@ainyc/canonry 4.85.0 → 4.86.0

package/dist/{chunk-3K3QRSYE.js → chunk-DLBQU3VG.js}

@@ -46,8 +46,10 @@ import {
   adsSummaryDtoSchema,
   adsSyncResponseSchema,
   agentProvidersResponseDtoSchema,
+  aggregateHarvestedQueries,
   apiKeyDtoSchema,
   apiKeyListDtoSchema,
+  applyHarvestSemanticNovelty,
   auditLogEntrySchema,
   authInvalid,
   authRequired,
@@ -70,6 +72,7 @@ import {
   brandKeyFromText,
   brandLabelFromDomain,
   brandMetricsDtoSchema,
+  buildHarvestAnchorTerms,
   categorizeSource,
   categorizeSourceWithCompetitors,
   categoryLabel,
@@ -104,6 +107,7 @@ import {
   deriveWinnabilityClass,
   determineAnswerMentioned,
   discoveryBucketSchema,
+  discoveryHarvestDtoSchema,
   discoveryPromotePreviewSchema,
   discoveryPromoteRequestSchema,
   discoveryPromoteResultSchema,
@@ -130,6 +134,7 @@ import {
   ga4SocialReferralHistoryEntrySchema,
   ga4StatusDtoSchema,
   ga4SyncResponseDtoSchema,
+  gateHarvestedSearchQueries,
   gbpAccountListResponseSchema,
   gbpDailyMetricListResponseSchema,
   gbpDiscoverRequestSchema,
@@ -248,7 +253,7 @@ import {
   wordpressSchemaDeployResultDtoSchema,
   wordpressSchemaStatusResultDtoSchema,
   wordpressStatusDtoSchema
-} from "./chunk-I2BJC3DT.js";
+} from "./chunk-23HGQV22.js";
 
