@ainyc/canonry 4.62.0 → 4.62.1

package/dist/{chunk-QU62IX7K.js → chunk-EECLXA5D.js}

@@ -207,7 +207,7 @@ import {
   wordpressSchemaDeployResultDtoSchema,
   wordpressSchemaStatusResultDtoSchema,
   wordpressStatusDtoSchema
-} from "./chunk-URPUUKLC.js";
+} from "./chunk-GWV4FKVY.js";
 
 // src/intelligence-service.ts
 import { eq as eq30, desc as desc15, asc as asc3, and as and23, ne as ne5, or as or5, inArray as inArray11, gte as gte6, lte as lte3 } from "drizzle-orm";
@@ -25950,6 +25950,9 @@ function verifyIpForRule(ip, ruleId) {
 
 // ../integration-traffic/src/rules.ts
 var LEGACY_CHATGPT_DOMAIN = "chat.openai.com";
+var SELF_TRAFFIC_USER_AGENT_PATTERNS = [
+  /AINYC-AEO-Audit\//i
+];
 var DEFAULT_AI_CRAWLER_RULES = [
   {
     id: "openai-gptbot",
@@ -26248,6 +26251,11 @@ function utmSourceFromUrl(value) {
     return null;
   }
 }
+function isSelfTraffic(event) {
+  const userAgent = event.userAgent?.trim();
+  if (!userAgent) return false;
+  return SELF_TRAFFIC_USER_AGENT_PATTERNS.some((pattern) => pattern.test(userAgent));
+}
 function classifyCrawler(event) {
   const userAgent = event.userAgent?.trim();
   if (!userAgent) return null;
@@ -26450,12 +26458,14 @@ function buildTrafficProbeReport(events, options = {}) {
   const topAiUserFetchPaths = /* @__PURE__ */ new Map();
   const topAiReferrers = /* @__PURE__ */ new Map();
   const topAiReferralLandingPaths = /* @__PURE__ */ new Map();
+  const includedEvents = events.filter((event) => !isSelfTraffic(event));
+  const selfTrafficExcluded = events.length - includedEvents.length;
   let crawlerHits = 0;
   let aiUserFetchHits = 0;
   let aiReferralHits = 0;
   let unknownHits = 0;
   const samples = [];
-  for (const event of events) {
+  for (const event of includedEvents) {
     const tsHour = hourBucket(event.observedAt);
     const pathNormalized = normalizeTrafficPathPattern(event.path);
     const crawler = classifyCrawler(event);
@@ -26586,7 +26596,8 @@ function buildTrafficProbeReport(events, options = {}) {
   return {
     generatedAt: options.generatedAt ?? (/* @__PURE__ */ new Date()).toISOString(),
     totals: {
-      normalizedEvents: events.length,
+      normalizedEvents: includedEvents.length,
+      selfTrafficExcluded,
       crawlerHits,
       aiUserFetchHits,
       aiReferralSessions: aiReferralSessions.size,
@@ -27051,7 +27062,7 @@ async function drainVercelTrafficEvents(options) {
   const events = [];
   const seenEventIds = /* @__PURE__ */ new Set();
   if (endMs <= startMs) {
-    return { events, subWindowCount: 0, effectiveStartMs: startMs, retentionClamped: false };
+    return { events, subWindowCount: 0, effectiveStartMs: startMs, retentionClamped: false, truncatedSliceCount: 0, truncatedSliceStartsMs: [] };
