@ainyc/canonry 4.76.1 → 4.76.2

package/dist/{chunk-WFMEK34V.js → chunk-42TDEKKM.js}

@@ -22,7 +22,7 @@ import {
   trafficConnectVercelRequestSchema,
   trafficConnectWordpressRequestSchema,
   trafficEventKindSchema
-} from "./chunk-JNAKRK77.js";
+} from "./chunk-ETM5EPQH.js";
 
 // src/config.ts
 import fs from "fs";

package/dist/{chunk-2FV5GIUB.js → chunk-BIT3JPMW.js}

@@ -11,6 +11,7 @@ import {
   ContentActions,
   DEFAULT_DISCOVERY_PROMOTE_BUCKETS,
   DEFAULT_DISCOVERY_PROMOTE_COMPETITOR_TYPES,
+  DISCOVERY_DEFAULT_DEDUP_THRESHOLD,
   DISCOVERY_PROMOTE_COMPETITOR_CAP,
   DISCOVERY_PROMOTE_COMPETITOR_MIN_HITS,
   DiscoveryBuckets,
@@ -186,6 +187,7 @@ import {
   schedulableRunKindSchema,
   scheduleDtoSchema,
   scheduleUpsertRequestSchema,
+  seedCollapseWarning,
   serializeRunError,
   settingsDtoSchema,
   siteAuditPagesResponseSchema,
@@ -230,7 +232,7 @@ import {
   wordpressSchemaDeployResultDtoSchema,
   wordpressSchemaStatusResultDtoSchema,
   wordpressStatusDtoSchema
-} from "./chunk-JNAKRK77.js";
+} from "./chunk-ETM5EPQH.js";
 