 // src/intelligence-service.ts
 import { eq as eq37, desc as desc18, asc as asc5, and as and27, ne as ne5, or as or5, inArray as inArray14, gte as gte7, lte as lte4 } from "drizzle-orm";
@@ -13577,6 +13582,7 @@ var SCHEMA_TABLE = {
   DomainClassificationsResponseDto: domainClassificationsResponseDtoSchema,
   RecommendationBriefDto: recommendationBriefDtoSchema,
   RecommendationExplanationDto: recommendationExplanationDtoSchema,
+  DiscoveryHarvestDto: discoveryHarvestDtoSchema,
   DiscoveryPromotePreview: discoveryPromotePreviewSchema,
   DiscoveryPromoteResult: discoveryPromoteResultSchema,
   DiscoverySessionDetailDto: discoverySessionDetailDtoSchema,
@@ -17432,6 +17438,23 @@ var routeCatalog = [
       404: errorResponse("Project or session not found.")
     }
   },
+  {
+    method: "get",
+    path: "/api/v1/projects/{name}/discover/sessions/{id}/harvest",
+    summary: "Harvest issued search queries (grounding fan-out) from a session",
+    description: "Reads the search queries the answer engine actually issued to answer each probe (Gemini's `groundingMetadata.webSearchQueries` fan-out) back out of the session's stored probe payloads, then runs a mandatory quality gate and returns the survivors as candidate seeds, ranked by how many distinct probes issued each one. The gate drops navigational/phone lookups, over-specific outliers, off-subject acronym collisions, exact already-tracked matches, and \u2014 via an embedding cosine pass over the project's tracked queries \u2014 semantic duplicates (paraphrases/synonyms an exact match can't see). `semanticNoveltyApplied` reports whether that embedding pass ran (it falls back to exact-match when embeddings are unavailable). These are a THIRD signal \u2014 *issued retrieval queries* \u2014 distinct from `mention` (answer text) and `cited` (source list); they carry no demand of their own. Read-only and derived: nothing is probed, tracked, or promoted. `minProbeHits` raises the recurrence floor; `anchor=false` disables the subject anchor for new-subject discovery on a well-scoped project. `stats` carries the raw count and a per-reason rejection tally. Issue #713.",
+    tags: ["discovery"],
+    parameters: [
+      nameParameter,
+      { name: "id", in: "path", required: true, description: "Discovery session ID.", schema: stringSchema },
+      { name: "minProbeHits", in: "query", required: false, description: "Minimum number of distinct probes a candidate must appear in to be admitted (recurrence floor). Default 1.", schema: stringSchema },
+      { name: "anchor", in: "query", required: false, description: 'Set to "false" to disable the subject-anchor filter. Default applies it (when the subject corpus is rich enough).', schema: stringSchema }
+    ],
+    responses: {
+      200: jsonResponse("Harvested candidate seeds + gate stats returned.", "DiscoveryHarvestDto"),
+      404: errorResponse("Project or session not found.")
+    }
+  },
   {
     method: "get",
     path: "/api/v1/projects/{name}/discover/sessions/{id}/promote",
@@ -32896,6 +32919,72 @@ async function discoveryRoutes(app, opts) {
       return reply.send(detail);
     }
   );
+  app.get(
+    "/projects/:name/discover/sessions/:id/harvest",
+    async (request, reply) => {
+      const project = resolveProject(app.db, request.params.name);
+      const session = app.db.select().from(discoverySessions).where(eq34(discoverySessions.id, request.params.id)).get();
+      if (!session || session.projectId !== project.id) {
+        throw notFound("Discovery session", request.params.id);
+      }
+      const parsedFloor = parseInt(request.query.minProbeHits ?? "", 10);
+      const minProbeHits = Number.isNaN(parsedFloor) || parsedFloor < 1 ? 1 : parsedFloor;
+      const applyAnchor = request.query.anchor !== "false";
+      const provider = session.seedProvider ?? "gemini";
+      const probeRows = app.db.select().from(discoveryProbes).where(eq34(discoveryProbes.sessionId, session.id)).all();
+      const extract = opts.harvestSearchQueries;
+      const probesWithQueries = probeRows.map((row) => {
+        if (!extract || !row.rawResponse) return { searchQueries: [] };
+        try {
+          const raw = JSON.parse(row.rawResponse);
+          return { searchQueries: extract({ provider, rawResponse: raw }) };
+        } catch {
+          return { searchQueries: [] };
+        }
+      });
+      const trackedQueries = app.db.select({ query: queries.query }).from(queries).where(eq34(queries.projectId, project.id)).all().map((r) => r.query);
+      const anchorTerms = buildHarvestAnchorTerms(
+        [session.icpDescription ?? "", ...trackedQueries],
+        effectiveDomains(project)
+      );
+      const aggregated = aggregateHarvestedQueries(probesWithQueries);
+      let result = gateHarvestedSearchQueries({
+        candidates: aggregated,
+        trackedQueries,
+        anchorTerms,
+        minProbeHits,
+        applyAnchor
+      });
+      let semanticNoveltyApplied = false;
+      if (opts.embedQueries && result.admitted.length > 0 && trackedQueries.length > 0) {
+        try {
+          const candidateTexts = result.admitted.map((c) => c.query);
+          const vectors = await opts.embedQueries([...candidateTexts, ...trackedQueries]);
+          if (vectors.length === candidateTexts.length + trackedQueries.length) {
+            result = applyHarvestSemanticNovelty({
+              result,
+              candidateVectors: vectors.slice(0, candidateTexts.length),
+              trackedVectors: vectors.slice(candidateTexts.length)
+            });
+            semanticNoveltyApplied = true;
+          }
+        } catch {
+        }
+      }
+      const harvest = {
+        sessionId: session.id,
+        projectId: project.id,
+        provider,
+        status: session.status,
+        minProbeHits,
+        anchorApplied: result.anchorApplied,
+        semanticNoveltyApplied,
+        candidates: result.admitted,
+        stats: result.stats
+      };
+      return reply.send(harvest);
+    }
+  );
   app.get(
     "/projects/:name/discover/sessions/:id/promote",
     async (request, reply) => {
@@ -33581,7 +33670,9 @@ async function apiRoutes(app, opts) {
       discoverLatestRelease: opts.discoverLatestRelease
     });
     await api.register(discoveryRoutes, {
-      onDiscoveryRunRequested: opts.onDiscoveryRunRequested
+      onDiscoveryRunRequested: opts.onDiscoveryRunRequested,
+      harvestSearchQueries: opts.harvestSearchQueries,
+      embedQueries: opts.embedQueries
     });
     await api.register(technicalAeoRoutes, {
       onSiteAuditRequested: opts.onSiteAuditRequested

package/dist/{chunk-62YB3ML7.js → chunk-LLJPZKHG.js}

@@ -23,7 +23,7 @@ import {
   trafficConnectVercelRequestSchema,
   trafficConnectWordpressRequestSchema,
   trafficEventKindSchema
-} from "./chunk-I2BJC3DT.js";
+} from "./chunk-23HGQV22.js";
 
