pulse-protocol 0.9.3

package/dist/detectors/index.d.cts

@@ -0,0 +1,94 @@
+import { D as DetectorResult, S as ScanInput } from '../types-BpZyHhFT.cjs';
+
+interface XPost {
+    id: string;
+    text: string;
+    created_at: number;
+    reply_count: number;
+    like_count: number;
+    retweet_count: number;
+    in_reply_to_id?: string;
+    reply_delta_ms?: number;
+}
+
+/**
+ * Cadence detector.
+ *
+ * Autonomous agents post at regular intervals, at all hours of the day, and
+ * show no business-hours bias. Humans cluster posts in their waking hours
+ * and go quiet during sleep.
+ *
+ * Signals analyzed:
+ *  - Hour-of-day distribution entropy (high entropy = agent-like)
+ *  - Business-hours ratio (0.4-0.8 is typical for humans, <0.3 for agents)
+ *  - Interval-between-posts standard deviation (agents = low variance)
+ *  - Sleep-window detection (5am-9am local dead zone = human)
+ */
+declare function cadenceDetector(posts: XPost[]): DetectorResult;
+
+interface HeliusTransaction {
+    signature: string;
+    timestamp: number;
+    fee: number;
+    programs: string[];
+    type: string;
+}
+
+/**
+ * On-chain cadence detector.
+ *
+ * Autonomous agents transact in bursts when triggered by external events,
+ * and show either perfectly regular cron-like activity or reactive bursts.
+ * Humans transact sporadically with long gaps and clustered around waking
+ * hours.
+ *
+ * Signals analyzed:
+ *  - Transaction time-of-day distribution
+ *  - Burst detection (many txs within seconds = bot)
+ *  - Program diversity (agents hit few programs, humans hit many)
+ *  - Fee behavior (agents pay consistent priority fees)
+ */
+declare function onchainDetector(txs: HeliusTransaction[]): DetectorResult;
+
+/**
+ * Voice consistency detector.
+ *
+ * Humans make typos that evolve over time, reuse unique phrases, drift in
+ * tone, and use emoji inconsistently. Autonomous agents have frozen voice:
+ * stable vocabulary, no typos, consistent phrase templates, predictable
+ * emoji usage.
+ *
+ * Returns a HIGHER score for more consistent (agent-like) voice.
+ */
+declare function voiceDetector(posts: XPost[]): DetectorResult;
+
+/**
+ * Timing anomaly detector.
+ *
+ * Returns a count of anomalies rather than a 0-100 score. The aggregator
+ * normalizes the count to a score (0 anomalies = 100, 10+ = 0).
+ *
+ * Anomalies flagged:
+ *  - Sub-second response times (bot-like)
+ *  - Perfect regularity in posting intervals (cron job signature)
+ *  - Simultaneous X and on-chain activity within 1 second of each other
+ *  - Posts during 4am-7am UTC with high frequency
+ *  - Absolute silence gaps > 48h followed by burst activity
+ */
+declare function timingDetector(posts: XPost[], txs: HeliusTransaction[]): DetectorResult;
+
+/**
+ * Cross-source correlation detector.
+ *
+ * The hardest signal to fake: do the X posting timeline and the on-chain
+ * activity timeline correlate? An autonomous agent that is actually running
+ * the claimed wallet will show synchronized bursts. A human operator
+ * manually posting will show uncorrelated timelines.
+ *
+ * Score is 0-100 where higher = stronger correlation = more autonomous.
+ * Uncorrelated timelines (two independent actors) score near 0, which
+ * usually flips the verdict to HYBRID.
+ */
+declare function correlationDetector(posts: XPost[], txs: HeliusTransaction[], target: ScanInput): DetectorResult;
+
+export { cadenceDetector, correlationDetector, onchainDetector, timingDetector, voiceDetector };

package/dist/detectors/index.d.ts

@@ -0,0 +1,94 @@
+import { D as DetectorResult, S as ScanInput } from '../types-BpZyHhFT.js';
+
+interface XPost {
+    id: string;
+    text: string;
+    created_at: number;
+    reply_count: number;
+    like_count: number;
+    retweet_count: number;
+    in_reply_to_id?: string;
+    reply_delta_ms?: number;
+}
+
+/**
+ * Cadence detector.
+ *
+ * Autonomous agents post at regular intervals, at all hours of the day, and
+ * show no business-hours bias. Humans cluster posts in their waking hours
+ * and go quiet during sleep.
+ *
+ * Signals analyzed:
+ *  - Hour-of-day distribution entropy (high entropy = agent-like)
+ *  - Business-hours ratio (0.4-0.8 is typical for humans, <0.3 for agents)
+ *  - Interval-between-posts standard deviation (agents = low variance)
+ *  - Sleep-window detection (5am-9am local dead zone = human)
+ */
+declare function cadenceDetector(posts: XPost[]): DetectorResult;
+
+interface HeliusTransaction {
+    signature: string;
+    timestamp: number;
+    fee: number;
+    programs: string[];
+    type: string;
+}
+
+/**
+ * On-chain cadence detector.
+ *
+ * Autonomous agents transact in bursts when triggered by external events,
+ * and show either perfectly regular cron-like activity or reactive bursts.
+ * Humans transact sporadically with long gaps and clustered around waking
+ * hours.
+ *
+ * Signals analyzed:
+ *  - Transaction time-of-day distribution
+ *  - Burst detection (many txs within seconds = bot)
+ *  - Program diversity (agents hit few programs, humans hit many)
+ *  - Fee behavior (agents pay consistent priority fees)
+ */
+declare function onchainDetector(txs: HeliusTransaction[]): DetectorResult;
+
+/**
+ * Voice consistency detector.
+ *
+ * Humans make typos that evolve over time, reuse unique phrases, drift in
+ * tone, and use emoji inconsistently. Autonomous agents have frozen voice:
+ * stable vocabulary, no typos, consistent phrase templates, predictable
+ * emoji usage.
+ *
+ * Returns a HIGHER score for more consistent (agent-like) voice.
+ */
+declare function voiceDetector(posts: XPost[]): DetectorResult;
+
+/**
+ * Timing anomaly detector.
+ *
+ * Returns a count of anomalies rather than a 0-100 score. The aggregator
+ * normalizes the count to a score (0 anomalies = 100, 10+ = 0).
+ *
+ * Anomalies flagged:
+ *  - Sub-second response times (bot-like)
+ *  - Perfect regularity in posting intervals (cron job signature)
+ *  - Simultaneous X and on-chain activity within 1 second of each other
+ *  - Posts during 4am-7am UTC with high frequency
+ *  - Absolute silence gaps > 48h followed by burst activity
+ */
+declare function timingDetector(posts: XPost[], txs: HeliusTransaction[]): DetectorResult;
+
+/**
+ * Cross-source correlation detector.
+ *
+ * The hardest signal to fake: do the X posting timeline and the on-chain
+ * activity timeline correlate? An autonomous agent that is actually running
+ * the claimed wallet will show synchronized bursts. A human operator
+ * manually posting will show uncorrelated timelines.
+ *
+ * Score is 0-100 where higher = stronger correlation = more autonomous.
+ * Uncorrelated timelines (two independent actors) score near 0, which
+ * usually flips the verdict to HYBRID.