 // src/intelligence-service.ts
 import { eq as eq33, desc as desc17, asc as asc4, and as and24, ne as ne5, or as or5, inArray as inArray12, gte as gte6, lte as lte3 } from "drizzle-orm";
@@ -1014,6 +1016,7 @@ var discoverySessions = sqliteTable("discovery_sessions", {
   aspirationalCount: integer("aspirational_count"),
   wastedCount: integer("wasted_count"),
   competitorMap: text("competitor_map", { mode: "json" }).$type().notNull().default([]),
+  warning: text("warning"),
   error: text("error"),
   startedAt: text("started_at"),
   finishedAt: text("finished_at"),
@@ -2863,6 +2866,16 @@ var MIGRATION_VERSIONS = [
       `CREATE INDEX IF NOT EXISTS idx_site_audit_pages_run ON site_audit_pages(run_id)`,
       `CREATE INDEX IF NOT EXISTS idx_site_audit_pages_project_score ON site_audit_pages(project_id, overall_score)`
     ]
+  },
+  {
+    // Non-fatal operator warning on a discovery session (e.g. the seed-dedup
+    // degenerate-collapse guard). The session still completes; the warning
+    // flags that its coverage may be misleading.
+    version: 76,
+    name: "discovery-session-warning",
+    statements: [
+      `ALTER TABLE discovery_sessions ADD COLUMN warning TEXT`
+    ]
   }
 ];
 function isDuplicateColumnError(err) {
@@ -16652,7 +16665,7 @@ var routeCatalog = [
     method: "post",
     path: "/api/v1/projects/{name}/discover/run",
     summary: "Start a tracked-basket discovery session",
-    description: 'Kicks off a discovery session for the project. The pipeline: ICP description \u2192 Gemini grounded seed prompt \u2192 embed + cluster (cosine \u2265 0.85 by default) \u2192 pick canonical representatives \u2192 probe each canonical via Gemini grounding \u2192 classify into cited / aspirational / wasted-surface \u2192 aggregate competitor map. Returns immediately with `{ runId, sessionId, status: "running", consolidated }`; the actual work runs in the background. Poll `GET /projects/{name}/discover/sessions/{id}` until `status` is `completed` or `failed`. Concurrent/duplicate requests for the same (project, ICP) are consolidated onto a single in-flight session: the response carries `consolidated: true` and `200 OK` instead of `201`, and the request\'s `dedupThreshold` / `maxProbes` are ignored (the in-flight session keeps its original config).',
+    description: 'Kicks off a discovery session for the project. The pipeline: ICP description \u2192 Gemini grounded seed prompt \u2192 embed + cluster (cosine \u2265 0.95 by default) \u2192 pick canonical representatives \u2192 probe each canonical via Gemini grounding \u2192 classify into cited / aspirational / wasted-surface \u2192 aggregate competitor map. Returns immediately with `{ runId, sessionId, status: "running", consolidated }`; the actual work runs in the background. Poll `GET /projects/{name}/discover/sessions/{id}` until `status` is `completed` or `failed`. Concurrent/duplicate requests for the same (project, ICP) are consolidated onto a single in-flight session: the response carries `consolidated: true` and `200 OK` instead of `201`, and the request\'s `dedupThreshold` / `maxProbes` are ignored (the in-flight session keeps its original config).',
     tags: ["discovery"],
     parameters: [nameParameter],
     requestBody: {
@@ -16663,7 +16676,7 @@ var routeCatalog = [
             type: "object",
             properties: {
               icpDescription: { type: "string", description: "Free-text ICP. Required if the project does not have spec.icpDescription stored." },
-              dedupThreshold: { type: "number", description: "Cosine similarity threshold for clustering. Defaults to 0.85." },
+              dedupThreshold: { type: "number", description: "Cosine similarity threshold for clustering. Defaults to 0.95." },
               maxProbes: { type: "integer", description: "Max canonical queries to probe in this session. Default 100, hard cap 500." },
               locations: {
                 type: "array",
@@ -31655,6 +31668,7 @@ function serializeSession(row) {
     aspirationalCount: row.aspirationalCount ?? null,
     wastedCount: row.wastedCount ?? null,
     competitorMap: parseCompetitorMap(row.competitorMap),
+    warning: row.warning ?? null,
     error: row.error ?? null,
     startedAt: row.startedAt ?? null,
     finishedAt: row.finishedAt ?? null,
@@ -31690,7 +31704,6 @@ function selectEligibleCompetitors(competitorMap, competitorTypes) {
 // ../api-routes/src/discovery/orchestrate.ts
 import crypto26 from "crypto";
 import { eq as eq31 } from "drizzle-orm";
-var DEFAULT_DEDUP_THRESHOLD = 0.85;
 var DEFAULT_MAX_PROBES = 100;
 var ABSOLUTE_MAX_PROBES = 500;
 function classifyProbeBucket(input) {
@@ -31736,7 +31749,7 @@ async function pickCanonicals(candidates, deps, dedupThreshold) {
   return clusters.map(pickClusterRepresentative);
 }
 async function executeDiscovery(opts) {
-  const dedupThreshold = opts.dedupThreshold ?? DEFAULT_DEDUP_THRESHOLD;
+  const dedupThreshold = opts.dedupThreshold ?? DISCOVERY_DEFAULT_DEDUP_THRESHOLD;
   const requestedMax = opts.maxProbes ?? DEFAULT_MAX_PROBES;
   const maxProbes = Math.min(Math.max(1, requestedMax), ABSOLUTE_MAX_PROBES);
   const startedAt = (/* @__PURE__ */ new Date()).toISOString();
@@ -31757,13 +31770,19 @@ async function executeDiscovery(opts) {
     { embed: opts.deps.embed },
     dedupThreshold
   );
+  const warning = seedCollapseWarning({
+    seedCountRaw,
+    canonicalCount: canonicals.length,
+    dedupThreshold
+  });
   const probedCanonicals = canonicals.slice(0, maxProbes);
   const seedCount = probedCanonicals.length;
   opts.db.update(discoverySessions).set({
     status: DiscoverySessionStatuses.probing,
     seedProvider: seedResult.provider,
     seedCountRaw,
-    seedCount
+    seedCount,
+    warning
   }).where(eq31(discoverySessions.id, opts.sessionId)).run();
   const probeRows = [];
   const buckets = { cited: 0, aspirational: 0, "wasted-surface": 0 };

