dopple-ai 0.1.0

package/dist/chunk-KEWXLWAO.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/graph/query.ts"],"sourcesContent":["/**\n * Graph query utilities.\n *\n * Get neighborhoods, find patterns, compile persona-relevant context.\n */\n\nimport type { OceanVector, LLMProvider } from \"../types.js\";\nimport type { KnowledgeGraph, GraphNode, GraphEdge, GraphPattern } from \"./types.js\";\n\n/**\n * Get all nodes within N hops of a starting node.\n */\nexport function getNeighborhood(\n  graph: KnowledgeGraph,\n  nodeId: string,\n  depth: number = 2\n): GraphNode[] {\n  const visited = new Set<string>();\n  const queue: Array<{ id: string; d: number }> = [{ id: nodeId, d: 0 }];\n\n  while (queue.length > 0) {\n    const { id, d } = queue.shift()!;\n    if (visited.has(id) || d > depth) continue;\n    visited.add(id);\n\n    // Find connected nodes\n    for (const edge of graph.edges) {\n      if (edge.from === id && !visited.has(edge.to)) {\n        queue.push({ id: edge.to, d: d + 1 });\n      }\n      if (edge.to === id && !visited.has(edge.from)) {\n        queue.push({ id: edge.from, d: d + 1 });\n      }\n    }\n  }\n\n  return graph.nodes.filter((n) => visited.has(n.id));\n}\n\n/**\n * Find common patterns in the graph.\n * Looks for: frequent co-occurrences, hub nodes, churn patterns.\n */\nexport function findPatterns(graph: KnowledgeGraph): GraphPattern[] {\n  const patterns: GraphPattern[] = [];\n\n  // Find hub nodes (high degree)\n  const degree: Record<string, number> = {};\n  for (const edge of graph.edges) {\n    degree[edge.from] = (degree[edge.from] ?? 0) + 1;\n    degree[edge.to] = (degree[edge.to] ?? 0) + 1;\n  }\n\n  const hubs = Object.entries(degree)\n    .filter(([, d]) => d >= 3)\n    .sort((a, b) => b[1] - a[1])\n    .slice(0, 5);\n\n  for (const [nodeId, d] of hubs) {\n    const node = graph.nodes.find((n) => n.id === nodeId);\n    if (node) {\n      patterns.push({\n        description: `\"${node.label}\" is a hub node (${d} connections) — many users/entities relate to it`,\n        nodes: [nodeId],\n        edges: graph.edges.filter((e) => e.from === nodeId || e.to === nodeId).map((e) => e.id),\n        frequency: d,\n        significance: Math.min(1, d / 10),\n      });\n    }\n  }\n\n  // Find churn patterns (nodes connected to \"cancelled\" edges)\n  const churnEdges = graph.edges.filter((e) => e.type === \"cancelled\" || e.type === \"churned_because\");\n  if (churnEdges.length > 0) {\n    const churnReasons = churnEdges\n      .map((e) => {\n        const target = graph.nodes.find((n) => n.id === e.to);\n        return target?.label;\n      })\n      .filter(Boolean);\n\n    if (churnReasons.length > 0) {\n      patterns.push({\n        description: `Churn pattern: ${churnReasons.slice(0, 3).join(\", \")}`,\n        nodes: churnEdges.map((e) => e.to),\n        edges: churnEdges.map((e) => e.id),\n        frequency: churnEdges.length,\n        significance: Math.min(1, churnEdges.length / 5),\n      });\n    }\n  }\n\n  // Find complaint clusters\n  const complaintNodes = graph.nodes.filter((n) => n.type === \"complaint\");\n  if (complaintNodes.length >= 2) {\n    patterns.push({\n      description: `${complaintNodes.length} complaints identified: ${complaintNodes.slice(0, 3).map((n) => `\"${n.label}\"`).join(\", \")}`,\n      nodes: complaintNodes.map((n) => n.id),\n      edges: [],\n      frequency: complaintNodes.length,\n      significance: Math.min(1, complaintNodes.length / 5),\n    });\n  }\n\n  // Find feature usage patterns\n  const eventNodes = graph.nodes.filter((n) => n.type === \"event\");\n  const topEvents = eventNodes\n    .sort((a, b) => (Number(b.properties.totalOccurrences) ?? 0) - (Number(a.properties.totalOccurrences) ?? 0))\n    .slice(0, 5);\n\n  if (topEvents.length >= 2) {\n    patterns.push({\n      description: `Top features: ${topEvents.map((n) => `${n.label} (${n.properties.totalOccurrences}x)`).join(\", \")}`,\n      nodes: topEvents.map((n) => n.id),\n      edges: [],\n      frequency: topEvents.length,\n      significance: 0.7,\n    });\n  }\n\n  return patterns.sort((a, b) => b.significance - a.significance);\n}\n\n/**\n * Get persona-relevant context from the graph.\n *\n * Given a persona's OCEAN traits, find the most relevant graph neighborhood\n * and compile it into context strings the persona can reference.\n *\n * High openness personas → get innovation/feature nodes\n * High neuroticism personas → get complaint/churn nodes\n * High conscientiousness personas → get metric/reliability nodes\n */\nexport function getPersonaContext(\n  graph: KnowledgeGraph,\n  traits: OceanVector\n): string[] {\n  const context: string[] = [];\n\n  // Everyone gets the key metrics\n  const metricNodes = graph.nodes.filter((n) => n.type === \"metric\");\n  for (const m of metricNodes) {\n    context.push(m.label);\n  }\n\n  // High openness → features, innovation, new things\n  if (traits.openness > 0.6) {\n    const featureNodes = graph.nodes.filter(\n      (n) => n.type === \"feature\" || n.type === \"event\"\n    );\n    for (const f of featureNodes.slice(0, 5)) {\n      const count = f.properties.totalOccurrences;\n      context.push(\n        count\n          ? `Feature \"${f.label}\" is used ${count} times`\n          : `Feature: ${f.label}`\n      );\n    }\n  }\n\n  // High neuroticism → complaints, churn, negative signals\n  if (traits.neuroticism > 0.6) {\n    const complaints = graph.nodes.filter((n) => n.type === \"complaint\");\n    for (const c of complaints.slice(0, 5)) {\n      context.push(`User complaint: \"${c.label}\"`);\n    }\n\n    const churnEdges = graph.edges.filter(\n      (e) => e.type === \"cancelled\" || e.type === \"churned_because\"\n    );\n    if (churnEdges.length > 0) {\n      context.push(`${churnEdges.length} users have cancelled`);\n    }\n  }\n\n  // High conscientiousness → data, metrics, reliability\n  if (traits.conscientiousness > 0.6) {\n    const discovered = graph.edges.filter(\n      (e) => e.properties.discovered === true\n    );\n    for (const d of discovered.slice(0, 3)) {\n      const from = graph.nodes.find((n) => n.id === d.from);\n      const to = graph.nodes.find((n) => n.id === d.to);\n      if (from && to) {\n        context.push(`Pattern: \"${from.label}\" ${d.type} \"${to.label}\"`);\n      }\n    }\n  }\n\n  // High agreeableness → social proof, what others do\n  if (traits.agreeableness > 0.6) {\n    const hubEdges = Object.entries(\n      graph.edges.reduce(\n        (acc, e) => {\n          acc[e.to] = (acc[e.to] ?? 0) + 1;\n          return acc;\n        },\n        {} as Record<string, number>\n      )\n    )\n      .sort((a, b) => b[1] - a[1])\n      .slice(0, 3);\n\n    for (const [nodeId, count] of hubEdges) {\n      const node = graph.nodes.find((n) => n.id === nodeId);\n      if (node && count >= 2) {\n        context.push(`${count} users interact with \"${node.label}\"`);\n      }\n    }\n  }\n\n  // High extraversion → popular features, social features\n  if (traits.extraversion > 0.6) {\n    const socialEvents = graph.nodes.filter(\n      (n) =>\n        n.type === \"event\" &&\n        /share|invite|collab|team|social/i.test(n.label)\n    );\n    for (const s of socialEvents.slice(0, 3)) {\n      context.push(`Social feature: \"${s.label}\" (${s.properties.totalOccurrences ?? 0} uses)`);\n    }\n  }\n\n  return context;\n}\n\n/**\n * Human-readable graph summary.\n */\nexport function graphSummary(graph: KnowledgeGraph): string {\n  const typeCounts: Record<string, number> = {};\n  for (const node of graph.nodes) {\n    typeCounts[node.type] = (typeCounts[node.type] ?? 