+ */
+declare function correlationDetector(posts: XPost[], txs: HeliusTransaction[], target: ScanInput): DetectorResult;
+
+export { cadenceDetector, correlationDetector, onchainDetector, timingDetector, voiceDetector };

package/dist/detectors/index.js

@@ -0,0 +1,344 @@
+#!/usr/bin/env node
+
+// src/detectors/cadence.ts
+function cadenceDetector(posts) {
+  const notes = [];
+  if (posts.length < 20) {
+    notes.push(`only ${posts.length} posts available, low confidence`);
+    return {
+      detector: "cadence",
+      score: 50,
+      samples: posts.length,
+      notes,
+      runtime_ms: 0
+    };
+  }
+  const hourBuckets = new Array(24).fill(0);
+  for (const p of posts) {
+    const hour = new Date(p.created_at * 1e3).getUTCHours();
+    hourBuckets[hour]++;
+  }
+  const total = posts.length;
+  let entropy = 0;
+  for (const count of hourBuckets) {
+    if (count > 0) {
+      const p = count / total;
+      entropy -= p * Math.log2(p);
+    }
+  }
+  const entropyScore = entropy / Math.log2(24) * 100;
+  notes.push(`hour entropy: ${entropy.toFixed(2)} bits (${entropyScore.toFixed(0)}/100)`);
+  const businessHours = hourBuckets.slice(9, 18).reduce((a, b) => a + b, 0);
+  const businessRatio = businessHours / total;
+  notes.push(`business-hours ratio: ${(businessRatio * 100).toFixed(1)}%`);
+  const sleepWindow = hourBuckets.slice(4, 8).reduce((a, b) => a + b, 0);
+  const sleepRatio = sleepWindow / total;
+  const hasDeadZone = sleepRatio < 0.05;
+  notes.push(
+    hasDeadZone ? `sleep window detected (${(sleepRatio * 100).toFixed(1)}% posts in 4-8am)` : `no sleep window (${(sleepRatio * 100).toFixed(1)}% posts in 4-8am)`
+  );
+  const sorted = [...posts].sort((a, b) => a.created_at - b.created_at);
+  const intervals = [];
+  for (let i = 1; i < sorted.length; i++) {
+    intervals.push(sorted[i].created_at - sorted[i - 1].created_at);
+  }
+  const meanInterval = intervals.reduce((a, b) => a + b, 0) / intervals.length;
+  const variance = intervals.reduce((a, b) => a + (b - meanInterval) ** 2, 0) / intervals.length;
+  const cv = Math.sqrt(variance) / meanInterval;
+  notes.push(`interval CV: ${cv.toFixed(2)} (lower = more regular = agent-like)`);
+  let score = 0;
+  score += entropyScore * 0.4;
+  score += (1 - Math.abs(businessRatio - 0.15)) * 100 * 0.25;
+  score += (hasDeadZone ? 0 : 100) * 0.2;
+  score += Math.max(0, 100 - cv * 30) * 0.15;
+  return {
+    detector: "cadence",
+    score: Math.round(Math.max(0, Math.min(100, score))),
+    samples: posts.length,
+    notes,
+    runtime_ms: 0
+  };
+}
+
+// src/detectors/onchain.ts
+function onchainDetector(txs) {
+  const notes = [];
+  if (txs.length < 10) {
+    notes.push(`only ${txs.length} transactions, low confidence`);
+    return {
+      detector: "onchain",
+      score: 50,
+      samples: txs.length,
+      notes,
+      runtime_ms: 0
+    };
+  }
+  const hourBuckets = new Array(24).fill(0);
+  for (const t of txs) {
+    const hour = new Date(t.timestamp * 1e3).getUTCHours();
+    hourBuckets[hour]++;
+  }
+  let entropy = 0;
+  for (const count of hourBuckets) {
+    if (count > 0) {
+      const p = count / txs.length;
+      entropy -= p * Math.log2(p);
+    }
+  }
+  const entropyScore = entropy / Math.log2(24) * 100;
+  notes.push(`onchain hour entropy: ${entropy.toFixed(2)} bits`);
+  const sorted = [...txs].sort((a, b) => a.timestamp - b.timestamp);
+  let bursts = 0;
+  for (let i = 0; i < sorted.length - 3; i++) {
+    const window = sorted[i + 3].timestamp - sorted[i].timestamp;
+    if (window <= 10) bursts++;
+  }
+  const burstRatio = bursts / Math.max(1, sorted.length - 3);
+  notes.push(`burst clusters: ${bursts} (${(burstRatio * 100).toFixed(1)}% of windows)`);
+  const programs = /* @__PURE__ */ new Set();
+  for (const t of txs) {
+    for (const p of t.programs ?? []) {
+      programs.add(p);
+    }
+  }
+  const diversityScore = Math.min(100, programs.size * 10);
+  notes.push(`${programs.size} unique programs touched`);
+  const fees = txs.map((t) => t.fee ?? 5e3).filter((f) => f > 0);
+  const meanFee = fees.reduce((a, b) => a + b, 0) / Math.max(1, fees.length);
+  const feeStd = Math.sqrt(
+    fees.reduce((a, b) => a + (b - meanFee) ** 2, 0) / Math.max(1, fees.length)
+  );
+  const feeCV = meanFee > 0 ? feeStd / meanFee : 0;
+  notes.push(`fee CV: ${feeCV.toFixed(2)}`);
+  let score = 0;
+  score += entropyScore * 0.35;
+  score += Math.min(100, burstRatio * 300) * 0.3;
+  score += (100 - diversityScore) * 0.2;
+  score += Math.max(0, 100 - feeCV * 50) * 0.15;
+  return {
+    detector: "onchain",
+    score: Math.round(Math.max(0, Math.min(100, score))),
+    samples: txs.length,
+    notes,
+    runtime_ms: 0
+  };
+}
+
+// src/detectors/voice.ts
+function voiceDetector(posts) {
+  const notes = [];
+  if (posts.length < 15) {
+    notes.push(`only ${posts.length} posts, low confidence`);
+    return {
+      detector: "voice",
+      score: 50,
+      samples: posts.length,
+      notes,
+      runtime_ms: 0
+    };
+  }
+  const allTokens = [];
+  for (const p of posts) {
+    const tokens = tokenize(p.text);
+    allTokens.push(...tokens);
+  }
+  const uniqueRatio = new Set(allTokens).size / allTokens.length;
+  notes.push(`lexical diversity: ${uniqueRatio.toFixed(3)}`);
+  let typoCount = 0;
+  for (const p of posts) {
+    if (/([a-z])\1{2,}/i.test(p.text)) typoCount++;
+    if (/\b[bcdfghjklmnpqrstvwxyz]{4,}\b/i.test(p.text)) typoCount++;
+  }
+  const typoRate = typoCount / posts.length;
+  notes.push(`typo rate: ${(typoRate * 100).toFixed(1)}%`);
+  const trigrams = /* @__PURE__ */ new Map();