package/dist/{chunk-JNAKRK77.js → chunk-ETM5EPQH.js}

@@ -1870,6 +1870,12 @@ var discoverySessionDtoSchema = z17.object({
   aspirationalCount: z17.number().int().nullable().default(null),
   wastedCount: z17.number().int().nullable().default(null),
   competitorMap: z17.array(discoveryCompetitorMapEntrySchema).default([]),
+  /**
+   * Non-fatal operator warning recorded while the session ran (currently the
+   * seed dedup collapse guard). The session still completes; the warning flags
+   * that its coverage may be misleading.
+   */
+  warning: z17.string().nullable().optional(),
   error: z17.string().nullable().optional(),
   startedAt: z17.string().nullable().optional(),
   finishedAt: z17.string().nullable().optional(),
@@ -1879,6 +1885,16 @@ var discoverySessionDetailDtoSchema = discoverySessionDtoSchema.extend({
   probes: z17.array(discoveryProbeDtoSchema).default([])
 });
 var DISCOVERY_MAX_PROBES_CAP = 500;
+var DISCOVERY_DEFAULT_DEDUP_THRESHOLD = 0.95;
+var DISCOVERY_SEED_COLLAPSE_RATIO = 0.2;
+var DISCOVERY_SEED_COLLAPSE_MIN_RAW = 10;
+function seedCollapseWarning(input) {
+  const { seedCountRaw, canonicalCount, dedupThreshold } = input;
+  if (seedCountRaw < DISCOVERY_SEED_COLLAPSE_MIN_RAW) return null;
+  if (canonicalCount / seedCountRaw >= DISCOVERY_SEED_COLLAPSE_RATIO) return null;
+  const noun = canonicalCount === 1 ? "query" : "queries";
+  return `Seed dedup collapsed ${seedCountRaw} raw candidates into ${canonicalCount} canonical ${noun} at threshold ${dedupThreshold}. Distinct intents were likely merged into one cluster; re-run with a higher --dedup-threshold.`;
+}
 var discoveryRunRequestSchema = z17.object({
   icpDescription: z17.string().min(1).optional(),
   dedupThreshold: z17.number().min(0).max(1).optional(),
@@ -4440,6 +4456,8 @@ export {
   discoverySessionDtoSchema,
   discoverySessionDetailDtoSchema,
   DISCOVERY_MAX_PROBES_CAP,
+  DISCOVERY_DEFAULT_DEDUP_THRESHOLD,
+  seedCollapseWarning,
   discoveryRunRequestSchema,
   discoveryPromoteRequestSchema,
   discoveryPromotePreviewSchema,

package/dist/{chunk-5QINOJ26.js → chunk-ZUQK4TLL.js}

@@ -9,7 +9,7 @@ import {
   loadConfig,
   loadConfigRaw,
   saveConfigPatch
-} from "./chunk-WFMEK34V.js";
+} from "./chunk-42TDEKKM.js";
 import {
   CC_CACHE_DIR,
   DUCKDB_SPEC,
@@ -97,7 +97,7 @@ import {
   siteAuditPages,
   siteAuditSnapshots,
   usageCounters
-} from "./chunk-2FV5GIUB.js";
+} from "./chunk-BIT3JPMW.js";
 import {
   AGENT_MEMORY_VALUE_MAX_BYTES,
   AGENT_PROVIDER_IDS,
@@ -149,7 +149,7 @@ import {
   validationError,
   winnabilityClassLabel,
   withRetry
-} from "./chunk-JNAKRK77.js";
+} from "./chunk-ETM5EPQH.js";
 