 // src/config.ts
 import fs from "fs";
@@ -3454,6 +3454,18 @@ var getApiV1ProjectsByNameDiscoverSessionsById = (options) => {
     ...options
   });
 };
+var getApiV1ProjectsByNameDiscoverSessionsByIdHarvest = (options) => {
+  return (options.client ?? client).get({
+    security: [
+      {
+        scheme: "bearer",
+        type: "http"
+      }
+    ],
+    url: "/api/v1/projects/{name}/discover/sessions/{id}/harvest",
+    ...options
+  });
+};
 var getApiV1ProjectsByNameDiscoverSessionsByIdPromote = (options) => {
   return (options.client ?? client).get({
     security: [
@@ -4750,6 +4762,19 @@ var ApiClient = class {
       })
     );
   }
+  async getDiscoveryHarvest(project, sessionId, opts) {
+    return this.invoke(
+      () => getApiV1ProjectsByNameDiscoverSessionsByIdHarvest({
+        client: this.heyClient,
+        path: { name: project, id: sessionId },
+        query: {
+          minProbeHits: opts?.minProbeHits !== void 0 ? String(opts.minProbeHits) : void 0,
+          // The server treats anchor=false as "disable"; omit otherwise.
+          anchor: opts?.anchor === false ? "false" : void 0
+        }
+      })
+    );
+  }
   async previewDiscoveryPromote(project, sessionId) {
     return this.invoke(
       () => getApiV1ProjectsByNameDiscoverSessionsByIdPromote({
@@ -5494,6 +5519,12 @@ var discoverySessionIdInputSchema = z2.object({
   project: projectNameSchema,
   sessionId: z2.string().min(1).describe("Discovery session ID returned by canonry_discover_run_start.")
 });
+var discoveryHarvestInputSchema = z2.object({
+  project: projectNameSchema,
+  sessionId: z2.string().min(1).describe("Discovery session ID returned by canonry_discover_run_start."),
+  minProbeHits: z2.number().int().positive().optional().describe("Recurrence floor \u2014 a candidate must have appeared in at least this many distinct probes to be admitted. Default 1."),
+  anchor: z2.boolean().optional().describe("Apply the subject-anchor filter that drops off-topic acronym collisions. Default true; pass false for new-subject discovery on a well-scoped project.")
+});
 var discoveryPromoteInputSchema = z2.object({
   project: projectNameSchema,
   sessionId: z2.string().min(1).describe("Discovery session ID returned by canonry_discover_run_start."),
@@ -6749,6 +6780,20 @@ var canonryMcpTools = [
     openApiOperations: ["GET /api/v1/projects/{name}/discover/sessions/{id}"],
     handler: (client2, input) => client2.getDiscoverySession(input.project, input.sessionId)
   }),
+  defineTool({
+    name: "canonry_discover_harvest",
+    title: "Harvest discovery search queries",
+    description: `Read the search queries the answer engine actually issued (Gemini's grounding fan-out) back out of a session's stored probes, gate them for buyer-intent + novelty, and return the survivors as candidate seeds ranked by how many distinct probes issued each one. These are a THIRD signal \u2014 issued retrieval queries \u2014 distinct from mention (answer text) and cited (source list); they carry no demand of their own. Read-only and derived: nothing is probed, tracked, or promoted. Use it to surface "queries the model searched for that you aren't tracking yet"; the operator/agent then decides what to add via canonry_query_add. minProbeHits raises the recurrence floor; anchor=false disables the subject filter. stats carries the raw count and per-reason rejection tally.`,
+    access: "read",
+    tier: "discovery",
+    inputSchema: discoveryHarvestInputSchema,
+    annotations: readAnnotations(),
+    openApiOperations: ["GET /api/v1/projects/{name}/discover/sessions/{id}/harvest"],
+    handler: (client2, input) => client2.getDiscoveryHarvest(input.project, input.sessionId, {
+      minProbeHits: input.minProbeHits,
+      anchor: input.anchor
+    })
+  }),
   defineTool({
     name: "canonry_discover_promote_preview",
     title: "Preview discovery promotion",

package/dist/{chunk-7BMSWI2K.js → chunk-SELXBOAP.js}

@@ -9,7 +9,7 @@ import {
   loadConfig,
   loadConfigRaw,
   saveConfigPatch
-} from "./chunk-62YB3ML7.js";
+} from "./chunk-LLJPZKHG.js";
 import {
   CC_CACHE_DIR,
   DUCKDB_SPEC,
@@ -104,7 +104,7 @@ import {
   siteAuditPages,
   siteAuditSnapshots,
   usageCounters
-} from "./chunk-3K3QRSYE.js";
+} from "./chunk-DLBQU3VG.js";
 import {
   AGENT_MEMORY_VALUE_MAX_BYTES,
   AGENT_PROVIDER_IDS,
@@ -160,7 +160,7 @@ import {
   validationError,
   winnabilityClassLabel,
   withRetry
-} from "./chunk-I2BJC3DT.js";
+} from "./chunk-23HGQV22.js";
 