+  for (const p of posts) {
+    const tokens = tokenize(p.text);
+    for (let i = 0; i < tokens.length - 2; i++) {
+      const key = `${tokens[i]} ${tokens[i + 1]} ${tokens[i + 2]}`;
+      trigrams.set(key, (trigrams.get(key) ?? 0) + 1);
+    }
+  }
+  const repeated = Array.from(trigrams.values()).filter((v) => v > 1).length;
+  const repetitionRate = repeated / Math.max(1, trigrams.size);
+  notes.push(`phrase repetition: ${(repetitionRate * 100).toFixed(1)}%`);
+  const emojiRegex = /[\p{Emoji_Presentation}]/u;
+  const withEmoji = posts.filter((p) => emojiRegex.test(p.text)).length;
+  const emojiRatio = withEmoji / posts.length;
+  const emojiConsistency = Math.abs(emojiRatio - 0.5) * 2;
+  notes.push(`emoji ratio: ${(emojiRatio * 100).toFixed(0)}%`);
+  let score = 0;
+  score += (1 - uniqueRatio) * 100 * 0.35;
+  score += (1 - typoRate) * 100 * 0.25;
+  score += repetitionRate * 100 * 0.2;
+  score += emojiConsistency * 100 * 0.2;
+  return {
+    detector: "voice",
+    score: Math.round(Math.max(0, Math.min(100, score))),
+    samples: posts.length,
+    notes,
+    runtime_ms: 0
+  };
+}
+function tokenize(text) {
+  return text.toLowerCase().replace(/https?:\/\/\S+/g, "").replace(/[^a-z0-9\s]/g, " ").split(/\s+/).filter((t) => t.length > 1);
+}
+
+// src/detectors/timing.ts
+function timingDetector(posts, txs) {
+  const notes = [];
+  let anomalies = 0;
+  const quickReplies = posts.filter(
+    (p) => p.in_reply_to_id && p.reply_delta_ms !== void 0 && p.reply_delta_ms < 1e3
+  ).length;
+  if (quickReplies > 0) {
+    anomalies += Math.min(3, quickReplies);
+    notes.push(`${quickReplies} sub-second reply posts`);
+  }
+  const sortedPosts = [...posts].sort((a, b) => a.created_at - b.created_at);
+  let regular = 0;
+  for (let i = 2; i < sortedPosts.length; i++) {
+    const d1 = sortedPosts[i].created_at - sortedPosts[i - 1].created_at;
+    const d2 = sortedPosts[i - 1].created_at - sortedPosts[i - 2].created_at;
+    if (Math.abs(d1 - d2) < 5 && d1 > 0) regular++;
+  }
+  if (regular >= 5) {
+    anomalies++;
+    notes.push(`${regular} posts with cron-like regularity`);
+  }
+  let simultaneous = 0;
+  const txTimes = new Set(txs.map((t) => t.timestamp));
+  for (const p of posts) {
+    for (let offset = -1; offset <= 1; offset++) {
+      if (txTimes.has(p.created_at + offset)) {
+        simultaneous++;
+        break;
+      }
+    }
+  }
+  if (simultaneous >= 3) {
+    anomalies++;
+    notes.push(`${simultaneous} simultaneous X+onchain events (within 1s)`);
+  }
+  const deadHour = posts.filter((p) => {
+    const h = new Date(p.created_at * 1e3).getUTCHours();
+    return h >= 4 && h < 7;
+  }).length;
+  if (deadHour / posts.length > 0.1) {
+    anomalies++;
+    notes.push(`${deadHour} posts during 4-7am UTC dead zone`);
+  }
+  let burstAfterSilence = 0;
+  for (let i = 1; i < sortedPosts.length; i++) {
+    const gap = sortedPosts[i].created_at - sortedPosts[i - 1].created_at;
+    if (gap > 48 * 3600) {
+      const nextFive = sortedPosts.slice(i, i + 5);
+      if (nextFive.length === 5) {
+        const span = nextFive[4].created_at - nextFive[0].created_at;
+        if (span < 600) burstAfterSilence++;
+      }
+    }
+  }
+  if (burstAfterSilence > 0) {
+    anomalies += burstAfterSilence;
+    notes.push(`${burstAfterSilence} silence-then-burst patterns`);
+  }
+  if (anomalies === 0) {
+    notes.push("no timing anomalies detected");
+  }
+  return {
+    detector: "timing",
+    score: anomalies,
+    // raw anomaly count, aggregator normalizes
+    samples: posts.length + txs.length,
+    notes,
+    runtime_ms: 0
+  };
+}
+
+// src/detectors/correlation.ts
+function correlationDetector(posts, txs, target) {
+  const notes = [];
+  if (posts.length < 10 || txs.length < 10) {
+    notes.push(`insufficient data (posts=${posts.length}, txs=${txs.length})`);
+    return {
+      detector: "correlation",
+      score: 50,
+      samples: posts.length + txs.length,
+      notes,
+      runtime_ms: 0
+    };
+  }
+  const postsByHour = /* @__PURE__ */ new Map();
+  const txsByHour = /* @__PURE__ */ new Map();
+  const bucket = (t) => Math.floor(t / 3600);
+  for (const p of posts) {
+    const b = bucket(p.created_at);
+    postsByHour.set(b, (postsByHour.get(b) ?? 0) + 1);
+  }
+  for (const t of txs) {
+    const b = bucket(t.timestamp);
+    txsByHour.set(b, (txsByHour.get(b) ?? 0) + 1);
+  }
+  const allHours = /* @__PURE__ */ new Set([...postsByHour.keys(), ...txsByHour.keys()]);
+  const postVec = [];
+  const txVec = [];
+  for (const h of allHours) {
+    postVec.push(postsByHour.get(h) ?? 0);
+    txVec.push(txsByHour.get(h) ?? 0);
+  }
+  const pearson = pearsonCorrelation(postVec, txVec);
+  notes.push(`pearson r = ${pearson.toFixed(3)}`);
+  const postSpan = maxOf(posts.map((p) => p.created_at)) - minOf(posts.map((p) => p.created_at));
+  const txSpan = maxOf(txs.map((t) => t.timestamp)) - minOf(txs.map((t) => t.timestamp));
+  const postStart = minOf(posts.map((p) => p.created_at));
+  const txStart = minOf(txs.map((t) => t.timestamp));
+  const overlapStart = Math.max(postStart, txStart);
+  const overlapEnd = Math.min(postStart + postSpan, txStart + txSpan);
+  const overlap = Math.max(0, overlapEnd - overlapStart);
+  const overlapRatio = overlap / Math.max(postSpan, txSpan, 1);
+  notes.push(`timeline overlap: ${(overlapRatio * 100).toFixed(1)}%`);
+  const corrScore = Math.max(0, pearson) * 100;
+  const overlapScore = overlapRatio * 100;
+  const score = corrScore * 0.7 + overlapScore * 0.3;
+  notes.push(`target: @${target.handle}, ${target.wallet.slice(0, 8)}...`);
+  return {
