose (mg/dL)',\n type: 'gaussian_float',\n description: 'Fast fasting blood glucose (μ=108.4, σ=22.1)',\n params: { mean: 108.4, stdDev: 22.1, unit: 'mg/dL' },\n },\n {\n id: 'f-h-bp',\n name: 'systolic_bp',\n label: 'Systolic BP',\n type: 'gaussian_float',\n description: 'Arterial blood pressure (μ=124.0, σ=15.5)',\n params: { mean: 124.0, stdDev: 15.5, unit: 'mmHg' },\n },\n {\n id: 'f-h-tier',\n name: 'triage_tier',\n label: 'Triage Severity',\n type: 'category_weights',\n description: 'Clinical priority classification',\n params: {\n categories: ['Routine', 'Elevated', 'Urgent', 'Critical'],\n weights: [0.52, 0.28, 0.14, 0.06],\n },\n },\n {\n id: 'f-h-date',\n name: 'admission_date',\n label: 'Admission Date',\n type: 'date_range',\n description: 'Encounter timestamps',\n params: { dateStart: '2024-06-01', dateEnd: '2025-08-20' },\n },\n ],\n },\n 'fintech-fraud': {\n id: 'fintech-fraud',\n name: 'Fintech Fraud Detection & Risk Scoring',\n description: 'Transaction monitoring telemetry for anti-fraud classifiers with skewed anomalous patterns.',\n category: 'Fintech & Risk',\n fields: [\n {\n id: 'f-ff-id',\n name: 'tx_token',\n label: 'Transaction ID',\n type: 'name',\n description: 'Cryptographic ledger token',\n params: {},\n },\n {\n id: 'f-ff-name',\n name: 'merchant_name',\n label: 'Merchant Name',\n type: 'name',\n description: 'Synthetic counterparty vendor',\n params: {},\n },\n {\n id: 'f-ff-email',\n name: 'auth_email',\n label: 'Auth Email',\n type: 'email',\n description: 'Tokenized transaction origin email',\n params: { domain: 'syn-pay.net' },\n },\n {\n id: 'f-ff-card',\n name: 'card_pan',\n label: 'Card PAN',\n type: 'credit_card',\n description: 'Tokenized 16-digit card surrogate',\n params: {},\n },\n {\n id: 'f-ff-amt',\n name: 'settlement_usd',\n label: 'Settlement ($)',\n type: 'gaussian_float',\n description: 'Transaction gross volume (μ=$1420.00, σ=$480.00)',\n params: { mean: 1420.0, stdDev: 480.0, currency: '$' },\n },\n {\n id: 'f-ff-verdict',\n name: 'risk_verdict',\n label: 'Risk Verdict',\n type: 'category_weights',\n description: 'Ground truth classifier target',\n params: {\n categories: ['Legitimate', 'Suspicious', 'Flagged_Fraud'],\n weights: [0.89, 0.08, 0.03],\n },\n },\n {\n id: 'f-ff-ts',\n name: 'settled_at',\n label: 'Settlement Time',\n type: 'date_range',\n description: 'High-precision event timestamp',\n params: { dateStart: '2025-01-01', dateEnd: '2025-08-20' },\n },\n ],\n },\n 'saas-telemetry': {\n id: 'saas-telemetry',\n name: 'SaaS Product Telemetry & Usage Metrics',\n description: 'B2B subscription telemetry with active seat distribution, recurring revenue, and subscription tiers.',\n category: 'Cloud Software & B2B',\n fields: [\n {\n id: 'f-st-id',\n name: 'org_uuid',\n label: 'Organization UUID',\n type: 'name',\n description: 'Tenant identifier',\n params: {},\n },\n {\n id: 'f-st-name',\n name: 'tenant_name',\n label: 'Tenant Name',\n type: 'name',\n description: 'Synthetic enterprise tenant moniker',\n params: {},\n },\n {\n id: 'f-st-email',\n name: 