@ainyc/canonry 4.33.0 → 4.33.1

package/dist/{chunk-JJHBPITI.js → chunk-DZENHID5.js}

@@ -10388,7 +10388,7 @@ 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" }`; the actual work runs in the background. Poll `GET /projects/{name}/discover/sessions/{id}` until `status` is `completed` or `failed`.',
+    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).',
     tags: ["discovery"],
     parameters: [nameParameter],
     requestBody: {
@@ -10412,7 +10412,8 @@ var routeCatalog = [
       }
     },
     responses: {
-      201: { description: "Discovery session enqueued; returns { runId, sessionId, status }." },
+      200: { description: "An in-flight session with the same project + ICP was reused; returns { runId, sessionId, status, consolidated: true }. The request's dedupThreshold / maxProbes are ignored." },
+      201: { description: "New discovery session enqueued; returns { runId, sessionId, status, consolidated: false }." },
       400: { description: "Missing or invalid ICP / parameters." },
       404: { description: "Project not found." }
     }
@@ -20009,7 +20010,8 @@ async function doctorRoutes(app, opts) {
 
 // ../api-routes/src/discovery/routes.ts
 import crypto21 from "crypto";
-import { eq as eq25, desc as desc13 } from "drizzle-orm";
+import { and as and16, desc as desc13, eq as eq25, gte as gte4, inArray as inArray8 } from "drizzle-orm";
+var MAX_INFLIGHT_DISCOVERY_AGE_MS = 2 * 60 * 60 * 1e3;
 async function discoveryRoutes(app, opts) {
   app.post("/projects/:name/discover/run", async (request, reply) => {
     const project = resolveProject(app.db, request.params.name);
@@ -20038,9 +20040,23 @@ async function discoveryRoutes(app, opts) {
       });
     }
     const now = (/* @__PURE__ */ new Date()).toISOString();
-    const sessionId = crypto21.randomUUID();
-    const runId = crypto21.randomUUID();
-    app.db.transaction((tx) => {
+    const ageFloorIso = new Date(Date.now() - MAX_INFLIGHT_DISCOVERY_AGE_MS).toISOString();
+    const decision = app.db.transaction((tx) => {
+      const existing = tx.select({ id: discoverySessions.id, runId: discoverySessions.runId }).from(discoverySessions).where(and16(
+        eq25(discoverySessions.projectId, project.id),
+        eq25(discoverySessions.icpDescription, icpDescription),
+        inArray8(discoverySessions.status, [
+          DiscoverySessionStatuses.queued,
+          DiscoverySessionStatuses.seeding,
+          DiscoverySessionStatuses.probing
+        ]),
+        gte4(discoverySessions.createdAt, ageFloorIso)
+      )).orderBy(desc13(discoverySessions.createdAt)).get();
+      if (existing && existing.runId) {
+        return { reused: true, sessionId: existing.id, runId: existing.runId };
+      }
+      const sessionId = crypto21.randomUUID();
+      const runId = crypto21.randomUUID();
       tx.insert(discoverySessions).values({
         id: sessionId,
         projectId: project.id,
@@ -20066,17 +20082,31 @@ async function discoveryRoutes(app, opts) {
         entityType: "discovery_session",
         entityId: sessionId
       });
+      return { reused: false, sessionId, runId };
     });
+    if (decision.reused) {
+      return reply.status(200).send({
+        runId: decision.runId,
+        sessionId: decision.sessionId,
+        status: "running",
+        consolidated: true
+      });
+    }
     opts.onDiscoveryRunRequested({
-      runId,
-      sessionId,
+      runId: decision.runId,
+      sessionId: decision.sessionId,
       projectId: project.id,
       icpDescription,
       dedupThreshold: parsed.data.dedupThreshold,
       maxProbes: parsed.data.maxProbes,
       locations
     });
-    return reply.status(201).send({ runId, sessionId, status: "running" });
+    return reply.status(201).send({
+      runId: decision.runId,
+      sessionId: decision.sessionId,
+      status: "running",
+      consolidated: false
+    });
   });
   app.get(
     "/projects/:name/discover/sessions",
@@ -23328,7 +23358,7 @@ import crypto24 from "crypto";
 import fs7 from "fs";
 import path9 from "path";
 import os5 from "os";
-import { and as and16, eq as eq27, inArray as inArray8, sql as sql10 } from "drizzle-orm";
+import { and as and17, eq as eq27, inArray as inArray9, sql as sql10 } from "drizzle-orm";
 