0) + 1;\n  }\n\n  const edgeTypes: Record<string, number> = {};\n  for (const edge of graph.edges) {\n    edgeTypes[edge.type] = (edgeTypes[edge.type] ?? 0) + 1;\n  }\n\n  const lines: string[] = [\n    `Knowledge graph: ${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`,\n    `Sources: ${graph.metadata.sources.join(\", \")}`,\n    `Node types: ${Object.entries(typeCounts).map(([t, c]) => `${t}(${c})`).join(\", \")}`,\n    `Edge types: ${Object.entries(edgeTypes).map(([t, c]) => `${t}(${c})`).join(\", \")}`,\n  ];\n\n  return lines.join(\"\\n\");\n}\n\n// ---------------------------------------------------------------------------\n// Natural language graph query\n// ---------------------------------------------------------------------------\n\nexport interface GraphQueryResult {\n  question: string;\n  answer: string;\n  evidence: Array<{\n    type: \"node\" | \"edge\" | \"pattern\";\n    label: string;\n    detail: string;\n    relevance: number;\n  }>;\n  confidence: number;\n  relatedNodes: string[];\n}\n\n/**\n * Ask the knowledge graph a natural language question.\n *\n * The LLM interprets the question, identifies relevant nodes/edges/patterns,\n * and synthesizes an answer grounded in the graph data.\n *\n * @example\n * ```ts\n * const result = await queryGraph(llm, graph, \"Why are users churning?\");\n * console.log(result.answer);     // \"Users are primarily churning due to...\"\n * console.log(result.evidence);   // specific nodes/edges that support the answer\n * ```\n */\nexport async function queryGraph(\n  llm: LLMProvider,\n  graph: KnowledgeGraph,\n  question: string\n): Promise<GraphQueryResult> {\n  // Serialize graph into a compact representation the LLM can reason over\n  const graphContext = serializeGraphForLLM(graph);\n  const patterns = findPatterns(graph);\n  const patternText = patterns.length > 0\n    ? `\\nDiscovered patterns:\\n${patterns.map((p) => `- ${p.description} (significance: ${p.significance.toFixed(2)})`).join(\"\\n\")}`\n    : \"\";\n\n  const result = await llm.generateJSON<{\n    answer: string;\n    evidence: Array<{\n      type: \"node\" | \"edge\" | \"pattern\";\n      label: string;\n      detail: string;\n      relevance: number;\n    }>;\n    confidence: number;\n    relatedNodes: string[];\n  }>(\n    `You are a data analyst answering questions about a product's user base using a knowledge graph. The graph contains real user data — users, events, features, complaints, payments, and their relationships.\n\nGround every claim in specific graph data. If the graph doesn't contain enough information to answer confidently, say so and explain what data would help.`,\n    `Knowledge Graph:\n${graphContext}\n${patternText}\n\nQuestion: \"${question}\"\n\nAnalyze the graph to answer this question. Return JSON:\n{\n  \"answer\": \"<2-4 sentence answer grounded in the graph data>\",\n  \"evidence\": [\n    {\"type\": \"node|edge|pattern\", \"label\": \"<entity or relationship>\", \"detail\": \"<how this supports the answer>\", \"relevance\": <0-1>}\n  ],\n  \"confidence\": <0.0-1.0 based on how much graph data supports the answer>,\n  \"relatedNodes\": [\"<labels of the most relevant nodes for follow-up questions>\"]\n}`\n  );\n\n  return {\n    question,\n    ...result,\n  };\n}\n\n/**\n * Serialize the graph into a compact text representation for LLM context.\n * Keeps it under ~4K tokens by summarizing large graphs.\n */\nfunction serializeGraphForLLM(graph: KnowledgeGraph): string {\n  const lines: string[] = [];\n  lines.push(`${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`);\n\n  // Group nodes by type\n  const byType: Record<string, GraphNode[]> = {};\n  for (const node of graph.nodes) {\n    if (!byType[node.type]) byType[node.type] = [];\n    byType[node.type].push(node);\n  }\n\n  for (const [type, nodes] of Object.entries(byType)) {\n    const limit = type === \"user\" ? 5 : 15; // Cap user nodes to avoid bloat\n    const shown = nodes.slice(0, limit);\n    const remaining = nodes.length - shown.length;\n\n    lines.push(`\\n[${type.toUpperCase()}] (${nodes.length} total)`);\n    for (const n of shown) {\n      const props = Object.entries(n.properties)\n        .filter(([k]) => !k.startsWith(\"_\") && k !== \"fullText\")\n        .slice(0, 3)\n        .map(([k, v]) => `${k}=${JSON.stringify(v)}`)\n        .join(\", \");\n      lines.push(`  - ${n.label}${props ? ` (${props})` : \"\"}`);\n    }\n    if (remaining > 0) {\n      lines.push(`  ... and ${remaining} more`);\n    }\n  }\n\n  // Show edges grouped by type\n  const edgesByType: Record<string, Array<{ from: string; to: string; weight: number }>> = {};\n  for (const edge of graph.edges) {\n    if (!edgesByType[edge.type]) edgesByType[edge.type] = [];\n    const from = graph.nodes.find((n) => n.id === edge.from);\n    const to = graph.nodes.find((n) => n.id === edge.to);\n    if (from && to) {\n      edgesByType[edge.type].push({ from: from.label, to: to.label, weight: edge.weight });\n    }\n  }\n\n  lines.push(\"\\n[RELATIONSHIPS]\");\n  for (const [type, edges] of Object.entries(edgesByType)) {\n    const shown = edges.slice(0, 8);\n    lines.push(`  ${type} (${edges.length} total):`);\n    for (const e of shown) {\n      lines.push(`    \"${e.from}\" → \"${e.to}\" (weight: ${e.weight.toFixed(1)})`);\n    }\n    if (edges.length > shown.length) {\n      lines.push(`    ... and ${edges.length - shown.length} more`);\n    }\n  }\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";AAYO,SAAS,gBACd,OACA,QACA,QAAgB,GACH;AACb,QAAM,UAAU,oBAAI,IAAY;AAChC,QAAM,QAA0C,CAAC,EAAE,IAAI,QAAQ,GAAG,EAAE,CAAC;AAErE,SAAO,MAAM,SAAS,GAAG;AACvB,UAAM,EAAE,IAAI,EAAE,IAAI,MAAM,MAAM;AAC9B,QAAI,QAAQ,IAAI,EAAE,KAAK,IAAI,MAAO;AAClC,YAAQ,IAAI,EAAE;AAGd,eAAW,QAAQ,MAAM,OAAO;AAC9B,UAAI,KAAK,SAAS,MAAM,CAAC,QAAQ,IAAI,KAAK,EAAE,GAAG;AAC7C,cAAM,KAAK,EAAE,IAAI,KAAK,IAAI,GAAG,IAAI,EAAE,CAAC;AAAA,MACtC;AACA,UAAI,KAAK,OAAO,MAAM,CAAC,QAAQ,IAAI,KAAK,IAAI,GAAG;AAC7C,cAAM,KAAK,EAAE,IAAI,KAAK,MAAM,GAAG,IAAI,EAAE,CAAC;AAAA,MACxC;AAAA,IACF;AAAA,EACF;AAEA,SAAO,MAAM,MAAM,OAAO,CAAC,MAAM,QAAQ,IAAI,EAAE,EAAE,CAAC;AACpD;AAMO,SAAS,aAAa,OAAuC;AAClE,QAAM,WAA2B,CAAC;AAGlC,QAAM,SAAiC,CAAC;AACxC,aAAW,QAAQ,MAAM,OAAO;AAC9B,WAAO,KAAK,IAAI,KAAK,OAAO,KAAK,IAAI,KAAK,KAAK;AAC/C,WAAO,KAAK,EAAE,KAAK,OAAO,KAAK,EAAE,KAAK,KAAK;AAAA,EAC7C;AAEA,QAAM,OAAO,OAAO,QAAQ,MAAM,EAC/B,OAAO,CAAC,CAAC,EAAE,CAAC,MAAM,KAAK,CAAC,EACxB,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,EAC1B,MAAM,GAAG,CAAC;AAEb,aAAW,CAAC,QAAQ,CAAC,KAAK,MAAM;AAC9B,UAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,MAAM;AACpD,QAAI,MAAM;AACR,eAAS,KAAK;AAAA,QACZ,aAAa,IAAI,KAAK,KAAK,oBAAoB,CAAC;AAAA,QAChD,OAAO,CAAC,MAAM;AAAA,QACd,OAAO,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,OAAO,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACtF,WAAW;AAAA,QACX,cAAc,KAAK,IAAI,GAAG,IAAI,EAAE;AAAA,MAClC,CAAC;AAAA,IACH;AAAA,EACF;AAGA,QAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,eAAe,EAAE,SAAS,iBAAiB;AACnG,MAAI,WAAW,SAAS,GAAG;AACzB,UAAM,eAAe,WAClB,IAAI,CAAC,MAAM;AACV,YAAM,SAAS,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE;AACpD,aAAO,QAAQ;AAAA,IACjB,CAAC,EACA,OAAO,OAAO;AAEjB,QAAI,aAAa,SAAS,GAAG;AAC3B,eAAS,KAAK;AAAA,QACZ,aAAa,kBAAkB,aAAa,MAAM,GAAG,CAAC,EAAE,KAAK,IAAI,CAAC;AAAA,QAClE,OAAO,WAAW,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACjC,OAAO,WAAW,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACjC,WAAW,WAAW;AAAA,QACtB,cAAc,KAAK,IAAI,GAAG,WAAW,SAAS