 // src/telemetry.ts
 import crypto from "crypto";
@@ -5761,7 +5761,7 @@ function readStoredGroundingSources(rawResponse) {
   return result;
 }
 async function backfillInsightsCommand(project, opts) {
-  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-ITD2CTKH.js");
+  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-QIGX7IVQ.js");
   const config = loadConfig();
   const db = createClient(config.database);
   migrate(db);

package/dist/cli.js

@@ -27,7 +27,7 @@ import {
   setTelemetrySource,
   showFirstRunNotice,
   trackEvent
-} from "./chunk-5QINOJ26.js";
+} from "./chunk-ZUQK4TLL.js";
 import {
   CliError,
   EXIT_SYSTEM_ERROR,
@@ -44,7 +44,7 @@ import {
   saveConfig,
   saveConfigPatch,
   usageError
-} from "./chunk-WFMEK34V.js";
+} from "./chunk-42TDEKKM.js";
 import {
   apiKeys,
   createClient,
@@ -52,7 +52,7 @@ import {
   projects,
   queries,
   renderReportHtml
-} from "./chunk-2FV5GIUB.js";
+} from "./chunk-BIT3JPMW.js";
 import {
   CcReleaseSyncStatuses,
   CheckScopes,
@@ -72,7 +72,7 @@ import {
   providerQuotaPolicySchema,
   resolveProviderInput,
   winnabilityClassSchema
-} from "./chunk-JNAKRK77.js";
+} from "./chunk-ETM5EPQH.js";
 
 // src/cli.ts
 import { pathToFileURL } from "url";
@@ -1797,6 +1797,7 @@ function printSessionDetail(session) {
       console.log(`    - ${c.domain} (${c.hits} hits, ${c.competitorType})`);
     }
   }
+  if (session.warning) console.log(`  Warning:       ${session.warning}`);
   if (session.error) console.log(`  Error:         ${session.error}`);
   if (session.startedAt) console.log(`  Started:       ${session.startedAt}`);
   if (session.finishedAt) console.log(`  Finished:      ${session.finishedAt}`);
@@ -1920,7 +1921,7 @@ Usage: ${usage}`,
 var DISCOVER_CLI_COMMANDS = [
   {
     path: ["discover", "run"],
-    usage: 'canonry discover run <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.85] [--max-probes 100] [--wait] [--format json]',
+    usage: 'canonry discover run <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.95] [--max-probes 100] [--wait] [--format json]',
     options: {
       icp: stringOption(),
       "icp-angle": multiStringOption(),
@@ -1930,7 +1931,7 @@ var DISCOVER_CLI_COMMANDS = [
       wait: { type: "boolean", default: false }
     },
     run: async (input) => {
-      const usage = 'canonry discover run <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.85] [--max-probes 100] [--wait] [--format json]';
+      const usage = 'canonry discover run <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.95] [--max-probes 100] [--wait] [--format json]';
       const project = requireProject(input, "discover.run", usage);
       await discoverRun(project, {
         icp: getString(input.values, "icp"),
@@ -1949,7 +1950,7 @@ var DISCOVER_CLI_COMMANDS = [
   },
   {
     path: ["discover", "seed"],
-    usage: 'canonry discover seed <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.85] [--max-probes 100] [--wait] [--format json]',
+    usage: 'canonry discover seed <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.95] [--max-probes 100] [--wait] [--format json]',
     options: {
       icp: stringOption(),
       "icp-angle": multiStringOption(),
@@ -1959,7 +1960,7 @@ var DISCOVER_CLI_COMMANDS = [
       wait: { type: "boolean", default: false }
     },
     run: async (input) => {
-      const usage = 'canonry discover seed <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.85] [--max-probes 100] [--wait] [--format json]';
+      const usage = 'canonry discover seed <project> [--icp "..."] [--icp-angle "..."] [--locations michigan,florida] [--dedup-threshold 0.95] [--max-probes 100] [--wait] [--format json]';
       const project = requireProject(input, "discover.seed", usage);
       await discoverSeed(project, {
         icp: getString(input.values, "icp"),

package/dist/index.js

@@ -1,11 +1,11 @@
 import {
   createServer
-} from "./chunk-5QINOJ26.js";
+} from "./chunk-ZUQK4TLL.js";
 import {
   loadConfig
-} from "./chunk-WFMEK34V.js";
-import "./chunk-2FV5GIUB.js";
-import "./chunk-JNAKRK77.js";
+} from "./chunk-42TDEKKM.js";
+import "./chunk-BIT3JPMW.js";
+import "./chunk-ETM5EPQH.js";
 export {
   createServer,
   loadConfig

package/dist/{intelligence-service-ITD2CTKH.js → intelligence-service-QIGX7IVQ.js}