 // src/telemetry.ts
 import crypto from "crypto";
@@ -6289,7 +6289,7 @@ function readStoredGroundingSources(rawResponse) {
   return result;
 }
 async function backfillInsightsCommand(project, opts) {
-  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-AHHBQKRD.js");
+  const { IntelligenceService: IntelligenceService2 } = await import("./intelligence-service-ZHUJKZRO.js");
   const config = loadConfig();
   const db = createClient(config.database);
   migrate(db);
@@ -10877,6 +10877,21 @@ async function createServer(opts) {
         app.log.error({ runId: input.runId, err }, "Discovery run failed");
       });
     },
+    // Read issued search queries (fan-out) back out of a stored probe payload.
+    // Discovery is Gemini-only today, so the Gemini extractor handles every
+    // probe; the provider arg lets a future multi-provider discovery dispatch.
+    harvestSearchQueries: ({ rawResponse }) => extractSearchQueriesFromRaw(rawResponse),
+    // Embed seam for the harvest's semantic novelty pass — the same Gemini
+    // embedder the discovery seed pipeline uses. Resolved at call time so a
+    // provider key set after boot is picked up; rejects (→ route degrades to
+    // exact-match novelty) when no Gemini key is configured.
+    embedQueries: (queriesToEmbed) => {
+      const cfg = registry.get("gemini")?.config;
+      if (!cfg?.apiKey) {
+        return Promise.reject(new Error("Gemini API key not configured; harvest semantic novelty unavailable"));
+      }
+      return embedQueries(queriesToEmbed, { apiKey: cfg.apiKey, baseUrl: cfg.baseUrl });
+    },
     onSiteAuditRequested: (runId, projectId, auditOpts) => {
       runSiteAudit(runId, projectId, auditOpts);
     },

package/dist/cli.js

@@ -27,7 +27,7 @@ import {
   setTelemetrySource,
   showFirstRunNotice,
   trackEvent
-} from "./chunk-7BMSWI2K.js";
+} from "./chunk-SELXBOAP.js";
 import {
   CliError,
   EXIT_SYSTEM_ERROR,
@@ -44,7 +44,7 @@ import {
   saveConfig,
   saveConfigPatch,
   usageError
-} from "./chunk-62YB3ML7.js";
+} from "./chunk-LLJPZKHG.js";
 import {
   apiKeys,
   createClient,
@@ -52,7 +52,7 @@ import {
   projects,
   queries,
   renderReportHtml
-} from "./chunk-3K3QRSYE.js";
+} from "./chunk-DLBQU3VG.js";
 import {
   BacklinkSources,
   CcReleaseSyncStatuses,
@@ -77,7 +77,7 @@ import {
   providerQuotaPolicySchema,
   resolveProviderInput,
   winnabilityClassSchema
-} from "./chunk-I2BJC3DT.js";
+} from "./chunk-23HGQV22.js";
 