,CAAC;AAAA,MACjD,CAAC;AAAA,IACH;AAAA,EACF;AAGA,QAAM,iBAAiB,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,WAAW;AACvE,MAAI,eAAe,UAAU,GAAG;AAC9B,aAAS,KAAK;AAAA,MACZ,aAAa,GAAG,eAAe,MAAM,2BAA2B,eAAe,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,MAAM,IAAI,EAAE,KAAK,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,MAChI,OAAO,eAAe,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,MACrC,OAAO,CAAC;AAAA,MACR,WAAW,eAAe;AAAA,MAC1B,cAAc,KAAK,IAAI,GAAG,eAAe,SAAS,CAAC;AAAA,IACrD,CAAC;AAAA,EACH;AAGA,QAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,OAAO;AAC/D,QAAM,YAAY,WACf,KAAK,CAAC,GAAG,OAAO,OAAO,EAAE,WAAW,gBAAgB,KAAK,MAAM,OAAO,EAAE,WAAW,gBAAgB,KAAK,EAAE,EAC1G,MAAM,GAAG,CAAC;AAEb,MAAI,UAAU,UAAU,GAAG;AACzB,aAAS,KAAK;AAAA,MACZ,aAAa,iBAAiB,UAAU,IAAI,CAAC,MAAM,GAAG,EAAE,KAAK,KAAK,EAAE,WAAW,gBAAgB,IAAI,EAAE,KAAK,IAAI,CAAC;AAAA,MAC/G,OAAO,UAAU,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,MAChC,OAAO,CAAC;AAAA,MACR,WAAW,UAAU;AAAA,MACrB,cAAc;AAAA,IAChB,CAAC;AAAA,EACH;AAEA,SAAO,SAAS,KAAK,CAAC,GAAG,MAAM,EAAE,eAAe,EAAE,YAAY;AAChE;AAYO,SAAS,kBACd,OACA,QACU;AACV,QAAM,UAAoB,CAAC;AAG3B,QAAM,cAAc,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,QAAQ;AACjE,aAAW,KAAK,aAAa;AAC3B,YAAQ,KAAK,EAAE,KAAK;AAAA,EACtB;AAGA,MAAI,OAAO,WAAW,KAAK;AACzB,UAAM,eAAe,MAAM,MAAM;AAAA,MAC/B,CAAC,MAAM,EAAE,SAAS,aAAa,EAAE,SAAS;AAAA,IAC5C;AACA,eAAW,KAAK,aAAa,MAAM,GAAG,CAAC,GAAG;AACxC,YAAM,QAAQ,EAAE,WAAW;AAC3B,cAAQ;AAAA,QACN,QACI,YAAY,EAAE,KAAK,aAAa,KAAK,WACrC,YAAY,EAAE,KAAK;AAAA,MACzB;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,cAAc,KAAK;AAC5B,UAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,WAAW;AACnE,eAAW,KAAK,WAAW,MAAM,GAAG,CAAC,GAAG;AACtC,cAAQ,KAAK,oBAAoB,EAAE,KAAK,GAAG;AAAA,IAC7C;AAEA,UAAM,aAAa,MAAM,MAAM;AAAA,MAC7B,CAAC,MAAM,EAAE,SAAS,eAAe,EAAE,SAAS;AAAA,IAC9C;AACA,QAAI,WAAW,SAAS,GAAG;AACzB,cAAQ,KAAK,GAAG,WAAW,MAAM,uBAAuB;AAAA,IAC1D;AAAA,EACF;AAGA,MAAI,OAAO,oBAAoB,KAAK;AAClC,UAAM,aAAa,MAAM,MAAM;AAAA,MAC7B,CAAC,MAAM,EAAE,WAAW,eAAe;AAAA,IACrC;AACA,eAAW,KAAK,WAAW,MAAM,GAAG,CAAC,GAAG;AACtC,YAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI;AACpD,YAAM,KAAK,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE;AAChD,UAAI,QAAQ,IAAI;AACd,gBAAQ,KAAK,aAAa,KAAK,KAAK,KAAK,EAAE,IAAI,KAAK,GAAG,KAAK,GAAG;AAAA,MACjE;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,gBAAgB,KAAK;AAC9B,UAAM,WAAW,OAAO;AAAA,MACtB,MAAM,MAAM;AAAA,QACV,CAAC,KAAK,MAAM;AACV,cAAI,EAAE,EAAE,KAAK,IAAI,EAAE,EAAE,KAAK,KAAK;AAC/B,iBAAO;AAAA,QACT;AAAA,QACA,CAAC;AAAA,MACH;AAAA,IACF,EACG,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,EAC1B,MAAM,GAAG,CAAC;AAEb,eAAW,CAAC,QAAQ,KAAK,KAAK,UAAU;AACtC,YAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,MAAM;AACpD,UAAI,QAAQ,SAAS,GAAG;AACtB,gBAAQ,KAAK,GAAG,KAAK,yBAAyB,KAAK,KAAK,GAAG;AAAA,MAC7D;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,eAAe,KAAK;AAC7B,UAAM,eAAe,MAAM,MAAM;AAAA,MAC/B,CAAC,MACC,EAAE,SAAS,WACX,mCAAmC,KAAK,EAAE,KAAK;AAAA,IACnD;AACA,eAAW,KAAK,aAAa,MAAM,GAAG,CAAC,GAAG;AACxC,cAAQ,KAAK,oBAAoB,EAAE,KAAK,MAAM,EAAE,WAAW,oBAAoB,CAAC,QAAQ;AAAA,IAC1F;AAAA,EACF;AAEA,SAAO;AACT;AAKO,SAAS,aAAa,OAA+B;AAC1D,QAAM,aAAqC,CAAC;AAC5C,aAAW,QAAQ,MAAM,OAAO;AAC9B,eAAW,KAAK,IAAI,KAAK,WAAW,KAAK,IAAI,KAAK,KAAK;AAAA,EACzD;AAEA,QAAM,YAAoC,CAAC;AAC3C,aAAW,QAAQ,MAAM,OAAO;AAC9B,cAAU,KAAK,IAAI,KAAK,UAAU,KAAK,IAAI,KAAK,KAAK;AAAA,EACvD;AAEA,QAAM,QAAkB;AAAA,IACtB,oBAAoB,MAAM,SAAS,SAAS,WAAW,MAAM,SAAS,SAAS;AAAA,IAC/E,YAAY,MAAM,SAAS,QAAQ,KAAK,IAAI,CAAC;AAAA,IAC7C,eAAe,OAAO,QAAQ,UAAU,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,IAClF,eAAe,OAAO,QAAQ,SAAS,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,EACnF;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;AAgCA,eAAsB,WACpB,KACA,OACA,UAC2B;AAE3B,QAAM,eAAe,qBAAqB,KAAK;AAC/C,QAAM,WAAW,aAAa,KAAK;AACnC,QAAM,cAAc,SAAS,SAAS,IAClC;AAAA;AAAA,EAA2B,SAAS,IAAI,CAAC,MAAM,KAAK,EAAE,WAAW,mBAAmB,EAAE,aAAa,QAAQ,CAAC,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC,KAC5H;AAEJ,QAAM,SAAS,MAAM,IAAI;AAAA,IAWvB;AAAA;AAAA;AAAA,IAGA;AAAA,EACF,YAAY;AAAA,EACZ,WAAW;AAAA;AAAA,aAEA,QAAQ;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWnB;AAEA,SAAO;AAAA,IACL;AAAA,IACA,GAAG;AAAA,EACL;AACF;AAMA,SAAS,qBAAqB,OAA+B;AAC3D,QAAM,QAAkB,CAAC;AACzB,QAAM,KAAK,GAAG,MAAM,SAAS,SAAS,WAAW,MAAM,SAAS,SAAS,QAAQ;AAGjF,QAAM,SAAsC,CAAC;AAC7C,aAAW,QAAQ,MAAM,OAAO;AAC9B,QAAI,CAAC,OAAO,KAAK,IAAI,EAAG,QAAO,KAAK,IAAI,IAAI,CAAC;AAC7C,WAAO,KAAK,IAAI,EAAE,KAAK,IAAI;AAAA,EAC7B;AAEA,aAAW,CAAC,MAAM,KAAK,KAAK,OAAO,QAAQ,MAAM,GAAG;AAClD,UAAM,QAAQ,SAAS,SAAS,IAAI;AACpC,UAAM,QAAQ,MAAM,MAAM,GAAG,KAAK;AAClC,UAAM,YAAY,MAAM,SAAS,MAAM;AAEvC,UAAM,KAAK;AAAA,GAAM,KAAK,YAAY,CAAC,MAAM,MAAM,MAAM,SAAS;AAC9D,eAAW,KAAK,OAAO;AACrB,YAAM,QAAQ,OAAO,QAAQ,EAAE,UAAU,EACtC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,KAAK,MAAM,UAAU,EACtD,MAAM,GAAG,CAAC,EACV,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,UAAU,CAAC,CAAC,EAAE,EAC3C,KAAK,IAAI;AACZ,YAAM,KAAK,OAAO,EAAE,KAAK,GAAG,QAAQ,KAAK,KAAK,MAAM,EAAE,EAAE;AAAA,IAC1D;AACA,QAAI,YAAY,GAAG;AACjB,YAAM,KAAK,aAAa,SAAS,OAAO;AAAA,IAC1C;AAAA,EACF;AAGA,QAAM,cAAmF,CAAC;AAC1F,aAAW,QAAQ,MAAM,OAAO;AAC9B,QAAI,CAAC,YAAY,KAAK,IAAI,EAAG,aAAY,KAAK,IAAI,IAAI,CAAC;AACvD,UAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,KAAK,IAAI;AACvD,UAAM,KAAK,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,KAAK,EAAE;AACnD,QAAI,QAAQ,IAAI;AACd,kBAAY,KAAK,IAAI,EAAE,KAAK,EAAE,MAAM,KAAK,OAAO,IAAI,GAAG,OAAO,QAAQ,KAAK,OAAO,CAAC;AAAA,IACrF;AAAA,EACF;AAEA,QAAM,KAAK,mBAAmB;AAC9B,aAAW,CAAC,MAAM,KAAK,KAAK,OAAO,QAAQ,WAAW,GAAG;AACvD,UAAM,QAAQ,MAAM,MAAM,GAAG,CAAC;AAC9B,UAAM,KAAK,KAAK,IAAI,KAAK,MAAM,MAAM,UAAU;AAC/C,eAAW,KAAK,OAAO;AACrB,YAAM,KAAK,QAAQ,EAAE,IAAI,aAAQ,EAAE,EAAE,cAAc,EAAE,OAAO,QAAQ,CAAC,CAAC,GAAG;AAAA,IAC3E;AACA,QAAI,MAAM,SAAS,MAAM,QAAQ;AAC/B,YAAM,KAAK,eAAe,MAAM,SAAS,MAAM,MAAM,OAAO;AAAA,IAC9D;AAAA,EACF;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/chunk-PGZVVIL6.js

@@ -0,0 +1,249 @@
+#!/usr/bin/env node
+
+// src/graph/query.ts
+function getNeighborhood(graph, nodeId, depth = 2) {
+  const visited = /* @__PURE__ */ new Set();
+  const queue = [{ id: nodeId, d: 0 }];
+  while (queue.length > 0) {
+    const { id, d } = queue.shift();
+    if (visited.has(id) || d > depth) continue;
+    visited.add(id);
+    for (const edge of graph.edges) {
+      if (edge.from === id && !visited.has(edge.to)) {
+        queue.push({ id: edge.to, d: d + 1 });
+      }
+      if (edge.to === id && !visited.has(edge.from)) {
+        queue.push({ id: edge.from, d: d + 1 });
+      }
+    }
+  }
+  return graph.nodes.filter((n) => visited.has(n.id));
+}
+function findPatterns(graph) {
+  const patterns = [];
+  const degree = {};
+  for (const edge of graph.edges) {
+    degree[edge.from] = (degree[edge.from] ?? 0) + 1;
+    degree[edge.to] = (degree[edge.to] ?? 0) + 1;
+  }
+  const hubs = Object.entries(degree).filter(([, d]) => d >= 3).sort((a, b) => b[1] - a[1]).slice(0, 5);
+  for (const [nodeId, d] of hubs) {
+    const node = graph.nodes.find((n) => n.id === nodeId);
+    if (node) {
+      patterns.push({
+        description: `"${node.label}" is a hub node (${d} connections) \u2014 many users/entities relate to it`,
+        nodes: [nodeId],
+        edges: graph.edges.filter((e) => e.from === nodeId || e.to === nodeId).map((e) => e.id),
+        frequency: d,
+        significance: Math.min(1, d / 10)
+      });
+    }
+  }
+  const churnEdges = graph.edges.filter((e) => e.type === "cancelled" || e.type === "churned_because");
+  if (churnEdges.length > 0) {
+    const churnReasons = churnEdges.map((e) => {
+      const target = graph.nodes.find((n) => n.id === e.to);
+      return target?.label;