   }
   let cursorMs = startMs;
   let spanMs = endMs - startMs;
@@ -27061,6 +27072,8 @@ async function drainVercelTrafficEvents(options) {
   let retentionResolved = false;
   let floorSpanProbeCountdown = 0;
   let floorPageBudgetCountdown = 0;
+  let truncatedSliceCount = 0;
+  const truncatedSliceStartsMs = [];
   while (cursorMs < endMs) {
     if (subWindowCount >= options.maxSubWindows) {
       throw new Error(
@@ -27103,27 +27116,28 @@ async function drainVercelTrafficEvents(options) {
         }
         continue;
       }
-      if (pageBudget >= FLOOR_SLICE_MAX_PAGES) {
-        throw new Error(
-          `Vercel ${MIN_SUB_WINDOW_MS / 1e3}-second slice holds more than ${FLOOR_SLICE_MAX_PAGES} pages and cannot be drained further`
-        );
+      if (pageBudget < FLOOR_SLICE_MAX_PAGES) {
+        page = await options.pull({
+          token: options.token,
+          projectId: options.projectId,
+          teamId: options.teamId,
+          environment: options.environment,
+          startDate: cursorMs,
+          endDate: subEndMs,
+          maxPages: FLOOR_SLICE_MAX_PAGES
+        });
+        subWindowCount += 1;
+        floorSpanProbeCountdown = FLOOR_CONGESTION_PROBE_INTERVAL;
+        floorPageBudgetCountdown = FLOOR_CONGESTION_PROBE_INTERVAL;
       }
-      page = await options.pull({
-        token: options.token,
-        projectId: options.projectId,
-        teamId: options.teamId,
-        environment: options.environment,
-        startDate: cursorMs,
-        endDate: subEndMs,
-        maxPages: FLOOR_SLICE_MAX_PAGES
-      });
-      subWindowCount += 1;
-      floorSpanProbeCountdown = FLOOR_CONGESTION_PROBE_INTERVAL;
-      floorPageBudgetCountdown = FLOOR_CONGESTION_PROBE_INTERVAL;
       if (page.hasMore) {
-        throw new Error(
-          `Vercel ${MIN_SUB_WINDOW_MS / 1e3}-second slice holds more than ${FLOOR_SLICE_MAX_PAGES} pages and cannot be drained further`
-        );
+        if (options.abortOnTruncation) {
+          throw new Error(
+            `Vercel ${MIN_SUB_WINDOW_MS / 1e3}-second slice starting ${new Date(cursorMs).toISOString()} holds more than ${FLOOR_SLICE_MAX_PAGES} pages and cannot be drained further`
+          );
+        }
+        truncatedSliceCount += 1;
+        truncatedSliceStartsMs.push(cursorMs);
       }
     }
     for (const event of page.events) {
@@ -27147,7 +27161,7 @@ async function drainVercelTrafficEvents(options) {
       }
     }
   }
-  return { events, subWindowCount, effectiveStartMs, retentionClamped };
+  return { events, subWindowCount, effectiveStartMs, retentionClamped, truncatedSliceCount, truncatedSliceStartsMs };
 }
 