'admin_email',\n label: 'Admin Email',\n type: 'email',\n description: 'Primary owner email',\n params: { domain: 'syn-saas.io' },\n },\n {\n id: 'f-st-seats',\n name: 'active_seats',\n label: 'Active Seats',\n type: 'gaussian_float',\n description: 'Provisioned enterprise seats (μ=48.0, σ=18.0)',\n params: { mean: 48.0, stdDev: 18.0, unit: 'seats' },\n },\n {\n id: 'f-st-mrr',\n name: 'mrr_usd',\n label: 'Monthly Revenue',\n type: 'gaussian_float',\n description: 'MRR ledger (μ=$2850.00, σ=$950.00)',\n params: { mean: 2850.0, stdDev: 950.0, currency: '$' },\n },\n {\n id: 'f-st-plan',\n name: 'plan_tier',\n label: 'Plan Tier',\n type: 'category_weights',\n description: 'Subscription entitlement level',\n params: {\n categories: ['Starter', 'Growth', 'Enterprise'],\n weights: [0.48, 0.36, 0.16],\n },\n },\n {\n id: 'f-st-date',\n name: 'provisioned_at',\n label: 'Provisioned Date',\n type: 'date_range',\n description: 'Cohort subscription activation',\n params: { dateStart: '2024-01-01', dateEnd: '2025-08-20' },\n },\n ],\n },\n}\n\nexport const FIRST_NAMES = [\n 'Elena',\n 'Malik',\n 'Siddharth',\n 'Claire',\n 'Liam',\n 'Amara',\n 'Kenji',\n 'Sofia',\n 'Devon',\n 'Maya',\n 'Tobias',\n 'Fatima',\n 'Mateo',\n 'Astrid',\n 'Zara',\n]\n\nexport const LAST_NAMES = [\n 'Rostova',\n 'Al-Mansoor',\n 'Patel',\n 'Chen',\n \"O'Connor\",\n 'Okafor',\n 'Sato',\n 'Alvarez',\n 'Vance',\n 'Lin',\n 'Richter',\n 'Zahra',\n 'Silva',\n 'Lindqvist',\n 'Novak',\n]\n\nexport const CARD_BRANDS = ['Visa', 'Mastercard', 'Amex']\n\nexport function boxMullerRandom(\n mean: number,\n stdDev: number,\n seedOffset: number,\n seed: number,\n noiseMechanism: boolean,\n epsilon: number,\n): number {\n const u1 = Math.max(1e-6, Math.abs(Math.sin(seed + seedOffset * 17.3)))\n const u2 = Math.max(1e-6, Math.abs(Math.cos(seed + seedOffset * 31.7)))\n const z0 = Math.sqrt(-2.0 * Math.log(u1)) * Math.cos(2.0 * Math.PI * u2)\n\n const dpNoiseScale = noiseMechanism ? stdDev / (Math.max(0.1, epsilon) * 12) : 0\n const dpNoise = Math.sin(seed * 3.7 + seedOffset * 5.1) * dpNoiseScale\n\n return mean + z0 * stdDev + dpNoise\n}\n", "type": "registry:block", "target": "~/components/blocks/synthetic-data-templates.ts" } ], "dependencies": [ "lucide-react" ], "devDependencies": [], "registryDependencies": [ "https://uipkge.dev/r/react/badge.json", "https://uipkge.dev/r/react/button.json", "https://uipkge.dev/r/react/card.json", "https://uipkge.dev/r/react/input.json", "https://uipkge.dev/r/react/progress.json", "https://uipkge.dev/r/react/select.json", "https://uipkge.dev/r/react/separator.json", "https://uipkge.dev/r/react/slider.json", "https://uipkge.dev/r/react/switch.json", "https://uipkge.dev/r/react/table.json", "https://uipkge.dev/r/react/tabs.json" ], "description": "Gretel and Tonic-style LLM and mathematical synthetic dataset generator with differential privacy epsilon controls, custom schema field builders, K-anonymity policy toggles, live tabular CSV preview, and streaming JSON Lines export.", "categories": [ "ai", "data", "app", "security" ] }