+    }).filter(Boolean);
+    if (churnReasons.length > 0) {
+      patterns.push({
+        description: `Churn pattern: ${churnReasons.slice(0, 3).join(", ")}`,
+        nodes: churnEdges.map((e) => e.to),
+        edges: churnEdges.map((e) => e.id),
+        frequency: churnEdges.length,
+        significance: Math.min(1, churnEdges.length / 5)
+      });
+    }
+  }
+  const complaintNodes = graph.nodes.filter((n) => n.type === "complaint");
+  if (complaintNodes.length >= 2) {
+    patterns.push({
+      description: `${complaintNodes.length} complaints identified: ${complaintNodes.slice(0, 3).map((n) => `"${n.label}"`).join(", ")}`,
+      nodes: complaintNodes.map((n) => n.id),
+      edges: [],
+      frequency: complaintNodes.length,
+      significance: Math.min(1, complaintNodes.length / 5)
+    });
+  }
+  const eventNodes = graph.nodes.filter((n) => n.type === "event");
+  const topEvents = eventNodes.sort((a, b) => (Number(b.properties.totalOccurrences) ?? 0) - (Number(a.properties.totalOccurrences) ?? 0)).slice(0, 5);
+  if (topEvents.length >= 2) {
+    patterns.push({
+      description: `Top features: ${topEvents.map((n) => `${n.label} (${n.properties.totalOccurrences}x)`).join(", ")}`,
+      nodes: topEvents.map((n) => n.id),
+      edges: [],
+      frequency: topEvents.length,
+      significance: 0.7
+    });
+  }
+  return patterns.sort((a, b) => b.significance - a.significance);
+}
+function getPersonaContext(graph, traits) {
+  const context = [];
+  const metricNodes = graph.nodes.filter((n) => n.type === "metric");
+  for (const m of metricNodes) {
+    context.push(m.label);
+  }
+  if (traits.openness > 0.6) {
+    const featureNodes = graph.nodes.filter(
+      (n) => n.type === "feature" || n.type === "event"
+    );
+    for (const f of featureNodes.slice(0, 5)) {
+      const count = f.properties.totalOccurrences;
+      context.push(
+        count ? `Feature "${f.label}" is used ${count} times` : `Feature: ${f.label}`
+      );
+    }
+  }
+  if (traits.neuroticism > 0.6) {
+    const complaints = graph.nodes.filter((n) => n.type === "complaint");
+    for (const c of complaints.slice(0, 5)) {
+      context.push(`User complaint: "${c.label}"`);
+    }
+    const churnEdges = graph.edges.filter(
+      (e) => e.type === "cancelled" || e.type === "churned_because"
+    );
+    if (churnEdges.length > 0) {
+      context.push(`${churnEdges.length} users have cancelled`);
+    }
+  }
+  if (traits.conscientiousness > 0.6) {
+    const discovered = graph.edges.filter(
+      (e) => e.properties.discovered === true
+    );
+    for (const d of discovered.slice(0, 3)) {
+      const from = graph.nodes.find((n) => n.id === d.from);
+      const to = graph.nodes.find((n) => n.id === d.to);
+      if (from && to) {
+        context.push(`Pattern: "${from.label}" ${d.type} "${to.label}"`);
+      }
+    }
+  }
+  if (traits.agreeableness > 0.6) {
+    const hubEdges = Object.entries(
+      graph.edges.reduce(
+        (acc, e) => {
+          acc[e.to] = (acc[e.to] ?? 0) + 1;
+          return acc;
+        },
+        {}
+      )
+    ).sort((a, b) => b[1] - a[1]).slice(0, 3);
+    for (const [nodeId, count] of hubEdges) {
+      const node = graph.nodes.find((n) => n.id === nodeId);
+      if (node && count >= 2) {
+        context.push(`${count} users interact with "${node.label}"`);
+      }
+    }
+  }
+  if (traits.extraversion > 0.6) {
+    const socialEvents = graph.nodes.filter(
+      (n) => n.type === "event" && /share|invite|collab|team|social/i.test(n.label)
+    );
+    for (const s of socialEvents.slice(0, 3)) {
+      context.push(`Social feature: "${s.label}" (${s.properties.totalOccurrences ?? 0} uses)`);
+    }
+  }
+  return context;
+}
+function graphSummary(graph) {
+  const typeCounts = {};
+  for (const node of graph.nodes) {
+    typeCounts[node.type] = (typeCounts[node.type] ?? 0) + 1;
+  }
+  const edgeTypes = {};
+  for (const edge of graph.edges) {
+    edgeTypes[edge.type] = (edgeTypes[edge.type] ?? 0) + 1;
+  }
+  const lines = [
+    `Knowledge graph: ${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`,
+    `Sources: ${graph.metadata.sources.join(", ")}`,
+    `Node types: ${Object.entries(typeCounts).map(([t, c]) => `${t}(${c})`).join(", ")}`,
+    `Edge types: ${Object.entries(edgeTypes).map(([t, c]) => `${t}(${c})`).join(", ")}`
+  ];
+  return lines.join("\n");
+}
+async function queryGraph(llm, graph, question) {
+  const graphContext = serializeGraphForLLM(graph);
+  const patterns = findPatterns(graph);
+  const patternText = patterns.length > 0 ? `
+Discovered patterns:
+${patterns.map((p) => `- ${p.description} (significance: ${p.significance.toFixed(2)})`).join("\n")}` : "";
+  const result = await llm.generateJSON(
+    `You are a data analyst answering questions about a product's user base using a knowledge graph. The graph contains real user data \u2014 users, events, features, complaints, payments, and their relationships.
+
+Ground every claim in specific graph data. If the graph doesn't contain enough information to answer confidently, say so and explain what data would help.`,
+    `Knowledge Graph:
+${graphContext}
+${patternText}
+
+Question: "${question}"
+
+Analyze the graph to answer this question. Return JSON:
+{
+  "answer": "<2-4 sentence answer grounded in the graph data>",
+  "evidence": [
+    {"type": "node|edge|pattern", "label": "<entity or relationship>", "detail": "<how this supports the answer>", "relevance": <0-1>}
+  ],
+  "confidence": <0.0-1.0 based on how much graph data supports the answer>,
+  "relatedNodes": ["<labels of the most relevant nodes for follow-up questions>"]
+}`
+  );
+  return {
+    question,
+    ...result
+  };
+}
+function serializeGraphForLLM(graph) {
+  const lines = [];
+  lines.push(`${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`);
+  const byType = {};
+  for (const node of graph.nodes) {
+    if (!byType[node.type]) byType[node.type] = [];
+    byType[node.type].push(node);
+  }
+  for (const [type, nodes] of Object.entries(byType)) {
+    const limit = type === "user" ? 5 : 15;
+    const shown = nodes.slice(0, limit);
+    const remaining = nodes.length - shown.length;
+    lines.push(`
+[${type.toUpperCase()}] (${nodes.length} total)`);
+    for (const n of shown) {
+      const props = Object.entries(n.properties).filter(([k]) => !k.startsWith("_") && k !== "fullText").slice(0, 3).map(([k, v]) => `${k}=${JSON.stringify(v)}`).join(", ");
+      lines.push(`  - ${n.label}${props ? ` (${props})` : ""}`);
+    }
+    if (remaining > 0) {
+      lines.push(`  ... and ${remaining} more`);
+    }
+  }
+  const edgesByType = {};
+  for (const edge of graph.edges) {
+    if (!edgesByType[edge.type]) edgesByType[edge.type] = [];
+    const from = graph.nodes.find((n) => n.id === edge.from);
+    const to = graph.nodes.find((n) => n.id === edge.to);
+    if (from && to) {
+      edgesByType[edge.type].push({ from: from.label, to: to.label, weight: edge.weight });
+    }
+  }
+  lines.push("\n[RELATIONSHIPS]");
+  for (const [type, edges] of Object.entries(edgesByType)) {
+    const shown = edges.slice(0, 8);
+    lines.push(`  ${type} (${edges.length} total):`);
+    for (const e of shown) {
+      lines.push(`    "${e.from}" \u2192 "${e.to}" (weight: ${e.weight.toFixed(1)})`);
+    }
+    if (edges.length > shown.length) {
+      lines.push(`    ... and ${edges.length - shown.length} more`);
+    }
+  }
+  return lines.join("\n");
+}
+
+export {
+  getNeighborhood,