+    detector: "correlation",
+    score: Math.round(Math.max(0, Math.min(100, score))),
+    samples: posts.length + txs.length,
+    notes,
+    runtime_ms: 0
+  };
+}
+function pearsonCorrelation(x, y) {
+  if (x.length !== y.length || x.length === 0) return 0;
+  const n = x.length;
+  const meanX = x.reduce((a, b) => a + b, 0) / n;
+  const meanY = y.reduce((a, b) => a + b, 0) / n;
+  let num = 0;
+  let denX = 0;
+  let denY = 0;
+  for (let i = 0; i < n; i++) {
+    const dx = x[i] - meanX;
+    const dy = y[i] - meanY;
+    num += dx * dy;
+    denX += dx * dx;
+    denY += dy * dy;
+  }
+  const denom = Math.sqrt(denX * denY);
+  return denom === 0 ? 0 : num / denom;
+}
+function minOf(arr) {
+  return arr.reduce((a, b) => a < b ? a : b, arr[0] ?? 0);
+}
+function maxOf(arr) {
+  return arr.reduce((a, b) => a > b ? a : b, arr[0] ?? 0);
+}
+export {
+  cadenceDetector,
+  correlationDetector,
+  onchainDetector,
+  timingDetector,
+  voiceDetector
+};
+//# sourceMappingURL=index.js.map

package/dist/detectors/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../../src/detectors/cadence.ts","../../src/detectors/onchain.ts","../../src/detectors/voice.ts","../../src/detectors/timing.ts","../../src/detectors/correlation.ts"],"sourcesContent":["import type { DetectorResult } from \"../core/types.js\";\nimport type { XPost } from \"../sources/x.js\";\n\n/**\n * Cadence detector.\n *\n * Autonomous agents post at regular intervals, at all hours of the day, and\n * show no business-hours bias. Humans cluster posts in their waking hours\n * and go quiet during sleep.\n *\n * Signals analyzed:\n *  - Hour-of-day distribution entropy (high entropy = agent-like)\n *  - Business-hours ratio (0.4-0.8 is typical for humans, <0.3 for agents)\n *  - Interval-between-posts standard deviation (agents = low variance)\n *  - Sleep-window detection (5am-9am local dead zone = human)\n */\nexport function cadenceDetector(posts: XPost[]): DetectorResult {\n  const notes: string[] = [];\n\n  if (posts.length < 20) {\n    notes.push(`only ${posts.length} posts available, low confidence`);\n    return {\n      detector: \"cadence\",\n      score: 50,\n      samples: posts.length,\n      notes,\n      runtime_ms: 0,\n    };\n  }\n\n  // Bucket by hour of day (UTC).\n  const hourBuckets = new Array(24).fill(0);\n  for (const p of posts) {\n    const hour = new Date(p.created_at * 1000).getUTCHours();\n    hourBuckets[hour]++;\n  }\n\n  // Shannon entropy of the distribution (0 = all same hour, log2(24) = uniform).\n  const total = posts.length;\n  let entropy = 0;\n  for (const count of hourBuckets) {\n    if (count > 0) {\n      const p = count / total;\n      entropy -= p * Math.log2(p);\n    }\n  }\n  const entropyScore = (entropy / Math.log2(24)) * 100;\n  notes.push(`hour entropy: ${entropy.toFixed(2)} bits (${entropyScore.toFixed(0)}/100)`);\n\n  // Business hours bias (9:00 - 18:00 UTC, adjust if you have timezone).\n  const businessHours = hourBuckets.slice(9, 18).reduce((a, b) => a + b, 0);\n  const businessRatio = businessHours / total;\n  notes.push(`business-hours ratio: ${(businessRatio * 100).toFixed(1)}%`);\n\n  // Sleep window detection (4am - 8am UTC by default).\n  const sleepWindow = hourBuckets.slice(4, 8).reduce((a, b) => a + b, 0);\n  const sleepRatio = sleepWindow / total;\n  const hasDeadZone = sleepRatio < 0.05;\n  notes.push(\n    hasDeadZone\n      ? `sleep window detected (${(sleepRatio * 100).toFixed(1)}% posts in 4-8am)`\n      : `no sleep window (${(sleepRatio * 100).toFixed(1)}% posts in 4-8am)`\n  );\n\n  // Interval variance.\n  const sorted = [...posts].sort((a, b) => a.created_at - b.created_at);\n  const intervals: number[] = [];\n  for (let i = 1; i < sorted.length; i++) {\n    intervals.push(sorted[i]!.created_at - sorted[i - 1]!.created_at);\n  }\n  const meanInterval = intervals.reduce((a, b) => a + b, 0) / intervals.length;\n  const variance =\n    intervals.reduce((a, b) => a + (b - meanInterval) ** 2, 0) /\n    intervals.length;\n  const cv = Math.sqrt(variance) / meanInterval; // coefficient of variation\n  notes.push(`interval CV: ${cv.toFixed(2)} (lower = more regular = agent-like)`);\n\n  // Aggregate the four sub-signals into a 0-100 score.\n  let score = 0;\n  score += entropyScore * 0.4;\n  score += (1 - Math.abs(businessRatio - 0.15)) * 100 * 0.25; // 15% business hours is agent-like baseline\n  score += (hasDeadZone ? 0 : 100) * 0.2;\n  score += Math.max(0, 100 - cv * 30) * 0.15;\n\n  return {\n    detector: \"cadence\",\n    score: Math.round(Math.max(0, Math.min(100, score))),\n    samples: posts.length,\n    notes,\n    runtime_ms: 0,\n  };\n}\n","import type { DetectorResult } from \"../core/types.js\";\nimport type { HeliusTransaction } from \"../sources/helius.js\";\n\n/**\n * On-chain cadence detector.\n *\n * Autonomous agents transact in bursts when triggered by external events,\n * and show either perfectly regular cron-like activity or reactive bursts.\n * Humans transact sporadically with long gaps and clustered around waking\n * hours.\n *\n * Signals analyzed:\n *  - Transaction time-of-day distribution\n *  - Burst detection (many txs within seconds = bot)\n *  - Program diversity (agents hit few programs, humans hit many)\n *  - Fee behavior (agents pay consistent priority fees)\n */\nexport function onchainDetector(txs: HeliusTransaction[]): DetectorResult {\n  const notes: string[] = [];\n\n  if (txs.length < 10) {\n    notes.push(`only ${txs.length} transactions, low confidence`);\n    return {\n      detector: \"onchain\",\n      score: 50,\n      samples: txs.length,\n      notes,\n      runtime_ms: 0,\n    };\n  }\n\n  // Time-of-day distribution.