@@ -1,7 +1,7 @@
 import {
   IntelligenceService
-} from "./chunk-2FV5GIUB.js";
-import "./chunk-JNAKRK77.js";
+} from "./chunk-BIT3JPMW.js";
+import "./chunk-ETM5EPQH.js";
 export {
   IntelligenceService
 };

package/dist/mcp.js

@@ -3,8 +3,8 @@ import {
   PACKAGE_VERSION,
   canonryMcpTools,
   createApiClient
-} from "./chunk-WFMEK34V.js";
-import "./chunk-JNAKRK77.js";
+} from "./chunk-42TDEKKM.js";
+import "./chunk-ETM5EPQH.js";
 
 // src/mcp/cli.ts
 import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ainyc/canonry",
-  "version": "4.76.1",
+  "version": "4.76.2",
   "type": "module",
   "description": "Agent-first open-source AEO operating platform - track how answer engines cite your domain",
   "license": "FSL-1.1-ALv2",
@@ -62,26 +62,26 @@
     "@types/node-cron": "^3.0.11",
     "tsup": "^8.5.1",
     "tsx": "^4.19.0",
-    "@ainyc/canonry-api-routes": "0.0.0",
     "@ainyc/canonry-api-client": "0.0.0",
-    "@ainyc/canonry-config": "0.0.0",
-    "@ainyc/canonry-integration-bing": "0.0.0",
+    "@ainyc/canonry-api-routes": "0.0.0",
     "@ainyc/canonry-contracts": "0.0.0",
-    "@ainyc/canonry-integration-commoncrawl": "0.0.0",
-    "@ainyc/canonry-integration-cloud-run": "0.0.0",
     "@ainyc/canonry-db": "0.0.0",
+    "@ainyc/canonry-integration-bing": "0.0.0",
+    "@ainyc/canonry-config": "0.0.0",
+    "@ainyc/canonry-integration-cloud-run": "0.0.0",
+    "@ainyc/canonry-integration-commoncrawl": "0.0.0",
     "@ainyc/canonry-integration-google": "0.0.0",
-    "@ainyc/canonry-integration-google-business-profile": "0.0.0",
+    "@ainyc/canonry-integration-google-places": "0.0.0",
     "@ainyc/canonry-integration-traffic": "0.0.0",
-    "@ainyc/canonry-integration-wordpress": "0.0.0",
+    "@ainyc/canonry-integration-google-business-profile": "0.0.0",
+    "@ainyc/canonry-provider-cdp": "0.0.0",
     "@ainyc/canonry-intelligence": "0.0.0",
-    "@ainyc/canonry-integration-google-places": "0.0.0",
-    "@ainyc/canonry-provider-claude": "0.0.0",
+    "@ainyc/canonry-integration-wordpress": "0.0.0",
     "@ainyc/canonry-provider-gemini": "0.0.0",
-    "@ainyc/canonry-provider-local": "0.0.0",
-    "@ainyc/canonry-provider-cdp": "0.0.0",
+    "@ainyc/canonry-provider-claude": "0.0.0",
+    "@ainyc/canonry-provider-openai": "0.0.0",
     "@ainyc/canonry-provider-perplexity": "0.0.0",
-    "@ainyc/canonry-provider-openai": "0.0.0"
+    "@ainyc/canonry-provider-local": "0.0.0"
   },
   "scripts": {
     "build": "tsx scripts/copy-agent-assets.ts && tsup && tsx build-web.ts",