+  findPatterns,
+  getPersonaContext,
+  graphSummary,
+  queryGraph
+};
+//# sourceMappingURL=chunk-PGZVVIL6.js.map

package/dist/chunk-PGZVVIL6.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/graph/query.ts"],"sourcesContent":["/**\n * Graph query utilities.\n *\n * Get neighborhoods, find patterns, compile persona-relevant context.\n */\n\nimport type { OceanVector, LLMProvider } from \"../types.js\";\nimport type { KnowledgeGraph, GraphNode, GraphEdge, GraphPattern } from \"./types.js\";\n\n/**\n * Get all nodes within N hops of a starting node.\n */\nexport function getNeighborhood(\n  graph: KnowledgeGraph,\n  nodeId: string,\n  depth: number = 2\n): GraphNode[] {\n  const visited = new Set<string>();\n  const queue: Array<{ id: string; d: number }> = [{ id: nodeId, d: 0 }];\n\n  while (queue.length > 0) {\n    const { id, d } = queue.shift()!;\n    if (visited.has(id) || d > depth) continue;\n    visited.add(id);\n\n    // Find connected nodes\n    for (const edge of graph.edges) {\n      if (edge.from === id && !visited.has(edge.to)) {\n        queue.push({ id: edge.to, d: d + 1 });\n      }\n      if (edge.to === id && !visited.has(edge.from)) {\n        queue.push({ id: edge.from, d: d + 1 });\n      }\n    }\n  }\n\n  return graph.nodes.filter((n) => visited.has(n.id));\n}\n\n/**\n * Find common patterns in the graph.\n * Looks for: frequent co-occurrences, hub nodes, churn patterns.\n */\nexport function findPatterns(graph: KnowledgeGraph): GraphPattern[] {\n  const patterns: GraphPattern[] = [];\n\n  // Find hub nodes (high degree)\n  const degree: Record<string, number> = {};\n  for (const edge of graph.edges) {\n    degree[edge.from] = (degree[edge.from] ?? 0) + 1;\n    degree[edge.to] = (degree[edge.to] ?? 0) + 1;\n  }\n\n  const hubs = Object.entries(degree)\n    .filter(([, d]) => d >= 3)\n    .sort((a, b) => b[1] - a[1])\n    .slice(0, 5);\n\n  for (const [nodeId, d] of hubs) {\n    const node = graph.nodes.find((n) => n.id === nodeId);\n    if (node) {\n      patterns.push({\n        description: `\"${node.label}\" is a hub node (${d} connections) — many users/entities relate to it`,\n        nodes: [nodeId],\n        edges: graph.edges.filter((e) => e.from === nodeId || e.to === nodeId).map((e) => e.id),\n        frequency: d,\n        significance: Math.min(1, d / 10),\n      });\n    }\n  }\n\n  // Find churn patterns (nodes connected to \"cancelled\" edges)\n  const churnEdges = graph.edges.filter((e) => e.type === \"cancelled\" || e.type === \"churned_because\");\n  if (churnEdges.length > 0) {\n    const churnReasons = churnEdges\n      .map((e) => {\n        const target = graph.nodes.find((n) => n.id === e.to);\n        return target?.label;\n      })\n      .filter(Boolean);\n\n    if (churnReasons.length > 0) {\n      patterns.push({\n        description: `Churn pattern: ${churnReasons.slice(0, 3).join(\", \")}`,\n        nodes: churnEdges.map((e) => e.to),\n        edges: churnEdges.map((e) => e.id),\n        frequency: churnEdges.length,\n        significance: Math.min(1, churnEdges.length / 5),\n      });\n    }\n  }\n\n  // Find complaint clusters\n  const complaintNodes = graph.nodes.filter((n) => n.type === \"complaint\");\n  if (complaintNodes.length >= 2) {\n    patterns.push({\n      description: `${complaintNodes.length} complaints identified: ${complaintNodes.slice(0, 3).map((n) => `\"${n.label}\"`).join(\", \")}`,\n      nodes: complaintNodes.map((n) => n.id),\n      edges: [],\n      frequency: complaintNodes.length,\n      significance: Math.min(1, complaintNodes.length / 5),\n    });\n  }\n\n  // Find feature usage patterns\n  const eventNodes = graph.nodes.filter((n) => n.type === \"event\");\n  const topEvents = eventNodes\n    .sort((a, b) => (Number(b.properties.totalOccurrences) ?? 0) - (Number(a.properties.totalOccurrences) ?? 0))\n    .slice(0, 5);\n\n  if (topEvents.length >= 2) {\n    patterns.push({\n      description: `Top features: ${topEvents.map((n) => `${n.label} (${n.properties.totalOccurrences}x)`).join(\", \")}`,\n      nodes: topEvents.map((n) => n.id),\n      edges: [],\n      frequency: topEvents.length,\n      significance: 0.7,\n    });\n  }\n\n  return patterns.sort((a, b) => b.significance - a.significance);\n}\n\n/**\n * Get persona-relevant context from the graph.\n *\n * Given a persona's OCEAN traits, find the most relevant graph neighborhood\n * and compile it into context strings the persona can reference.\n *\n * High openness personas → get innovation/feature nodes\n * High neuroticism personas → get complaint/churn nodes\n * High conscientiousness personas → get metric/reliability nodes\n */\nexport function getPersonaContext(\n  graph: KnowledgeGraph,\n  traits: OceanVector\n): string[] {\n  const context: string[] = [];\n\n  // Everyone gets the key metrics\n  const metricNodes = graph.nodes.filter((n) => n.type === \"metric\");\n  for (const m of metricNodes) {\n    context.push(m.label);\n  }\n\n  // High openness → features, innovation, new things\n  if (traits.openness > 0.6) {\n    const featureNodes = graph.nodes.filter(\n      (n) => n.type === \"feature\" || n.type === \"event\"\n    );\n    for (const f of featureNodes.slice(0, 5)) {\n      const count = f.properties.totalOccurrences;\n      context.push(\n        count\n          ? `Feature \"${f.label}\" is used ${count} times`\n          : `Feature: ${f.label}`\n      );\n    }\n  }\n\n  // High neuroticism → complaints, churn, negative signals\n  if (traits.neuroticism > 0.6) {\n    const complaints = graph.nodes.filter((n) => n.type === \"complaint\");\n    for (const c of complaints.slice(0, 5)) {\n      context.push(`User complaint: \"${c.label}\"`);\n    }\n\n    const churnEdges = graph.edges.filter(\n      (e) => e.type === \"cancelled\" || e.type === \"churned_because\"\n    );\n    if (churnEdges.length > 0) {\n      context.push(`${churnEdges.length} users have cancelled`);\n    }\n  }\n\n  // High conscientiousness → data, metrics, reliability\n  if (traits.conscientiousness > 0.6) {\n    const discovered = graph.edges.filter(\n      (e) => e.properties.discovered === true\n    );\n    for (const d of discovered.slice(0, 3)) {\n      const from = graph.nodes.find((n) => n.id === d.from);\n      const to = graph.nodes.find((n) => n.id === d.to);\n      if (from && to) {\n        context.push(`Pattern: \"${from.label}\" ${d.type} \"${to.label}\"`);\n      }\n    }\n  }\n\n  // High agreeableness → social proof, what others do\n  if (traits.agreeableness > 0.6) {\n    const hubEdges = Object.entries(\n      graph.edges.reduce(\n        (acc, e) => {\n          acc[e.to] = (acc[e.to] ?? 0) + 1;\n          return acc;\n        },\n        {} as Record<string, number>\n      )\n    )\n      .sort((a, b) => b[1] - a[1])\n      .slice(0, 3);\n\n    for (const [nodeId, count] of hubEdges) {\n      const node = graph.nodes.find((n) => n.id === nodeId);\n      if (node && count >= 2) {\n        context.push(`${count} users interact with \"${node.label}\"`);\n      }\n    }\n  }\n\n  // High extraversion → popular features, social features\n  if (traits.extraversion > 0.6) {\n    const socialEvents = graph.nodes.filter(\n      (n) =>\n        n.type === \"event\" &&\n        /share|invite|collab|team|social/i.test(n.label)\n    );\n    for (const s of socialEvents.slice(0, 3)) {\n      context.push(`Social feature: \"${s.label}\" (${s.properties.totalOccurrences ?? 0} uses)`);\n    }\n  }\n\n  return context;\n}\n\n/**\n * Human-readable graph summary.