 // ../api-routes/src/traffic.ts
@@ -27237,10 +27251,16 @@ async function runBackfillTask(options) {
     return;
   }
   const report = buildTrafficProbeReport(allEvents, { sampleLimit: BACKFILL_SAMPLE_LIMIT });
+  if (report.totals.selfTrafficExcluded > 0) {
+    app.log.info(
+      { sourceId: sourceRow.id, selfTrafficExcluded: report.totals.selfTrafficExcluded },
+      "Backfill dropped Canonry self-traffic before rollup"
+    );
+  }
   const finishedAt = (/* @__PURE__ */ new Date()).toISOString();
   const windowStartIso = windowStart.toISOString();
   const windowEndIso = windowEnd.toISOString();
-  const newSorted = allEvents.slice().sort((a, b) => a.observedAt < b.observedAt ? 1 : a.observedAt > b.observedAt ? -1 : 0).map((e) => e.eventId);
+  const newSorted = allEvents.filter((e) => !isSelfTraffic(e)).sort((a, b) => a.observedAt < b.observedAt ? 1 : a.observedAt > b.observedAt ? -1 : 0).map((e) => e.eventId);
   const newRingBuffer = newSorted.slice(0, MAX_TRACKED_EVENT_IDS);
   const currentLastSyncedMs = sourceRow.lastSyncedAt ? new Date(sourceRow.lastSyncedAt).getTime() : Number.NEGATIVE_INFINITY;
   const nextLastSyncedAt = Math.max(currentLastSyncedMs, windowEnd.getTime()) === windowEnd.getTime() ? windowEndIso : sourceRow.lastSyncedAt;
@@ -27694,6 +27714,7 @@ async function trafficRoutes(app, opts) {
           sourceType: sourceRow.sourceType,
           sourceId: sourceRow.id,
           pulledEvents: 0,
+          selfTrafficExcluded: 0,
           crawlerHits: 0,
           aiUserFetchHits: 0,
           aiReferralHits: 0,
@@ -27850,6 +27871,16 @@ async function trafficRoutes(app, opts) {
         if (drained.retentionClamped) {
           throw vercelRetentionClampError(windowStart.getTime(), drained.effectiveStartMs);
         }
+        if (drained.truncatedSliceCount > 0) {
+          request.log.warn(
+            {
+              sourceId: sourceRow.id,
+              truncatedSlices: drained.truncatedSliceCount,
+              sliceStarts: drained.truncatedSliceStartsMs.map((ms) => new Date(ms).toISOString())
+            },
+            "Vercel drain truncated dense one-second slice(s); ingested a sample and advanced past them"
+          );
+        }
         allEvents = drained.events;
       } catch (e) {
         const msg = e instanceof Error ? e.message : String(e);
@@ -27863,6 +27894,7 @@ async function trafficRoutes(app, opts) {
     let sampleRows = 0;
     let finishedAt = (/* @__PURE__ */ new Date()).toISOString();
     let pulledEventsCount = 0;
+    let selfTrafficExcludedCount = 0;
     let crawlerHitsCount = 0;
     let aiUserFetchHitsCount = 0;
     let aiReferralHitsCount = 0;
@@ -27872,7 +27904,7 @@ async function trafficRoutes(app, opts) {
       const previousIds = latestRow.lastEventIds ?? [];
       const seenEventIds = new Set(previousIds);
       const dedupedEvents = seenEventIds.size === 0 ? allEvents : allEvents.filter((e) => !seenEventIds.has(e.eventId));
-      const newSorted = dedupedEvents.slice().sort((a, b) => a.observedAt < b.observedAt ? 1 : a.observedAt > b.observedAt ? -1 : 0).map((e) => e.eventId);
+      const newSorted = dedupedEvents.filter((e) => !isSelfTraffic(e)).sort((a, b) => a.observedAt < b.observedAt ? 1 : a.observedAt > b.observedAt ? -1 : 0).map((e) => e.eventId);
       const merged = [];
       const mergedSet = /* @__PURE__ */ new Set();
       for (const id of [...newSorted, ...previousIds]) {
@@ -27885,6 +27917,7 @@ async function trafficRoutes(app, opts) {
       const report = buildTrafficProbeReport(dedupedEvents, { sampleLimit });
       finishedAt = (/* @__PURE__ */ new Date()).toISOString();
       pulledEventsCount = report.totals.normalizedEvents;
+      selfTrafficExcludedCount = report.totals.selfTrafficExcluded;
       crawlerHitsCount = report.totals.crawlerHits;
       aiUserFetchHitsCount = report.totals.aiUserFetchHits;
       aiReferralHitsCount = report.totals.aiReferralHits;
@@ -28043,12 +28076,19 @@ async function trafficRoutes(app, opts) {
       entityType: "traffic_source",
       entityId: sourceRow.id
     });
+    if (selfTrafficExcludedCount > 0) {
+      request.log.info(
+        { sourceId: sourceRow.id, selfTrafficExcluded: selfTrafficExcludedCount },
+        "Dropped Canonry self-traffic before rollup; excluded from pulledEvents"
+      );
+    }
     try {
       opts.onTrafficSynced?.({
         status: "completed",
         sourceType: sourceRow.sourceType,
         sourceId: sourceRow.id,
         pulledEvents: pulledEventsCount,
+        selfTrafficExcluded: selfTrafficExcludedCount,
         crawlerHits: crawlerHitsCount,
         aiUserFetchHits: aiUserFetchHitsCount,
         aiReferralHits: aiReferralHitsCount,
@@ -28061,6 +28101,7 @@ async function trafficRoutes(app, opts) {
       runId,
       syncedAt: finishedAt,
       pulledEvents: pulledEventsCount,
+      selfTrafficExcluded: selfTrafficExcludedCount,
       crawlerHits: crawlerHitsCount,
       aiUserFetchHits: aiUserFetchHitsCount,
       aiReferralHits: aiReferralHitsCount,
@@ -28210,7 +28251,8 @@ async function trafficRoutes(app, opts) {
             startDate: chunkStartMs,
             endDate: chunkEndMs,
             pagesPerSubWindow: BACKFILL_MAX_PAGES,
-            maxSubWindows: VERCEL_MAX_SUB_WINDOWS
+            maxSubWindows: VERCEL_MAX_SUB_WINDOWS,
+            abortOnTruncation: true
           });
           if (drained.retentionClamped) {
             throw vercelRetentionClampError(chunkStartMs, drained.effectiveStartMs);