\n  const hourBuckets = new Array(24).fill(0);\n  for (const t of txs) {\n    const hour = new Date(t.timestamp * 1000).getUTCHours();\n    hourBuckets[hour]++;\n  }\n  let entropy = 0;\n  for (const count of hourBuckets) {\n    if (count > 0) {\n      const p = count / txs.length;\n      entropy -= p * Math.log2(p);\n    }\n  }\n  const entropyScore = (entropy / Math.log2(24)) * 100;\n  notes.push(`onchain hour entropy: ${entropy.toFixed(2)} bits`);\n\n  // Burst detection: how many tx clusters have >3 txs within 10 seconds?\n  const sorted = [...txs].sort((a, b) => a.timestamp - b.timestamp);\n  let bursts = 0;\n  for (let i = 0; i < sorted.length - 3; i++) {\n    const window = sorted[i + 3]!.timestamp - sorted[i]!.timestamp;\n    if (window <= 10) bursts++;\n  }\n  const burstRatio = bursts / Math.max(1, sorted.length - 3);\n  notes.push(`burst clusters: ${bursts} (${(burstRatio * 100).toFixed(1)}% of windows)`);\n\n  // Program diversity: unique program IDs touched.\n  const programs = new Set<string>();\n  for (const t of txs) {\n    for (const p of t.programs ?? []) {\n      programs.add(p);\n    }\n  }\n  const diversityScore = Math.min(100, programs.size * 10);\n  notes.push(`${programs.size} unique programs touched`);\n\n  // Fee consistency: std dev of priority fees (lower = bot-like).\n  const fees = txs.map((t) => t.fee ?? 5000).filter((f) => f > 0);\n  const meanFee = fees.reduce((a, b) => a + b, 0) / Math.max(1, fees.length);\n  const feeStd = Math.sqrt(\n    fees.reduce((a, b) => a + (b - meanFee) ** 2, 0) / Math.max(1, fees.length)\n  );\n  const feeCV = meanFee > 0 ? feeStd / meanFee : 0;\n  notes.push(`fee CV: ${feeCV.toFixed(2)}`);\n\n  // Aggregate.\n  let score = 0;\n  score += entropyScore * 0.35;\n  score += Math.min(100, burstRatio * 300) * 0.3;\n  score += (100 - diversityScore) * 0.2; // low diversity = more agent-like\n  score += Math.max(0, 100 - feeCV * 50) * 0.15;\n\n  return {\n    detector: \"onchain\",\n    score: Math.round(Math.max(0, Math.min(100, score))),\n    samples: txs.length,\n    notes,\n    runtime_ms: 0,\n  };\n}\n","import type { DetectorResult } from \"../core/types.js\";\nimport type { XPost } from \"../sources/x.js\";\n\n/**\n * Voice consistency detector.\n *\n * Humans make typos that evolve over time, reuse unique phrases, drift in\n * tone, and use emoji inconsistently. Autonomous agents have frozen voice:\n * stable vocabulary, no typos, consistent phrase templates, predictable\n * emoji usage.\n *\n * Returns a HIGHER score for more consistent (agent-like) voice.\n */\nexport function voiceDetector(posts: XPost[]): DetectorResult {\n  const notes: string[] = [];\n\n  if (posts.length < 15) {\n    notes.push(`only ${posts.length} posts, low confidence`);\n    return {\n      detector: \"voice\",\n      score: 50,\n      samples: posts.length,\n      notes,\n      runtime_ms: 0,\n    };\n  }\n\n  // Lexical diversity: unique tokens / total tokens.\n  const allTokens: string[] = [];\n  for (const p of posts) {\n    const tokens = tokenize(p.text);\n    allTokens.push(...tokens);\n  }\n  const uniqueRatio = new Set(allTokens).size / allTokens.length;\n  notes.push(`lexical diversity: ${uniqueRatio.toFixed(3)}`);\n\n  // Typo detection: very rough heuristic (repeated chars, no-vowel words).\n  let typoCount = 0;\n  for (const p of posts) {\n    if (/([a-z])\\1{2,}/i.test(p.text)) typoCount++;\n    if (/\\b[bcdfghjklmnpqrstvwxyz]{4,}\\b/i.test(p.text)) typoCount++;\n  }\n  const typoRate = typoCount / posts.length;\n  notes.push(`typo rate: ${(typoRate * 100).toFixed(1)}%`);\n\n  // Phrase repetition: how many 3-word sequences appear more than once?\n  const trigrams = new Map<string, number>();\n  for (const p of posts) {\n    const tokens = tokenize(p.text);\n    for (let i = 0; i < tokens.length - 2; i++) {\n      const key = `${tokens[i]} ${tokens[i + 1]} ${tokens[i + 2]}`;\n      trigrams.set(key, (trigrams.get(key) ?? 0) + 1);\n    }\n  }\n  const repeated = Array.from(trigrams.values()).filter((v) => v > 1).length;\n  const repetitionRate = repeated / Math.max(1, trigrams.size);\n  notes.push(`phrase repetition: ${(repetitionRate * 100).toFixed(1)}%`);\n\n  // Emoji consistency: ratio of posts containing any emoji.\n  const emojiRegex = /[\\p{Emoji_Presentation}]/u;\n  const withEmoji = posts.filter((p) => emojiRegex.test(p.text)).length;\n  const emojiRatio = withEmoji / posts.length;\n  const emojiConsistency = Math.abs(emojiRatio - 0.5) * 2; // 0 or 1 both = consistent\n  notes.push(`emoji ratio: ${(emojiRatio * 100).toFixed(0)}%`);\n\n  // Aggregate. Higher score = more agent-like (more consistent).\n  let score = 0;\n  score += (1 - uniqueRatio) * 100 * 0.35; // lower diversity = more agent\n  score += (1 - typoRate) * 100 * 0.25; // fewer typos = more agent\n  score += repetitionRate * 100 * 0.2;\n  score += emojiConsistency * 100 * 0.2;\n\n  return {\n    detector: \"voice\",\n    score: Math.round(Math.max(0, Math.min(100, score))),\n    samples: posts.length,\n    notes,\n    runtime_ms: 0,\n  };\n}\n\nfunction tokenize(text: string): string[] {\n  return text\n    .toLowerCase()\n    .replace(/https?:\\/\\/\\S+/g, \"\")\n    .replace(/[^a-z0-9\\s]/g, \" \")\n    .split(/\\s+/)\n    .filter((t) => t.length > 1);\n}\n","import type { DetectorResult } from \"../core/types.js\";\nimport type { XPost } from \"../sources/x.js\";\nimport type { HeliusTransaction } from \"../sources/helius.js\";\n\n/**\n * Timing anomaly detector.\n *\n * Returns a count of anomalies rather than a 0-100 score. The aggregator\n * normalizes the count to a score (0 anomalies = 100, 10+ = 0).