 // src/cli.ts
 import { pathToFileURL } from "url";
@@ -1832,6 +1832,39 @@ async function discoverShow(project, sessionId, opts) {
   }
   printSessionDetail(session);
 }
+async function discoverHarvest(project, sessionId, opts) {
+  const client = getClient4();
+  const harvest = await client.getDiscoveryHarvest(project, sessionId, {
+    minProbeHits: opts.minProbeHits,
+    anchor: opts.anchor
+  });
+  if (opts.format === "json") {
+    console.log(JSON.stringify(harvest, null, 2));
+    return;
+  }
+  if (opts.format === "jsonl") {
+    const context = { project, sessionId, provider: harvest.provider };
+    emitJsonl(harvest.candidates.map((c) => ({ ...context, ...c })));
+    return;
+  }
+  const r = harvest.stats.rejected;
+  console.log(`Harvested issued search queries \u2014 session ${harvest.sessionId} (${harvest.provider})`);
+  console.log(
+    `  ${harvest.stats.admitted} admitted of ${harvest.stats.rawCandidates} raw \xB7 floor=${harvest.minProbeHits} \xB7 anchor=${harvest.anchorApplied ? "on" : "off"} \xB7 novelty=${harvest.semanticNoveltyApplied ? "semantic" : "exact-only"}`
+  );
+  console.log(
+    `  rejected: ${r.belowFloor} below-floor \xB7 ${r.length} length \xB7 ${r.navigational} navigational \xB7 ${r.duplicate} already-tracked \xB7 ${r.semanticDuplicate} synonym \xB7 ${r.offAnchor} off-subject`
+  );
+  if (harvest.candidates.length === 0) {
+    console.log("  (no candidates \u2014 nothing to track from this session)");
+    return;
+  }
+  console.log("");
+  console.log("  HITS  QUERY");
+  for (const candidate of harvest.candidates) {
+    console.log(`  ${String(candidate.probeHits).padStart(4)}  ${candidate.query}`);
+  }
+}
 async function discoverPromotePreview(project, sessionId, opts) {
   const client = getClient4();
   const preview = await client.previewDiscoveryPromote(project, sessionId);
@@ -2156,6 +2189,32 @@ var DISCOVER_CLI_COMMANDS = [
       });
     }
   },
+  {
+    path: ["discover", "harvest"],
+    usage: "canonry discover harvest <project> <session-id> [--min-probe-hits <n>] [--no-anchor] [--format json|jsonl]",
+    options: {
+      "min-probe-hits": stringOption(),
+      "no-anchor": { type: "boolean", default: false }
+    },
+    run: async (input) => {
+      const usage = "canonry discover harvest <project> <session-id> [--min-probe-hits <n>] [--no-anchor] [--format json|jsonl]";
+      const project = requireProject(input, "discover.harvest", usage);
+      const sessionId = requirePositional(input, 1, {
+        command: "discover.harvest",
+        usage,
+        message: "session ID is required"
+      });
+      await discoverHarvest(project, sessionId, {
+        minProbeHits: parseIntegerOption(input, "min-probe-hits", {
+          command: "discover.harvest",
+          usage,
+          message: "--min-probe-hits must be an integer"
+        }),
+        anchor: !getBoolean(input.values, "no-anchor"),
+        format: input.format
+      });
+    }
+  },
   {
     path: ["discover", "promote", "preview"],
     usage: "canonry discover promote preview <project> <session-id> [--format json]",

package/dist/index.js

@@ -1,11 +1,11 @@
 import {
   createServer
-} from "./chunk-7BMSWI2K.js";
+} from "./chunk-SELXBOAP.js";
 import {
   loadConfig
-} from "./chunk-62YB3ML7.js";
-import "./chunk-3K3QRSYE.js";
-import "./chunk-I2BJC3DT.js";
+} from "./chunk-LLJPZKHG.js";
+import "./chunk-DLBQU3VG.js";
+import "./chunk-23HGQV22.js";
 export {
   createServer,
   loadConfig

package/dist/{intelligence-service-AHHBQKRD.js → intelligence-service-ZHUJKZRO.js}

@@ -1,7 +1,7 @@
 import {
   IntelligenceService
-} from "./chunk-3K3QRSYE.js";
-import "./chunk-I2BJC3DT.js";
+} from "./chunk-DLBQU3VG.js";
+import "./chunk-23HGQV22.js";
 export {
   IntelligenceService
 };

package/dist/mcp.js

@@ -3,10 +3,10 @@ import {
   PACKAGE_VERSION,
   canonryMcpTools,
   createApiClient
-} from "./chunk-62YB3ML7.js";
+} from "./chunk-LLJPZKHG.js";
 import {
   isReadOnlyKey
-} from "./chunk-I2BJC3DT.js";
+} from "./chunk-23HGQV22.js";
 
 // src/mcp/cli.ts
 import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ainyc/canonry",
-  "version": "4.85.0",
+  "version": "4.86.0",
   "type": "module",
   "description": "Agent-first open-source AEO operating platform - track how answer engines cite your domain",
   "license": "FSL-1.1-ALv2",
@@ -62,14 +62,14 @@
     "@types/node-cron": "^3.0.11",
     "tsup": "^8.5.1",
     "tsx": "^4.19.0",
-    "@ainyc/canonry-config": "0.0.0",
     "@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-integration-openai-ads": "0.0.0",
     "@ainyc/canonry-integration-cloud-run": "0.0.0",
-    "@ainyc/canonry-contracts": "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",
@@ -78,11 +78,11 @@
     "@ainyc/canonry-intelligence": "0.0.0",
     "@ainyc/canonry-integration-wordpress": "0.0.0",
     "@ainyc/canonry-provider-cdp": "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"
+    "@ainyc/canonry-provider-local": "0.0.0",
+    "@ainyc/canonry-provider-gemini": "0.0.0",
+    "@ainyc/canonry-provider-perplexity": "0.0.0",
+    "@ainyc/canonry-provider-openai": "0.0.0"
   },
   "scripts": {
     "build": "tsx scripts/copy-agent-assets.ts && tsup && tsx build-web.ts",