\n */\nexport function graphSummary(graph: KnowledgeGraph): string {\n  const typeCounts: Record<string, number> = {};\n  for (const node of graph.nodes) {\n    typeCounts[node.type] = (typeCounts[node.type] ?? 0) + 1;\n  }\n\n  const edgeTypes: Record<string, number> = {};\n  for (const edge of graph.edges) {\n    edgeTypes[edge.type] = (edgeTypes[edge.type] ?? 0) + 1;\n  }\n\n  const lines: string[] = [\n    `Knowledge graph: ${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`,\n    `Sources: ${graph.metadata.sources.join(\", \")}`,\n    `Node types: ${Object.entries(typeCounts).map(([t, c]) => `${t}(${c})`).join(\", \")}`,\n    `Edge types: ${Object.entries(edgeTypes).map(([t, c]) => `${t}(${c})`).join(\", \")}`,\n  ];\n\n  return lines.join(\"\\n\");\n}\n\n// ---------------------------------------------------------------------------\n// Natural language graph query\n// ---------------------------------------------------------------------------\n\nexport interface GraphQueryResult {\n  question: string;\n  answer: string;\n  evidence: Array<{\n    type: \"node\" | \"edge\" | \"pattern\";\n    label: string;\n    detail: string;\n    relevance: number;\n  }>;\n  confidence: number;\n  relatedNodes: string[];\n}\n\n/**\n * Ask the knowledge graph a natural language question.\n *\n * The LLM interprets the question, identifies relevant nodes/edges/patterns,\n * and synthesizes an answer grounded in the graph data.\n *\n * @example\n * ```ts\n * const result = await queryGraph(llm, graph, \"Why are users churning?\");\n * console.log(result.answer);     // \"Users are primarily churning due to...\"\n * console.log(result.evidence);   // specific nodes/edges that support the answer\n * ```\n */\nexport async function queryGraph(\n  llm: LLMProvider,\n  graph: KnowledgeGraph,\n  question: string\n): Promise<GraphQueryResult> {\n  // Serialize graph into a compact representation the LLM can reason over\n  const graphContext = serializeGraphForLLM(graph);\n  const patterns = findPatterns(graph);\n  const patternText = patterns.length > 0\n    ? `\\nDiscovered patterns:\\n${patterns.map((p) => `- ${p.description} (significance: ${p.significance.toFixed(2)})`).join(\"\\n\")}`\n    : \"\";\n\n  const result = await llm.generateJSON<{\n    answer: string;\n    evidence: Array<{\n      type: \"node\" | \"edge\" | \"pattern\";\n      label: string;\n      detail: string;\n      relevance: number;\n    }>;\n    confidence: number;\n    relatedNodes: string[];\n  }>(\n    `You are a data analyst answering questions about a product's user base using a knowledge graph. The graph contains real user data — users, events, features, complaints, payments, and their relationships.\n\nGround every claim in specific graph data. If the graph doesn't contain enough information to answer confidently, say so and explain what data would help.`,\n    `Knowledge Graph:\n${graphContext}\n${patternText}\n\nQuestion: \"${question}\"\n\nAnalyze the graph to answer this question. Return JSON:\n{\n  \"answer\": \"<2-4 sentence answer grounded in the graph data>\",\n  \"evidence\": [\n    {\"type\": \"node|edge|pattern\", \"label\": \"<entity or relationship>\", \"detail\": \"<how this supports the answer>\", \"relevance\": <0-1>}\n  ],\n  \"confidence\": <0.0-1.0 based on how much graph data supports the answer>,\n  \"relatedNodes\": [\"<labels of the most relevant nodes for follow-up questions>\"]\n}`\n  );\n\n  return {\n    question,\n    ...result,\n  };\n}\n\n/**\n * Serialize the graph into a compact text representation for LLM context.\n * Keeps it under ~4K tokens by summarizing large graphs.\n */\nfunction serializeGraphForLLM(graph: KnowledgeGraph): string {\n  const lines: string[] = [];\n  lines.push(`${graph.metadata.nodeCount} nodes, ${graph.metadata.edgeCount} edges`);\n\n  // Group nodes by type\n  const byType: Record<string, GraphNode[]> = {};\n  for (const node of graph.nodes) {\n    if (!byType[node.type]) byType[node.type] = [];\n    byType[node.type].push(node);\n  }\n\n  for (const [type, nodes] of Object.entries(byType)) {\n    const limit = type === \"user\" ? 5 : 15; // Cap user nodes to avoid bloat\n    const shown = nodes.slice(0, limit);\n    const remaining = nodes.length - shown.length;\n\n    lines.push(`\\n[${type.toUpperCase()}] (${nodes.length} total)`);\n    for (const n of shown) {\n      const props = Object.entries(n.properties)\n        .filter(([k]) => !k.startsWith(\"_\") && k !== \"fullText\")\n        .slice(0, 3)\n        .map(([k, v]) => `${k}=${JSON.stringify(v)}`)\n        .join(\", \");\n      lines.push(`  - ${n.label}${props ? ` (${props})` : \"\"}`);\n    }\n    if (remaining > 0) {\n      lines.push(`  ... and ${remaining} more`);\n    }\n  }\n\n  // Show edges grouped by type\n  const edgesByType: Record<string, Array<{ from: string; to: string; weight: number }>> = {};\n  for (const edge of graph.edges) {\n    if (!edgesByType[edge.type]) edgesByType[edge.type] = [];\n    const from = graph.nodes.find((n) => n.id === edge.from);\n    const to = graph.nodes.find((n) => n.id === edge.to);\n    if (from && to) {\n      edgesByType[edge.type].push({ from: from.label, to: to.label, weight: edge.weight });\n    }\n  }\n\n  lines.push(\"\\n[RELATIONSHIPS]\");\n  for (const [type, edges] of Object.entries(edgesByType)) {\n    const shown = edges.slice(0, 8);\n    lines.push(`  ${type} (${edges.length} total):`);\n    for (const e of shown) {\n      lines.push(`    \"${e.from}\" → \"${e.to}\" (weight: ${e.weight.toFixed(1)})`);\n    }\n    if (edges.length > shown.length) {\n      lines.push(`    ... and ${edges.length - shown.length} more`);\n    }\n  }\n\n  return lines.join(\"\\n\");\n}\n"],"mappings":";;;AAYO,SAAS,gBACd,OACA,QACA,QAAgB,GACH;AACb,QAAM,UAAU,oBAAI,IAAY;AAChC,QAAM,QAA0C,CAAC,EAAE,IAAI,QAAQ,GAAG,EAAE,CAAC;AAErE,SAAO,MAAM,SAAS,GAAG;AACvB,UAAM,EAAE,IAAI,EAAE,IAAI,MAAM,MAAM;AAC9B,QAAI,QAAQ,IAAI,EAAE,KAAK,IAAI,MAAO;AAClC,YAAQ,IAAI,EAAE;AAGd,eAAW,QAAQ,MAAM,OAAO;AAC9B,UAAI,KAAK,SAAS,MAAM,CAAC,QAAQ,IAAI,KAAK,EAAE,GAAG;AAC7C,cAAM,KAAK,EAAE,IAAI,KAAK,IAAI,GAAG,IAAI,EAAE,CAAC;AAAA,MACtC;AACA,UAAI,KAAK,OAAO,MAAM,CAAC,QAAQ,IAAI,KAAK,IAAI,GAAG;AAC7C,cAAM,KAAK,EAAE,IAAI,KAAK,MAAM,GAAG,IAAI,EAAE,CAAC;AAAA,MACxC;AAAA,IACF;AAAA,EACF;AAEA,SAAO,MAAM,MAAM,OAAO,CAAC,MAAM,QAAQ,IAAI,EAAE,EAAE,CAAC;AACpD;AAMO,SAAS,aAAa,OAAuC;AAClE,QAAM,WAA2B,CAAC;AAGlC,QAAM,SAAiC,CAAC;AACxC,aAAW,QAAQ,MAAM,OAAO;AAC9B,WAAO,KAAK,IAAI,KAAK,OAAO,KAAK,IAAI,KAAK,KAAK;AAC/C,WAAO,KAAK,EAAE,KAAK,OAAO,KAAK,EAAE,KAAK,KAAK;AAAA,EAC7C;AAEA,QAAM,OAAO,OAAO,QAAQ,MAAM,EAC/B,OAAO,CAAC,CAAC,EAAE,CAAC,MAAM,KAAK,CAAC,EACxB,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,EAC1B,MAAM,GAAG,CAAC;AAEb,aAAW,CAAC,QAAQ,CAAC,KAAK,MAAM;AAC9B,UAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,MAAM;AACpD,QAAI,MAAM;AACR,eAAS,KAAK;AAAA,QACZ,aAAa,IAAI,KAAK,KAAK,oBAAoB,CAAC;AAAA,QAChD,OAAO,CAAC,MAAM;AAAA,QACd,OAAO,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,OAAO,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACtF,WAAW;AAAA,QACX,cAAc,KAAK,IAAI,GAAG,IAAI,EAAE;AAAA,MAClC,CAAC;AAAA,IACH;AAAA,EACF;AAGA,QAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,eAAe,EAAE,SAAS,iBAAiB;AACnG,MAAI,WAAW,SAAS,GAAG;AACzB,UAAM,eAAe,WAClB,IAAI,CAAC,MAAM;AACV,YAAM,SAAS,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE;AACpD,aAAO,QAAQ;AAAA,IACjB,CAAC,EACA,OAAO,OAAO;AAEjB,QAAI,aAAa,SAAS,GAAG;AAC3B,eAAS,KAAK;AAAA,QACZ,aAAa,kBAAkB,aAAa,MAAM,GAAG,CAAC,EAAE,KAAK,IAAI,CAAC;AAAA,QAClE,OAAO,WAAW,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACjC,OAAO,WAAW,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,QACjC,WAAW,WAAW;AAAA,QACtB,cAAc,KAAK,IAAI,GAAG,WAAW,SAAS,CAAC;AAAA,MACjD,CAAC;AAAA,IACH;AAAA,EACF;AAGA,QAAM,iBAAiB,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,WAAW;AACvE,MAAI,eAAe,UAAU,GAAG;AAC9B,aAAS,KAAK;AAAA,MACZ,aAAa,GAAG,eAAe,MAAM,2BAA2B,eAAe,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,MAAM,IAAI,EAAE,KAAK,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,MAChI,OAAO,eAAe,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,MACrC,OAAO,CAAC;AAAA,MACR,WAAW,eAAe;AAAA,MAC1B,cAAc,KAAK,IAAI,GAAG,eAAe,SAAS,CAAC;AAAA,IACrD,CAAC;AAAA,EACH;AAGA,QAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,OAAO;AAC/D,QAAM,YAAY,WACf,KAAK,CAAC,GAAG,OAAO,OAAO,EAAE,WAAW,gBAAgB,KAAK,MAAM,OAAO,EAAE,WAAW,gBAAgB,KAAK,EAAE,EAC1G,MAAM,GAAG,CAAC;AAEb,MAAI,UAAU,UAAU,GAAG;AACzB,aAAS,KAAK;AAAA,MACZ,aAAa,iBAAiB,UAAU,IAAI,CAAC,MAAM,GAAG,EAAE,KAAK,KAAK,EAAE,WAAW,gBAAgB,IAAI,EAAE,KAAK,IAAI,CAAC;AAAA,MAC/G,OAAO,UAAU,IAAI,CAAC,MAAM,EAAE,EAAE;AAAA,MAChC,OAAO,CAAC;AAAA,MACR,WAAW,UAAU;AAAA,MACrB,cAAc;AAAA,IAChB,CAAC;AAAA,EACH;AAEA,SAAO,SAAS,KAAK,CAAC,GAAG,MAAM,EAAE,eAAe,EAAE,YAAY;AAChE;AAYO,SAAS,kBACd,OACA,QACU;AACV,QAAM,UAAoB,CAAC;AAG3B,QAAM,cAAc,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,QAAQ;AACjE,aAAW,KAAK,aAAa;AAC3B,YAAQ,KAAK,EAAE,KAAK;AAAA,EACtB;AAGA,MAAI,OAAO,WAAW,KAAK;AACzB,UAAM,eAAe,MAAM,MAAM;AAAA,MAC/B,CAAC,MAAM,EAAE,SAAS,aAAa,EAAE,SAAS;AAAA,IAC5C;AACA,eAAW,KAAK,aAAa,MAAM,GAAG,CAAC,GAAG;AACxC,YAAM,QAAQ,EAAE,WAAW;AAC3B,cAAQ;AAAA,QACN,QACI,YAAY,EAAE,KAAK,aAAa,KAAK,WACrC,YAAY,EAAE,KAAK;AAAA,MACzB;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,cAAc,KAAK;AAC5B,UAAM,aAAa,MAAM,MAAM,OAAO,CAAC,MAAM,EAAE,SAAS,WAAW;AACnE,eAAW,KAAK,WAAW,MAAM,GAAG,CAAC,GAAG;AACtC,cAAQ,KAAK,oBAAoB,EAAE,KAAK,GAAG;AAAA,IAC7C;AAEA,UAAM,aAAa,MAAM,MAAM;AAAA,MAC7B,CAAC,MAAM,EAAE,SAAS,eAAe,EAAE,SAAS;AAAA,IAC9C;AACA,QAAI,WAAW,SAAS,GAAG;AACzB,cAAQ,KAAK,GAAG,WAAW,MAAM,uBAAuB;AAAA,IAC1D;AAAA,EACF;AAGA,MAAI,OAAO,oBAAoB,KAAK;AAClC,UAAM,aAAa,MAAM,MAAM;AAAA,MAC7B,CAAC,MAAM,EAAE,WAAW,eAAe;AAAA,IACrC;AACA,eAAW,KAAK,WAAW,MAAM,GAAG,CAAC,GAAG;AACtC,YAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI;AACpD,YAAM,KAAK,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,EAAE,EAAE;AAChD,UAAI,QAAQ,IAAI;AACd,gBAAQ,KAAK,aAAa,KAAK,KAAK,KAAK,EAAE,IAAI,KAAK,GAAG,KAAK,GAAG;AAAA,MACjE;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,gBAAgB,KAAK;AAC9B,UAAM,WAAW,OAAO;AAAA,MACtB,MAAM,MAAM;AAAA,QACV,CAAC,KAAK,MAAM;AACV,cAAI,EAAE,EAAE,KAAK,IAAI,EAAE,EAAE,KAAK,KAAK;AAC/B,iBAAO;AAAA,QACT;AAAA,QACA,CAAC;AAAA,MACH;AAAA,IACF,EACG,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,EAC1B,MAAM,GAAG,CAAC;AAEb,eAAW,CAAC,QAAQ,KAAK,KAAK,UAAU;AACtC,YAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,MAAM;AACpD,UAAI,QAAQ,SAAS,GAAG;AACtB,gBAAQ,KAAK,GAAG,KAAK,yBAAyB,KAAK,KAAK,GAAG;AAAA,MAC7D;AAAA,IACF;AAAA,EACF;AAGA,MAAI,OAAO,eAAe,KAAK;AAC7B,UAAM,eAAe,MAAM,MAAM;AAAA,MAC/B,CAAC,MACC,EAAE,SAAS,WACX,mCAAmC,KAAK,EAAE,KAAK;AAAA,IACnD;AACA,eAAW,KAAK,aAAa,MAAM,GAAG,CAAC,GAAG;AACxC,cAAQ,KAAK,oBAAoB,EAAE,KAAK,MAAM,EAAE,WAAW,oBAAoB,CAAC,QAAQ;AAAA,IAC1F;AAAA,EACF;AAEA,SAAO;AACT;AAKO,SAAS,aAAa,OAA+B;AAC1D,QAAM,aAAqC,CAAC;AAC5C,aAAW,QAAQ,MAAM,OAAO;AAC9B,eAAW,KAAK,IAAI,KAAK,WAAW,KAAK,IAAI,KAAK,KAAK;AAAA,EACzD;AAEA,QAAM,YAAoC,CAAC;AAC3C,aAAW,QAAQ,MAAM,OAAO;AAC9B,cAAU,KAAK,IAAI,KAAK,UAAU,KAAK,IAAI,KAAK,KAAK;AAAA,EACvD;AAEA,QAAM,QAAkB;AAAA,IACtB,oBAAoB,MAAM,SAAS,SAAS,WAAW,MAAM,SAAS,SAAS;AAAA,IAC/E,YAAY,MAAM,SAAS,QAAQ,KAAK,IAAI,CAAC;AAAA,IAC7C,eAAe,OAAO,QAAQ,UAAU,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,IAClF,eAAe,OAAO,QAAQ,SAAS,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC;AAAA,EACnF;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;AAgCA,eAAsB,WACpB,KACA,OACA,UAC2B;AAE3B,QAAM,eAAe,qBAAqB,KAAK;AAC/C,QAAM,WAAW,aAAa,KAAK;AACnC,QAAM,cAAc,SAAS,SAAS,IAClC;AAAA;AAAA,EAA2B,SAAS,IAAI,CAAC,MAAM,KAAK,EAAE,WAAW,mBAAmB,EAAE,aAAa,QAAQ,CAAC,CAAC,GAAG,EAAE,KAAK,IAAI,CAAC,KAC5H;AAEJ,QAAM,SAAS,MAAM,IAAI;AAAA,IAWvB;AAAA;AAAA;AAAA,IAGA;AAAA,EACF,YAAY;AAAA,EACZ,WAAW;AAAA;AAAA,aAEA,QAAQ;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWnB;AAEA,SAAO;AAAA,IACL;AAAA,IACA,GAAG;AAAA,EACL;AACF;AAMA,SAAS,qBAAqB,OAA+B;AAC3D,QAAM,QAAkB,CAAC;AACzB,QAAM,KAAK,GAAG,MAAM,SAAS,SAAS,WAAW,MAAM,SAAS,SAAS,QAAQ;AAGjF,QAAM,SAAsC,CAAC;AAC7C,aAAW,QAAQ,MAAM,OAAO;AAC9B,QAAI,CAAC,OAAO,KAAK,IAAI,EAAG,QAAO,KAAK,IAAI,IAAI,CAAC;AAC7C,WAAO,KAAK,IAAI,EAAE,KAAK,IAAI;AAAA,EAC7B;AAEA,aAAW,CAAC,MAAM,KAAK,KAAK,OAAO,QAAQ,MAAM,GAAG;AAClD,UAAM,QAAQ,SAAS,SAAS,IAAI;AACpC,UAAM,QAAQ,MAAM,MAAM,GAAG,KAAK;AAClC,UAAM,YAAY,MAAM,SAAS,MAAM;AAEvC,UAAM,KAAK;AAAA,GAAM,KAAK,YAAY,CAAC,MAAM,MAAM,MAAM,SAAS;AAC9D,eAAW,KAAK,OAAO;AACrB,YAAM,QAAQ,OAAO,QAAQ,EAAE,UAAU,EACtC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,KAAK,MAAM,UAAU,EACtD,MAAM,GAAG,CAAC,EACV,IAAI,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,UAAU,CAAC,CAAC,EAAE,EAC3C,KAAK,IAAI;AACZ,YAAM,KAAK,OAAO,EAAE,KAAK,GAAG,QAAQ,KAAK,KAAK,MAAM,EAAE,EAAE;AAAA,IAC1D;AACA,QAAI,YAAY,GAAG;AACjB,YAAM,KAAK,aAAa,SAAS,OAAO;AAAA,IAC1C;AAAA,EACF;AAGA,QAAM,cAAmF,CAAC;AAC1F,aAAW,QAAQ,MAAM,OAAO;AAC9B,QAAI,CAAC,YAAY,KAAK,IAAI,EAAG,aAAY,KAAK,IAAI,IAAI,CAAC;AACvD,UAAM,OAAO,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,KAAK,IAAI;AACvD,UAAM,KAAK,MAAM,MAAM,KAAK,CAAC,MAAM,EAAE,OAAO,KAAK,EAAE;AACnD,QAAI,QAAQ,IAAI;AACd,kBAAY,KAAK,IAAI,EAAE,KAAK,EAAE,MAAM,KAAK,OAAO,IAAI,GAAG,OAAO,QAAQ,KAAK,OAAO,CAAC;AAAA,IACrF;AAAA,EACF;AAEA,QAAM,KAAK,mBAAmB;AAC9B,aAAW,CAAC,MAAM,KAAK,KAAK,OAAO,QAAQ,WAAW,GAAG;AACvD,UAAM,QAAQ,MAAM,MAAM,GAAG,CAAC;AAC9B,UAAM,KAAK,KAAK,IAAI,KAAK,MAAM,MAAM,UAAU;AAC/C,eAAW,KAAK,OAAO;AACrB,YAAM,KAAK,QAAQ,EAAE,IAAI,aAAQ,EAAE,EAAE,cAAc,EAAE,OAAO,QAAQ,CAAC,CAAC,GAAG;AAAA,IAC3E;AACA,QAAI,MAAM,SAAS,MAAM,QAAQ;AAC/B,YAAM,KAAK,eAAe,MAAM,SAAS,MAAM,MAAM,OAAO;AAAA,IAC9D;AAAA,EACF;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;","names":[]}

package/dist/chunk-QR5GEK27.js

@@ -0,0 +1,302 @@
+#!/usr/bin/env node
+
+// src/traits/ocean.ts
+var TRAIT_PROFILES = {
+  openness: {
+    trait: "openness",
+    facets: [
+      "imagination",
+      "artistic interests",
+      "emotionality",
+      "adventurousness",
+      "intellect",
+      "liberalism"
+    ],
+    highBehaviors: [
+      "Tries new products and features early",
+      "Values innovation and uniqueness over reliability",
+      "Open to unconventional solutions",
+      "Seeks variety \u2014 switches brands to explore",
+      "Drawn to aesthetics and design quality",
+      "Reads broadly, considers multiple perspectives before deciding"
+    ],
+    lowBehaviors: [
+      "Prefers familiar, proven products",
+      "Skeptical of novelty and trends",
+      "Sticks with what works \u2014 resists change",
+      "Values practical function over design",
+      "Makes decisions based on past experience, not curiosity"
+    ],
+    correlations: {
+      noveltySeekingScore: 0.41,
+      brandLoyaltyScore: -0.18,
+      priceSensitivityScore: -0.12,
+      socialInfluenceScore: 0.08,
+      complaintLikelihoodScore: 0.15,
+      decisionDeliberationScore: 0.22,
+      advocacyScore: 0.25,
+      riskToleranceScore: 0.38
+    }
+  },
+  conscientiousness: {
+    trait: "conscientiousness",
+    facets: [
+      "self-efficacy",
+      "orderliness",
+      "dutifulness",
+      "achievement-striving",
+      "self-discipline",
+      "cautiousness"
+    ],
+    highBehaviors: [
+      "Researches thoroughly before purchasing",
+      "Reads reviews, compares options methodically",
+      "Loyal to brands that consistently deliver",