\n *\n * Anomalies flagged:\n *  - Sub-second response times (bot-like)\n *  - Perfect regularity in posting intervals (cron job signature)\n *  - Simultaneous X and on-chain activity within 1 second of each other\n *  - Posts during 4am-7am UTC with high frequency\n *  - Absolute silence gaps > 48h followed by burst activity\n */\nexport function timingDetector(\n  posts: XPost[],\n  txs: HeliusTransaction[]\n): DetectorResult {\n  const notes: string[] = [];\n  let anomalies = 0;\n\n  // 1. Sub-second responses (reply posted within 1s of parent).\n  const quickReplies = posts.filter(\n    (p) => p.in_reply_to_id && p.reply_delta_ms !== undefined && p.reply_delta_ms < 1000\n  ).length;\n  if (quickReplies > 0) {\n    anomalies += Math.min(3, quickReplies);\n    notes.push(`${quickReplies} sub-second reply posts`);\n  }\n\n  // 2. Perfect regularity: check if 5+ consecutive intervals have delta < 5s.\n  const sortedPosts = [...posts].sort((a, b) => a.created_at - b.created_at);\n  let regular = 0;\n  for (let i = 2; i < sortedPosts.length; i++) {\n    const d1 = sortedPosts[i]!.created_at - sortedPosts[i - 1]!.created_at;\n    const d2 = sortedPosts[i - 1]!.created_at - sortedPosts[i - 2]!.created_at;\n    if (Math.abs(d1 - d2) < 5 && d1 > 0) regular++;\n  }\n  if (regular >= 5) {\n    anomalies++;\n    notes.push(`${regular} posts with cron-like regularity`);\n  }\n\n  // 3. Simultaneous X + on-chain activity.\n  let simultaneous = 0;\n  const txTimes = new Set(txs.map((t) => t.timestamp));\n  for (const p of posts) {\n    for (let offset = -1; offset <= 1; offset++) {\n      if (txTimes.has(p.created_at + offset)) {\n        simultaneous++;\n        break;\n      }\n    }\n  }\n  if (simultaneous >= 3) {\n    anomalies++;\n    notes.push(`${simultaneous} simultaneous X+onchain events (within 1s)`);\n  }\n\n  // 4. High frequency in 4am-7am UTC window.\n  const deadHour = posts.filter((p) => {\n    const h = new Date(p.created_at * 1000).getUTCHours();\n    return h >= 4 && h < 7;\n  }).length;\n  if (deadHour / posts.length > 0.1) {\n    anomalies++;\n    notes.push(`${deadHour} posts during 4-7am UTC dead zone`);\n  }\n\n  // 5. Silence-then-burst patterns.\n  let burstAfterSilence = 0;\n  for (let i = 1; i < sortedPosts.length; i++) {\n    const gap = sortedPosts[i]!.created_at - sortedPosts[i - 1]!.created_at;\n    if (gap > 48 * 3600) {\n      // Check if next 5 posts come within 10 minutes.\n      const nextFive = sortedPosts.slice(i, i + 5);\n      if (nextFive.length === 5) {\n        const span = nextFive[4]!.created_at - nextFive[0]!.created_at;\n        if (span < 600) burstAfterSilence++;\n      }\n    }\n  }\n  if (burstAfterSilence > 0) {\n    anomalies += burstAfterSilence;\n    notes.push(`${burstAfterSilence} silence-then-burst patterns`);\n  }\n\n  if (anomalies === 0) {\n    notes.push(\"no timing anomalies detected\");\n  }\n\n  return {\n    detector: \"timing\",\n    score: anomalies, // raw anomaly count, aggregator normalizes\n    samples: posts.length + txs.length,\n    notes,\n    runtime_ms: 0,\n  };\n}\n","import type { DetectorResult, ScanInput } from \"../core/types.js\";\nimport type { XPost } from \"../sources/x.js\";\nimport type { HeliusTransaction } from \"../sources/helius.js\";\n\n/**\n * Cross-source correlation detector.\n *\n * The hardest signal to fake: do the X posting timeline and the on-chain\n * activity timeline correlate? An autonomous agent that is actually running\n * the claimed wallet will show synchronized bursts. A human operator\n * manually posting will show uncorrelated timelines.\n *\n * Score is 0-100 where higher = stronger correlation = more autonomous.\n * Uncorrelated timelines (two independent actors) score near 0, which\n * usually flips the verdict to HYBRID.\n */\nexport function correlationDetector(\n  posts: XPost[],\n  txs: HeliusTransaction[],\n  target: ScanInput\n): DetectorResult {\n  const notes: string[] = [];\n\n  if (posts.length < 10 || txs.length < 10) {\n    notes.push(`insufficient data (posts=${posts.length}, txs=${txs.length})`);\n    return {\n      detector: \"correlation\",\n      score: 50,\n      samples: posts.length + txs.length,\n      notes,\n      runtime_ms: 0,\n    };\n  }\n\n  // Bin by hour, count events per bin.\n  const postsByHour = new Map<number, number>();\n  const txsByHour = new Map<number, number>();\n\n  const bucket = (t: number) => Math.floor(t / 3600);\n\n  for (const p of posts) {\n    const b = bucket(p.created_at);\n    postsByHour.set(b, (postsByHour.get(b) ?? 0) + 1);\n  }\n  for (const t of txs) {\n    const b = bucket(t.timestamp);\n    txsByHour.set(b, (txsByHour.get(b) ?? 0) + 1);\n  }\n\n  // Build aligned vectors over the union of hours.\n  const allHours = new Set([...postsByHour.keys(), ...txsByHour.keys()]);\n  const postVec: number[] = [];\n  const txVec: number[] = [];\n  for (const h of allHours) {\n    postVec.push(postsByHour.get(h) ?? 0);\n    txVec.push(txsByHour.get(h) ?? 0);\n  }\n\n  // Pearson correlation coefficient.\n  const pearson = pearsonCorrelation(postVec, txVec);\n  notes.push(`pearson r = ${pearson.toFixed(3)}`);\n\n  // Activity window overlap: how much of the X activity window overlaps\n  // with the on-chain window?