package/dist/{chunk-URPUUKLC.js → chunk-GWV4FKVY.js}

@@ -3454,6 +3454,8 @@ var trafficSyncResponseSchema = z24.object({
   runId: z24.string(),
   syncedAt: z24.string(),
   pulledEvents: z24.number().int().nonnegative(),
+  /** Self-traffic events (Canonry's own tooling) dropped before rollup. */
+  selfTrafficExcluded: z24.number().int().nonnegative(),
   crawlerHits: z24.number().int().nonnegative(),
   aiUserFetchHits: z24.number().int().nonnegative(),
   aiReferralHits: z24.number().int().nonnegative(),

package/dist/{chunk-4FLI5VSQ.js → chunk-UNDLLHDV.js}

@@ -8,7 +8,7 @@ import {
   loadConfig,
   loadConfigRaw,
   saveConfigPatch
-} from "./chunk-PITZUUFV.js";
+} from "./chunk-WW6NOJTP.js";
 import {
   CC_CACHE_DIR,
   DUCKDB_SPEC,
@@ -93,7 +93,7 @@ import {
   runs,
   schedules,
   usageCounters
-} from "./chunk-QU62IX7K.js";
+} from "./chunk-EECLXA5D.js";
 import {
   AGENT_MEMORY_VALUE_MAX_BYTES,
   AGENT_PROVIDER_IDS,
@@ -141,7 +141,7 @@ import {
   skillsClientSchema,
   validationError,
   withRetry
-} from "./chunk-URPUUKLC.js";
+} from "./chunk-GWV4FKVY.js";
 
 // src/telemetry.ts
 import crypto from "crypto";
@@ -5616,7 +5616,7 @@ function readStoredGroundingSources(rawResponse) {
   return result;
 }
 async function backfillInsightsCommand(project, opts) {
-  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-AZDX2EBS.js");
+  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-I5HMGDKS.js");
   const config = loadConfig();
   const db = createClient(config.database);
   migrate(db);
@@ -9987,6 +9987,7 @@ async function createServer(opts) {
         sourceType: event.sourceType,
         sourceId: event.sourceId,
         pulledEvents: event.pulledEvents,
+        selfTrafficExcluded: event.selfTrafficExcluded,
         crawlerHits: event.crawlerHits,
         aiReferralHits: event.aiReferralHits,
         durationMs: event.durationMs

package/dist/{chunk-PITZUUFV.js → chunk-WW6NOJTP.js}

@@ -22,7 +22,7 @@ import {
   trafficConnectVercelRequestSchema,
   trafficConnectWordpressRequestSchema,
   trafficEventKindSchema
-} from "./chunk-URPUUKLC.js";
+} from "./chunk-GWV4FKVY.js";
 
 // src/config.ts
 import fs from "fs";

package/dist/cli.js

@@ -27,7 +27,7 @@ import {
   setTelemetrySource,
   showFirstRunNotice,
   trackEvent
-} from "./chunk-4FLI5VSQ.js";
+} from "./chunk-UNDLLHDV.js";
 import {
   CliError,
   EXIT_SYSTEM_ERROR,
@@ -43,7 +43,7 @@ import {
   saveConfig,
   saveConfigPatch,
   usageError
-} from "./chunk-PITZUUFV.js";
+} from "./chunk-WW6NOJTP.js";
 import {
   apiKeys,
   createClient,
@@ -51,7 +51,7 @@ import {
   projects,
   queries,
   renderReportHtml
-} from "./chunk-QU62IX7K.js";
+} from "./chunk-EECLXA5D.js";
 import {
   CcReleaseSyncStatuses,
   CheckScopes,
@@ -67,7 +67,7 @@ import {
   notificationEventSchema,
   providerQuotaPolicySchema,
   resolveProviderInput
-} from "./chunk-URPUUKLC.js";
+} from "./chunk-GWV4FKVY.js";
 