+      "Plans purchases \u2014 rarely impulse buys",
+      "Values reliability and quality over novelty",
+      "Keeps subscriptions organized \u2014 cancels unused ones"
+    ],
+    lowBehaviors: [
+      "Impulse buyer \u2014 decides quickly on feel",
+      "Doesn't read the fine print",
+      "Forgets to cancel subscriptions",
+      "Less likely to leave reviews",
+      "Switches tools often without fully evaluating"
+    ],
+    correlations: {
+      noveltySeekingScore: -0.15,
+      brandLoyaltyScore: 0.34,
+      priceSensitivityScore: 0.19,
+      socialInfluenceScore: -0.08,
+      complaintLikelihoodScore: 0.12,
+      decisionDeliberationScore: 0.42,
+      advocacyScore: 0.18,
+      riskToleranceScore: -0.28
+    }
+  },
+  extraversion: {
+    trait: "extraversion",
+    facets: [
+      "friendliness",
+      "gregariousness",
+      "assertiveness",
+      "activity level",
+      "excitement-seeking",
+      "cheerfulness"
+    ],
+    highBehaviors: [
+      "Heavily influenced by social proof and peer behavior",
+      "Shares product experiences publicly \u2014 writes reviews, posts on social",
+      "Prefers products with community or social features",
+      "Talks about purchases with friends \u2014 word of mouth driver",
+      "Makes faster decisions \u2014 trusts gut and social signals",
+      "Values status and visibility in product choices"
+    ],
+    lowBehaviors: [
+      "Decides independently \u2014 not swayed by trends or peers",
+      "Researches alone, quietly",
+      "Rarely shares opinions publicly",
+      "Prefers solo-use products over social/collaborative tools",
+      "Values privacy in product interactions"
+    ],
+    correlations: {
+      noveltySeekingScore: 0.22,
+      brandLoyaltyScore: 0.05,
+      priceSensitivityScore: -0.1,
+      socialInfluenceScore: 0.38,
+      complaintLikelihoodScore: 0.2,
+      decisionDeliberationScore: -0.18,
+      advocacyScore: 0.42,
+      riskToleranceScore: 0.25
+    }
+  },
+  agreeableness: {
+    trait: "agreeableness",
+    facets: [
+      "trust",
+      "morality",
+      "altruism",
+      "cooperation",
+      "modesty",
+      "sympathy"
+    ],
+    highBehaviors: [
+      "Trusts brand promises and recommendations easily",
+      "Avoids confrontation \u2014 unlikely to complain publicly",
+      "Influenced by friend and family recommendations",
+      "Loyal even when slightly dissatisfied \u2014 gives second chances",
+      "Values ethical brands and social responsibility",
+      "Churns silently \u2014 switches without telling you why"
+    ],
+    lowBehaviors: [
+      "Skeptical of marketing claims",
+      "Will complain loudly if unsatisfied",
+      "Negotiates prices, seeks discounts aggressively",
+      "Makes decisions based on personal benefit, not brand relationship",
+      "Less influenced by others' opinions"
+    ],
+    correlations: {
+      noveltySeekingScore: -0.05,
+      brandLoyaltyScore: 0.28,
+      priceSensitivityScore: -0.08,
+      socialInfluenceScore: 0.3,
+      complaintLikelihoodScore: -0.35,
+      decisionDeliberationScore: 0.05,
+      advocacyScore: 0.2,
+      riskToleranceScore: -0.12
+    }
+  },
+  neuroticism: {
+    trait: "neuroticism",
+    facets: [
+      "anxiety",
+      "anger",
+      "depression",
+      "self-consciousness",
+      "immoderation",
+      "vulnerability"
+    ],
+    highBehaviors: [
+      "Anxious about purchase decisions \u2014 needs reassurance",
+      "Risk-averse \u2014 avoids unknown brands or untested products",
+      "Reads negative reviews disproportionately",
+      "Sensitive to price increases \u2014 feels loss acutely",
+      "May impulse-buy to cope with stress, then regret it",
+      "Seeks guarantees, refund policies, social proof before committing"
+    ],
+    lowBehaviors: [
+      "Confident in decisions \u2014 low buyer's remorse",
+      "Comfortable with risk and uncertainty",
+      "Less affected by negative reviews",
+      "Pragmatic about price changes \u2014 evaluates rationally",
+      "Stable preferences \u2014 less emotional in purchasing"
+    ],
+    correlations: {
+      noveltySeekingScore: -0.1,
+      brandLoyaltyScore: -0.05,
+      priceSensitivityScore: 0.28,
+      socialInfluenceScore: 0.18,
+      complaintLikelihoodScore: 0.22,
+      decisionDeliberationScore: -0.15,
+      advocacyScore: -0.12,
+      riskToleranceScore: -0.35
+    }
+  }
+};
+function randomOceanVector(constraints) {
+  const base = {
+    openness: Math.random(),
+    conscientiousness: Math.random(),
+    extraversion: Math.random(),
+    agreeableness: Math.random(),
+    neuroticism: Math.random()
+  };
+  return { ...base, ...constraints };
+}
+function traitDistance(a, b) {
+  const traits = [
+    "openness",
+    "conscientiousness",
+    "extraversion",
+    "agreeableness",
+    "neuroticism"
+  ];
+  let sum = 0;
+  for (const t of traits) {
+    sum += (a[t] - b[t]) ** 2;
+  }
+  return Math.sqrt(sum);
+}
+function averageOceanVector(vectors) {
+  const n = vectors.length;
+  if (n === 0) {
+    return {
+      openness: 0.5,
+      conscientiousness: 0.5,
+      extraversion: 0.5,
+      agreeableness: 0.5,
+      neuroticism: 0.5
+    };
+  }
+  const sum = {
+    openness: 0,
+    conscientiousness: 0,
+    extraversion: 0,
+    agreeableness: 0,
+    neuroticism: 0
+  };
+  for (const v of vectors) {
+    for (const t of Object.keys(sum)) {
+      sum[t] += v[t];
+    }
+  }
+  for (const t of Object.keys(sum)) {
+    sum[t] /= n;
+  }
+  return sum;
+}
+function clampVector(v) {
+  const clamped = { ...v };
+  for (const t of Object.keys(clamped)) {
+    clamped[t] = Math.max(0, Math.min(1, clamped[t]));
+  }
+  return clamped;
+}
+function compileBehavioralProfile(traits) {
+  const traitNames = [
+    "openness",
+    "conscientiousness",
+    "extraversion",
+    "agreeableness",
+    "neuroticism"
+  ];
+  const behaviorKeys = [
+    "noveltySeekingScore",
+    "brandLoyaltyScore",
+    "priceSensitivityScore",
+    "socialInfluenceScore",
+    "complaintLikelihoodScore",
+    "decisionDeliberationScore",
+    "advocacyScore",
+    "riskToleranceScore"
+  ];
+  const scores = {};
+  for (const bKey of behaviorKeys) {
+    let score = 0.5;
+    for (const tName of traitNames) {
+      const correlation = TRAIT_PROFILES[tName].correlations[bKey];
+      score += (traits[tName] - 0.5) * correlation;
+    }
+    scores[bKey] = Math.max(0, Math.min(1, score));
+  }
+  const rules = [];
+  for (const tName of traitNames) {
+    const profile = TRAIT_PROFILES[tName];
+    const value = traits[tName];
+    const behaviors = value >= 0.5 ? profile.highBehaviors : profile.lowBehaviors;
+    const intensity = Math.abs(value - 0.5) * 2;
+    const count = intensity > 0.6 ? 3 : intensity > 0.3 ? 2 : 1;
+    rules.push(...behaviors.slice(0, count));
+  }
+  return {
+    noveltySeekingScore: scores.noveltySeekingScore,
+    brandLoyaltyScore: scores.brandLoyaltyScore,
+    priceSensitivityScore: scores.priceSensitivityScore,
+    socialInfluenceScore: scores.socialInfluenceScore,
+    complaintLikelihoodScore: scores.complaintLikelihoodScore,
+    decisionDeliberationScore: scores.decisionDeliberationScore,
+    advocacyScore: scores.advocacyScore,
+    riskToleranceScore: scores.riskToleranceScore,
+    rules
+  };
+}
+
+export {
+  TRAIT_PROFILES,
+  randomOceanVector,
+  traitDistance,
+  averageOceanVector,
+  clampVector,
+  compileBehavioralProfile
+};
+//# sourceMappingURL=chunk-QR5GEK27.js.map