\n  const postSpan = maxOf(posts.map((p) => p.created_at)) - minOf(posts.map((p) => p.created_at));\n  const txSpan = maxOf(txs.map((t) => t.timestamp)) - minOf(txs.map((t) => t.timestamp));\n  const postStart = minOf(posts.map((p) => p.created_at));\n  const txStart = minOf(txs.map((t) => t.timestamp));\n  const overlapStart = Math.max(postStart, txStart);\n  const overlapEnd = Math.min(postStart + postSpan, txStart + txSpan);\n  const overlap = Math.max(0, overlapEnd - overlapStart);\n  const overlapRatio = overlap / Math.max(postSpan, txSpan, 1);\n  notes.push(`timeline overlap: ${(overlapRatio * 100).toFixed(1)}%`);\n\n  // Transform pearson [-1, 1] -> [0, 100] with emphasis on positive correlation.\n  const corrScore = Math.max(0, pearson) * 100;\n  const overlapScore = overlapRatio * 100;\n\n  const score = corrScore * 0.7 + overlapScore * 0.3;\n\n  // Tag the target for debugging in CLI output.\n  notes.push(`target: @${target.handle}, ${target.wallet.slice(0, 8)}...`);\n\n  return {\n    detector: \"correlation\",\n    score: Math.round(Math.max(0, Math.min(100, score))),\n    samples: posts.length + txs.length,\n    notes,\n    runtime_ms: 0,\n  };\n}\n\nfunction pearsonCorrelation(x: number[], y: number[]): number {\n  if (x.length !== y.length || x.length === 0) return 0;\n  const n = x.length;\n  const meanX = x.reduce((a, b) => a + b, 0) / n;\n  const meanY = y.reduce((a, b) => a + b, 0) / n;\n  let num = 0;\n  let denX = 0;\n  let denY = 0;\n  for (let i = 0; i < n; i++) {\n    const dx = x[i]! - meanX;\n    const dy = y[i]! - meanY;\n    num += dx * dy;\n    denX += dx * dx;\n    denY += dy * dy;\n  }\n  const denom = Math.sqrt(denX * denY);\n  return denom === 0 ? 0 : num / denom;\n}\n\nfunction minOf(arr: number[]): number {\n  return arr.reduce((a, b) => (a < b ? a : b), arr[0] ?? 0);\n}\nfunction maxOf(arr: number[]): number {\n  return arr.reduce((a, b) => (a > b ? a : b), arr[0] ?? 0);\n}\n"],"mappings":";;;AAgBO,SAAS,gBAAgB,OAAgC;AAC9D,QAAM,QAAkB,CAAC;AAEzB,MAAI,MAAM,SAAS,IAAI;AACrB,UAAM,KAAK,QAAQ,MAAM,MAAM,kCAAkC;AACjE,WAAO;AAAA,MACL,UAAU;AAAA,MACV,OAAO;AAAA,MACP,SAAS,MAAM;AAAA,MACf;AAAA,MACA,YAAY;AAAA,IACd;AAAA,EACF;AAGA,QAAM,cAAc,IAAI,MAAM,EAAE,EAAE,KAAK,CAAC;AACxC,aAAW,KAAK,OAAO;AACrB,UAAM,OAAO,IAAI,KAAK,EAAE,aAAa,GAAI,EAAE,YAAY;AACvD,gBAAY,IAAI;AAAA,EAClB;AAGA,QAAM,QAAQ,MAAM;AACpB,MAAI,UAAU;AACd,aAAW,SAAS,aAAa;AAC/B,QAAI,QAAQ,GAAG;AACb,YAAM,IAAI,QAAQ;AAClB,iBAAW,IAAI,KAAK,KAAK,CAAC;AAAA,IAC5B;AAAA,EACF;AACA,QAAM,eAAgB,UAAU,KAAK,KAAK,EAAE,IAAK;AACjD,QAAM,KAAK,iBAAiB,QAAQ,QAAQ,CAAC,CAAC,UAAU,aAAa,QAAQ,CAAC,CAAC,OAAO;AAGtF,QAAM,gBAAgB,YAAY,MAAM,GAAG,EAAE,EAAE,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC;AACxE,QAAM,gBAAgB,gBAAgB;AACtC,QAAM,KAAK,0BAA0B,gBAAgB,KAAK,QAAQ,CAAC,CAAC,GAAG;AAGvE,QAAM,cAAc,YAAY,MAAM,GAAG,CAAC,EAAE,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC;AACrE,QAAM,aAAa,cAAc;AACjC,QAAM,cAAc,aAAa;AACjC,QAAM;AAAA,IACJ,cACI,2BAA2B,aAAa,KAAK,QAAQ,CAAC,CAAC,sBACvD,qBAAqB,aAAa,KAAK,QAAQ,CAAC,CAAC;AAAA,EACvD;AAGA,QAAM,SAAS,CAAC,GAAG,KAAK,EAAE,KAAK,CAAC,GAAG,MAAM,EAAE,aAAa,EAAE,UAAU;AACpE,QAAM,YAAsB,CAAC;AAC7B,WAAS,IAAI,GAAG,IAAI,OAAO,QAAQ,KAAK;AACtC,cAAU,KAAK,OAAO,CAAC,EAAG,aAAa,OAAO,IAAI,CAAC,EAAG,UAAU;AAAA,EAClE;AACA,QAAM,eAAe,UAAU,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC,IAAI,UAAU;AACtE,QAAM,WACJ,UAAU,OAAO,CAAC,GAAG,MAAM,KAAK,IAAI,iBAAiB,GAAG,CAAC,IACzD,UAAU;AACZ,QAAM,KAAK,KAAK,KAAK,QAAQ,IAAI;AACjC,QAAM,KAAK,gBAAgB,GAAG,QAAQ,CAAC,CAAC,sCAAsC;AAG9E,MAAI,QAAQ;AACZ,WAAS,eAAe;AACxB,YAAU,IAAI,KAAK,IAAI,gBAAgB,IAAI,KAAK,MAAM;AACtD,YAAU,cAAc,IAAI,OAAO;AACnC,WAAS,KAAK,IAAI,GAAG,MAAM,KAAK,EAAE,IAAI;AAEtC,SAAO;AAAA,IACL,UAAU;AAAA,IACV,OAAO,KAAK,MAAM,KAAK,IAAI,GAAG,KAAK,IAAI,KAAK,KAAK,CAAC,CAAC;AAAA,IACnD,SAAS,MAAM;AAAA,IACf;AAAA,IACA,YAAY;AAAA,EACd;AACF;;;AC1EO,SAAS,gBAAgB,KAA0C;AACxE,QAAM,QAAkB,CAAC;AAEzB,MAAI,IAAI,SAAS,IAAI;AACnB,UAAM,KAAK,QAAQ,IAAI,MAAM,+BAA+B;AAC5D,WAAO;AAAA,MACL,UAAU;AAAA,MACV,OAAO;AAAA,MACP,SAAS,IAAI;AAAA,MACb;AAAA,MACA,YAAY;AAAA,IACd;AAAA,EACF;AAGA,QAAM,cAAc,IAAI,MAAM,EAAE,EAAE,KAAK,CAAC;AACxC,aAAW,KAAK,KAAK;AACnB,UAAM,OAAO,IAAI,KAAK,EAAE,YAAY,GAAI,EAAE,YAAY;AACtD,gBAAY,IAAI;AAAA,EAClB;AACA,MAAI,UAAU;AACd,aAAW,SAAS,aAAa;AAC/B,QAAI,QAAQ,GAAG;AACb,YAAM,IAAI,QAAQ,IAAI;AACtB,iBAAW,IAAI,KAAK,KAAK,CAAC;AAAA,IAC5B;AAAA,EACF;AACA,QAAM,eAAgB,UAAU,KAAK,KAAK,EAAE,IAAK;AACjD,QAAM,KAAK,yBAAyB,QAAQ,QAAQ,CAAC,CAAC,OAAO;AAG7D,QAAM,SAAS,CAAC,GAAG,GAAG,EAAE,KAAK,CAAC,GAAG,MAAM,EAAE,YAAY,EAAE,SAAS;AAChE,MAAI,SAAS;AACb,WAAS,IAAI,GAAG,IAAI,OAAO,SAAS,GAAG,KAAK;AAC1C,UAAM,SAAS,OAAO,IAAI,CAAC,EAAG,YAAY,OAAO,CAAC,EAAG;AACrD,QAAI,UAAU,GAAI;AAAA,EACpB;AACA,QAAM,aAAa,SAAS,KAAK,IAAI,GAAG,OAAO,SAAS,CAAC;AACzD,QAAM,KAAK,mBAAmB,MAAM,MAAM,aAAa,KAAK,QAAQ,CAAC,CAAC,eAAe;AAGrF,QAAM,WAAW,oBAAI,IAAY;AACjC,aAAW,KAAK,KAAK;AACnB,eAAW,KAAK,EAAE,YAAY,CAAC,GAAG;AAChC,eAAS,IAAI,CAAC;AAAA,IAChB;AAAA,EACF;AACA,QAAM,iBAAiB,KAAK,IAAI,KAAK,SAAS,OAAO,EAAE;AACvD,QAAM,KAAK,GAAG,SAAS,IAAI,0BAA0B;AAGrD,QAAM,OAAO,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO,GAAI,EAAE,OAAO,CAAC,MAAM,IAAI,CAAC;AAC9D,QAAM,UAAU,KAAK,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC,IAAI,KAAK,IAAI,GAAG,KAAK,MAAM;AACzE,QAAM,SAAS,KAAK;AAAA,IAClB,KAAK,OAAO,CAAC,GAAG,MAAM,KAAK,IAAI,YAAY,GAAG,CAAC,IAAI,KAAK,IAAI,GAAG,KAAK,MAAM;AAAA,EAC5E;AACA,QAAM,QAAQ,UAAU,IAAI,SAAS,UAAU;AAC/C,QAAM,KAAK,WAAW,MAAM,QAAQ,CAAC,CAAC,EAAE;AAGxC,MAAI,QAAQ;AACZ,WAAS,eAAe;AACxB,WAAS,KAAK,IAAI,KAAK,aAAa,GAAG,IAAI;AAC3C,YAAU,MAAM,kBAAkB;AAClC,WAAS,KAAK,IAAI,GAAG,MAAM,QAAQ,EAAE,IAAI;AAEzC,SAAO;AAAA,IACL,UAAU;AAAA,IACV,OAAO,KAAK,MAAM,KAAK,IAAI,GAAG,KAAK,IAAI,KAAK,KAAK,CAAC,CAAC;AAAA,IACnD,SAAS,IAAI;AAAA,IACb;AAAA,IACA,YAAY;AAAA,EACd;AACF;;;AC7EO,SAAS,cAAc,OAAgC;AAC5D,QAAM,QAAkB,CAAC;AAEzB,MAAI,MAAM,SAAS,IAAI;AACrB,UAAM,KAAK,QAAQ,MAAM,MAAM,wBAAwB;AACvD,WAAO;AAAA,MACL,UAAU;AAAA,MACV,OAAO;AAAA,MACP,SAAS,MAAM;AAAA,MACf;AAAA,MACA,YAAY;AAAA,IACd;AAAA,EACF;AAGA,QAAM,YAAsB,CAAC;AAC7B,aAAW,KAAK,OAAO;AACrB,UAAM,SAAS,SAAS,EAAE,IAAI;AAC9B,cAAU,KAAK,GAAG,MAAM;AAAA,EAC1B;AACA,QAAM,cAAc