 // src/cli.ts
 import { pathToFileURL } from "url";
@@ -3700,6 +3700,9 @@ async function trafficSync(project, opts) {
   console.log(`  Run ID:           ${result.runId}`);
   console.log(`  Window:           ${result.windowStart}  \u2192  ${result.windowEnd}`);
   console.log(`  Pulled events:    ${result.pulledEvents}`);
+  if (result.selfTrafficExcluded > 0) {
+    console.log(`  Self-traffic excl: ${result.selfTrafficExcluded}  (Canonry's own tooling, dropped before rollup)`);
+  }
   console.log(`  Crawler hits:     ${result.crawlerHits}  (${result.crawlerBucketRows} hourly bucket${result.crawlerBucketRows === 1 ? "" : "s"})`);
   console.log(`  AI referral hits: ${result.aiReferralHits}  (${result.aiReferralBucketRows} hourly bucket${result.aiReferralBucketRows === 1 ? "" : "s"})`);
   console.log(`  Unknown hits:     ${result.unknownHits}`);

package/dist/index.js

@@ -1,11 +1,11 @@
 import {
   createServer
-} from "./chunk-4FLI5VSQ.js";
+} from "./chunk-UNDLLHDV.js";
 import {
   loadConfig
-} from "./chunk-PITZUUFV.js";
-import "./chunk-QU62IX7K.js";
-import "./chunk-URPUUKLC.js";
+} from "./chunk-WW6NOJTP.js";
+import "./chunk-EECLXA5D.js";
+import "./chunk-GWV4FKVY.js";
 export {
   createServer,
   loadConfig

package/dist/{intelligence-service-AZDX2EBS.js → intelligence-service-I5HMGDKS.js}

@@ -1,7 +1,7 @@
 import {
   IntelligenceService
-} from "./chunk-QU62IX7K.js";
-import "./chunk-URPUUKLC.js";
+} from "./chunk-EECLXA5D.js";
+import "./chunk-GWV4FKVY.js";
 export {
   IntelligenceService
 };

package/dist/mcp.js

@@ -3,8 +3,8 @@ import {
   PACKAGE_VERSION,
   canonryMcpTools,
   createApiClient
-} from "./chunk-PITZUUFV.js";
-import "./chunk-URPUUKLC.js";
+} from "./chunk-WW6NOJTP.js";
+import "./chunk-GWV4FKVY.js";
 
 // src/mcp/cli.ts
 import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ainyc/canonry",
-  "version": "4.62.0",
+  "version": "4.62.1",
   "type": "module",
   "description": "Agent-first open-source AEO operating platform - track how answer engines cite your domain",
   "license": "FSL-1.1-ALv2",
@@ -62,25 +62,25 @@
     "tsup": "^8.5.1",
     "tsx": "^4.19.0",
     "@ainyc/canonry-api-client": "0.0.0",
-    "@ainyc/canonry-api-routes": "0.0.0",
-    "@ainyc/canonry-config": "0.0.0",
+    "@ainyc/canonry-contracts": "0.0.0",
     "@ainyc/canonry-db": "0.0.0",
     "@ainyc/canonry-integration-bing": "0.0.0",
-    "@ainyc/canonry-contracts": "0.0.0",
     "@ainyc/canonry-integration-cloud-run": "0.0.0",
+    "@ainyc/canonry-config": "0.0.0",
+    "@ainyc/canonry-api-routes": "0.0.0",
     "@ainyc/canonry-integration-commoncrawl": "0.0.0",
-    "@ainyc/canonry-integration-google-business-profile": "0.0.0",
+    "@ainyc/canonry-integration-google": "0.0.0",
     "@ainyc/canonry-integration-google-places": "0.0.0",
+    "@ainyc/canonry-integration-google-business-profile": "0.0.0",
     "@ainyc/canonry-integration-traffic": "0.0.0",
     "@ainyc/canonry-integration-wordpress": "0.0.0",
-    "@ainyc/canonry-integration-google": "0.0.0",
-    "@ainyc/canonry-provider-gemini": "0.0.0",
     "@ainyc/canonry-intelligence": "0.0.0",
-    "@ainyc/canonry-provider-claude": "0.0.0",
-    "@ainyc/canonry-provider-local": "0.0.0",
-    "@ainyc/canonry-provider-perplexity": "0.0.0",
     "@ainyc/canonry-provider-cdp": "0.0.0",
-    "@ainyc/canonry-provider-openai": "0.0.0"
+    "@ainyc/canonry-provider-gemini": "0.0.0",
+    "@ainyc/canonry-provider-local": "0.0.0",
+    "@ainyc/canonry-provider-claude": "0.0.0",
+    "@ainyc/canonry-provider-openai": "0.0.0",
+    "@ainyc/canonry-provider-perplexity": "0.0.0"
   },
   "scripts": {
     "build": "tsx scripts/copy-agent-assets.ts && tsup && tsx build-web.ts",