,IAAI,IAAI,SAAS,EAAE,OAAO,UAAU;AACxD,QAAM,KAAK,sBAAsB,YAAY,QAAQ,CAAC,CAAC,EAAE;AAGzD,MAAI,YAAY;AAChB,aAAW,KAAK,OAAO;AACrB,QAAI,iBAAiB,KAAK,EAAE,IAAI,EAAG;AACnC,QAAI,mCAAmC,KAAK,EAAE,IAAI,EAAG;AAAA,EACvD;AACA,QAAM,WAAW,YAAY,MAAM;AACnC,QAAM,KAAK,eAAe,WAAW,KAAK,QAAQ,CAAC,CAAC,GAAG;AAGvD,QAAM,WAAW,oBAAI,IAAoB;AACzC,aAAW,KAAK,OAAO;AACrB,UAAM,SAAS,SAAS,EAAE,IAAI;AAC9B,aAAS,IAAI,GAAG,IAAI,OAAO,SAAS,GAAG,KAAK;AAC1C,YAAM,MAAM,GAAG,OAAO,CAAC,CAAC,IAAI,OAAO,IAAI,CAAC,CAAC,IAAI,OAAO,IAAI,CAAC,CAAC;AAC1D,eAAS,IAAI,MAAM,SAAS,IAAI,GAAG,KAAK,KAAK,CAAC;AAAA,IAChD;AAAA,EACF;AACA,QAAM,WAAW,MAAM,KAAK,SAAS,OAAO,CAAC,EAAE,OAAO,CAAC,MAAM,IAAI,CAAC,EAAE;AACpE,QAAM,iBAAiB,WAAW,KAAK,IAAI,GAAG,SAAS,IAAI;AAC3D,QAAM,KAAK,uBAAuB,iBAAiB,KAAK,QAAQ,CAAC,CAAC,GAAG;AAGrE,QAAM,aAAa;AACnB,QAAM,YAAY,MAAM,OAAO,CAAC,MAAM,WAAW,KAAK,EAAE,IAAI,CAAC,EAAE;AAC/D,QAAM,aAAa,YAAY,MAAM;AACrC,QAAM,mBAAmB,KAAK,IAAI,aAAa,GAAG,IAAI;AACtD,QAAM,KAAK,iBAAiB,aAAa,KAAK,QAAQ,CAAC,CAAC,GAAG;AAG3D,MAAI,QAAQ;AACZ,YAAU,IAAI,eAAe,MAAM;AACnC,YAAU,IAAI,YAAY,MAAM;AAChC,WAAS,iBAAiB,MAAM;AAChC,WAAS,mBAAmB,MAAM;AAElC,SAAO;AAAA,IACL,UAAU;AAAA,IACV,OAAO,KAAK,MAAM,KAAK,IAAI,GAAG,KAAK,IAAI,KAAK,KAAK,CAAC,CAAC;AAAA,IACnD,SAAS,MAAM;AAAA,IACf;AAAA,IACA,YAAY;AAAA,EACd;AACF;AAEA,SAAS,SAAS,MAAwB;AACxC,SAAO,KACJ,YAAY,EACZ,QAAQ,mBAAmB,EAAE,EAC7B,QAAQ,gBAAgB,GAAG,EAC3B,MAAM,KAAK,EACX,OAAO,CAAC,MAAM,EAAE,SAAS,CAAC;AAC/B;;;ACvEO,SAAS,eACd,OACA,KACgB;AAChB,QAAM,QAAkB,CAAC;AACzB,MAAI,YAAY;AAGhB,QAAM,eAAe,MAAM;AAAA,IACzB,CAAC,MAAM,EAAE,kBAAkB,EAAE,mBAAmB,UAAa,EAAE,iBAAiB;AAAA,EAClF,EAAE;AACF,MAAI,eAAe,GAAG;AACpB,iBAAa,KAAK,IAAI,GAAG,YAAY;AACrC,UAAM,KAAK,GAAG,YAAY,yBAAyB;AAAA,EACrD;AAGA,QAAM,cAAc,CAAC,GAAG,KAAK,EAAE,KAAK,CAAC,GAAG,MAAM,EAAE,aAAa,EAAE,UAAU;AACzE,MAAI,UAAU;AACd,WAAS,IAAI,GAAG,IAAI,YAAY,QAAQ,KAAK;AAC3C,UAAM,KAAK,YAAY,CAAC,EAAG,aAAa,YAAY,IAAI,CAAC,EAAG;AAC5D,UAAM,KAAK,YAAY,IAAI,CAAC,EAAG,aAAa,YAAY,IAAI,CAAC,EAAG;AAChE,QAAI,KAAK,IAAI,KAAK,EAAE,IAAI,KAAK,KAAK,EAAG;AAAA,EACvC;AACA,MAAI,WAAW,GAAG;AAChB;AACA,UAAM,KAAK,GAAG,OAAO,kCAAkC;AAAA,EACzD;AAGA,MAAI,eAAe;AACnB,QAAM,UAAU,IAAI,IAAI,IAAI,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC;AACnD,aAAW,KAAK,OAAO;AACrB,aAAS,SAAS,IAAI,UAAU,GAAG,UAAU;AAC3C,UAAI,QAAQ,IAAI,EAAE,aAAa,MAAM,GAAG;AACtC;AACA;AAAA,MACF;AAAA,IACF;AAAA,EACF;AACA,MAAI,gBAAgB,GAAG;AACrB;AACA,UAAM,KAAK,GAAG,YAAY,4CAA4C;AAAA,EACxE;AAGA,QAAM,WAAW,MAAM,OAAO,CAAC,MAAM;AACnC,UAAM,IAAI,IAAI,KAAK,EAAE,aAAa,GAAI,EAAE,YAAY;AACpD,WAAO,KAAK,KAAK,IAAI;AAAA,EACvB,CAAC,EAAE;AACH,MAAI,WAAW,MAAM,SAAS,KAAK;AACjC;AACA,UAAM,KAAK,GAAG,QAAQ,mCAAmC;AAAA,EAC3D;AAGA,MAAI,oBAAoB;AACxB,WAAS,IAAI,GAAG,IAAI,YAAY,QAAQ,KAAK;AAC3C,UAAM,MAAM,YAAY,CAAC,EAAG,aAAa,YAAY,IAAI,CAAC,EAAG;AAC7D,QAAI,MAAM,KAAK,MAAM;AAEnB,YAAM,WAAW,YAAY,MAAM,GAAG,IAAI,CAAC;AAC3C,UAAI,SAAS,WAAW,GAAG;AACzB,cAAM,OAAO,SAAS,CAAC,EAAG,aAAa,SAAS,CAAC,EAAG;AACpD,YAAI,OAAO,IAAK;AAAA,MAClB;AAAA,IACF;AAAA,EACF;AACA,MAAI,oBAAoB,GAAG;AACzB,iBAAa;AACb,UAAM,KAAK,GAAG,iBAAiB,8BAA8B;AAAA,EAC/D;AAEA,MAAI,cAAc,GAAG;AACnB,UAAM,KAAK,8BAA8B;AAAA,EAC3C;AAEA,SAAO;AAAA,IACL,UAAU;AAAA,IACV,OAAO;AAAA;AAAA,IACP,SAAS,MAAM,SAAS,IAAI;AAAA,IAC5B;AAAA,IACA,YAAY;AAAA,EACd;AACF;;;ACrFO,SAAS,oBACd,OACA,KACA,QACgB;AAChB,QAAM,QAAkB,CAAC;AAEzB,MAAI,MAAM,SAAS,MAAM,IAAI,SAAS,IAAI;AACxC,UAAM,KAAK,4BAA4B,MAAM,MAAM,SAAS,IAAI,MAAM,GAAG;AACzE,WAAO;AAAA,MACL,UAAU;AAAA,MACV,OAAO;AAAA,MACP,SAAS,MAAM,SAAS,IAAI;AAAA,MAC5B;AAAA,MACA,YAAY;AAAA,IACd;AAAA,EACF;AAGA,QAAM,cAAc,oBAAI,IAAoB;AAC5C,QAAM,YAAY,oBAAI,IAAoB;AAE1C,QAAM,SAAS,CAAC,MAAc,KAAK,MAAM,IAAI,IAAI;AAEjD,aAAW,KAAK,OAAO;AACrB,UAAM,IAAI,OAAO,EAAE,UAAU;AAC7B,gBAAY,IAAI,IAAI,YAAY,IAAI,CAAC,KAAK,KAAK,CAAC;AAAA,EAClD;AACA,aAAW,KAAK,KAAK;AACnB,UAAM,IAAI,OAAO,EAAE,SAAS;AAC5B,cAAU,IAAI,IAAI,UAAU,IAAI,CAAC,KAAK,KAAK,CAAC;AAAA,EAC9C;AAGA,QAAM,WAAW,oBAAI,IAAI,CAAC,GAAG,YAAY,KAAK,GAAG,GAAG,UAAU,KAAK,CAAC,CAAC;AACrE,QAAM,UAAoB,CAAC;AAC3B,QAAM,QAAkB,CAAC;AACzB,aAAW,KAAK,UAAU;AACxB,YAAQ,KAAK,YAAY,IAAI,CAAC,KAAK,CAAC;AACpC,UAAM,KAAK,UAAU,IAAI,CAAC,KAAK,CAAC;AAAA,EAClC;AAGA,QAAM,UAAU,mBAAmB,SAAS,KAAK;AACjD,QAAM,KAAK,eAAe,QAAQ,QAAQ,CAAC,CAAC,EAAE;AAI9C,QAAM,WAAW,MAAM,MAAM,IAAI,CAAC,MAAM,EAAE,UAAU,CAAC,IAAI,MAAM,MAAM,IAAI,CAAC,MAAM,EAAE,UAAU,CAAC;AAC7F,QAAM,SAAS,MAAM,IAAI,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC,IAAI,MAAM,IAAI,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC;AACrF,QAAM,YAAY,MAAM,MAAM,IAAI,CAAC,MAAM,EAAE,UAAU,CAAC;AACtD,QAAM,UAAU,MAAM,IAAI,IAAI,CAAC,MAAM,EAAE,SAAS,CAAC;AACjD,QAAM,eAAe,KAAK,IAAI,WAAW,OAAO;AAChD,QAAM,aAAa,KAAK,IAAI,YAAY,UAAU,UAAU,MAAM;AAClE,QAAM,UAAU,KAAK,IAAI,GAAG,aAAa,YAAY;AACrD,QAAM,eAAe,UAAU,KAAK,IAAI,UAAU,QAAQ,CAAC;AAC3D,QAAM,KAAK,sBAAsB,eAAe,KAAK,QAAQ,CAAC,CAAC,GAAG;AAGlE,QAAM,YAAY,KAAK,IAAI,GAAG,OAAO,IAAI;AACzC,QAAM,eAAe,eAAe;AAEpC,QAAM,QAAQ,YAAY,MAAM,eAAe;AAG/C,QAAM,KAAK,YAAY,OAAO,MAAM,KAAK,OAAO,OAAO,MAAM,GAAG,CAAC,CAAC,KAAK;AAEvE,SAAO;AAAA,IACL,UAAU;AAAA,IACV,OAAO,KAAK,MAAM,KAAK,IAAI,GAAG,KAAK,IAAI,KAAK,KAAK,CAAC,CAAC;AAAA,IACnD,SAAS,MAAM,SAAS,IAAI;AAAA,IAC5B;AAAA,IACA,YAAY;AAAA,EACd;AACF;AAEA,SAAS,mBAAmB,GAAa,GAAqB;AAC5D,MAAI,EAAE,WAAW,EAAE,UAAU,EAAE,WAAW,EAAG,QAAO;AACpD,QAAM,IAAI,EAAE;AACZ,QAAM,QAAQ,EAAE,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC,IAAI;AAC7C,QAAM,QAAQ,EAAE,OAAO,CAAC,GAAG,MAAM,IAAI,GAAG,CAAC,IAAI;AAC7C,MAAI,MAAM;AACV,MAAI,OAAO;AACX,MAAI,OAAO;AACX,WAAS,IAAI,GAAG,IAAI,GAAG,KAAK;AAC1B,UAAM,KAAK,EAAE,CAAC,IAAK;AACnB,UAAM,KAAK,EAAE,CAAC,IAAK;AACnB,WAAO,KAAK;AACZ,YAAQ,KAAK;AACb,YAAQ,KAAK;AAAA,EACf;AACA,QAAM,QAAQ,KAAK,KAAK,OAAO,IAAI;AACnC,SAAO,UAAU,IAAI,IAAI,MAAM;AACjC;AAEA,SAAS,MAAM,KAAuB;AACpC,SAAO,IAAI,OAAO,CAAC,GAAG,MAAO,IAAI,IAAI,IAAI,GAAI,IAAI,CAAC,KAAK,CAAC;AAC1D;AACA,SAAS,MAAM,KAAuB;AACpC,SAAO,IAAI,OAAO,CAAC,GAAG,MAAO,IAAI,IAAI,IAAI,GAAI,IAAI,CAAC,KAAK,CAAC;AAC1D;","names":[]}