 // src/run-telemetry.ts
 import crypto23 from "crypto";
@@ -23669,7 +23699,7 @@ var JobRunner = class {
     this.registry = registry;
   }
   recoverStaleRuns() {
-    const stale = this.db.select({ id: runs.id, status: runs.status }).from(runs).where(inArray8(runs.status, ["running", "queued"])).all();
+    const stale = this.db.select({ id: runs.id, status: runs.status }).from(runs).where(inArray9(runs.status, ["running", "queued"])).all();
     if (stale.length === 0) return;
     const now = (/* @__PURE__ */ new Date()).toISOString();
     for (const run of stale) {
@@ -23707,7 +23737,7 @@ var JobRunner = class {
         throw new Error(`Run ${runId} is not executable from status '${existingRun.status}'`);
       }
       if (existingRun.status === "queued") {
-        this.db.update(runs).set({ status: "running", startedAt: now }).where(and16(eq27(runs.id, runId), eq27(runs.status, "queued"))).run();
+        this.db.update(runs).set({ status: "running", startedAt: now }).where(and17(eq27(runs.id, runId), eq27(runs.status, "queued"))).run();
       }
       this.throwIfRunCancelled(runId);
       const project = this.db.select().from(projects).where(eq27(projects.id, projectId)).get();
@@ -23732,7 +23762,7 @@ var JobRunner = class {
       }
       log.info("run.dispatch", { runId, providerCount: activeProviders.length, providers: activeProviders.map((p) => p.adapter.name) });
       const scopedQueryNames = parseJsonColumn(existingRun.queries, null);
-      projectQueries = scopedQueryNames ? this.db.select().from(queries).where(and16(eq27(queries.projectId, projectId), inArray8(queries.query, scopedQueryNames))).all() : this.db.select().from(queries).where(eq27(queries.projectId, projectId)).all();
+      projectQueries = scopedQueryNames ? this.db.select().from(queries).where(and17(eq27(queries.projectId, projectId), inArray9(queries.query, scopedQueryNames))).all() : this.db.select().from(queries).where(eq27(queries.projectId, projectId)).all();
       const projectCompetitors = this.db.select().from(competitors).where(eq27(competitors.projectId, projectId)).all();
       const competitorDomains = projectCompetitors.map((c) => c.domain);
       const allDomains = effectiveDomains({
@@ -24068,7 +24098,7 @@ function buildPhases(input) {
 
 // src/gsc-sync.ts
 import crypto25 from "crypto";
-import { eq as eq28, and as and17, sql as sql11 } from "drizzle-orm";
+import { eq as eq28, and as and18, sql as sql11 } from "drizzle-orm";
 var log2 = createLogger("GscSync");
 function formatDate3(d) {
   return d.toISOString().split("T")[0];
@@ -24120,7 +24150,7 @@ async function executeGscSync(db, runId, projectId, opts) {
     });
     log2.info("fetch.complete", { runId, projectId, rowCount: rows.length });
     db.delete(gscSearchData).where(
-      and17(
+      and18(
         eq28(gscSearchData.projectId, projectId),
         sql11`${gscSearchData.date} >= ${startDate}`,
         sql11`${gscSearchData.date} <= ${endDate}`
@@ -24209,7 +24239,7 @@ async function executeGscSync(db, runId, projectId, opts) {
       }
     }
     const snapshotDate = formatDate3(/* @__PURE__ */ new Date());
-    db.delete(gscCoverageSnapshots).where(and17(eq28(gscCoverageSnapshots.projectId, projectId), eq28(gscCoverageSnapshots.date, snapshotDate))).run();
+    db.delete(gscCoverageSnapshots).where(and18(eq28(gscCoverageSnapshots.projectId, projectId), eq28(gscCoverageSnapshots.date, snapshotDate))).run();
     db.insert(gscCoverageSnapshots).values({
       id: crypto25.randomUUID(),
       projectId,
@@ -24232,7 +24262,7 @@ async function executeGscSync(db, runId, projectId, opts) {
 
 // src/gsc-inspect-sitemap.ts
 import crypto26 from "crypto";
-import { eq as eq29, and as and18 } from "drizzle-orm";
+import { eq as eq29, and as and19 } from "drizzle-orm";
 
 // src/sitemap-parser.ts
 var log3 = createLogger("SitemapParser");
@@ -24448,7 +24478,7 @@ async function executeInspectSitemap(db, runId, projectId, opts) {
       }
     }
     const snapshotDate = (/* @__PURE__ */ new Date()).toISOString().split("T")[0];
-    db.delete(gscCoverageSnapshots).where(and18(eq29(gscCoverageSnapshots.projectId, projectId), eq29(gscCoverageSnapshots.date, snapshotDate))).run();
+    db.delete(gscCoverageSnapshots).where(and19(eq29(gscCoverageSnapshots.projectId, projectId), eq29(gscCoverageSnapshots.date, snapshotDate))).run();
     db.insert(gscCoverageSnapshots).values({
       id: crypto26.randomUUID(),
       projectId,
@@ -24659,7 +24689,7 @@ async function executeBingInspectSitemap(db, runId, projectId, opts) {
 // src/commoncrawl-sync.ts
 import crypto28 from "crypto";
 import path10 from "path";
-import { and as and19, eq as eq31, sql as sql12 } from "drizzle-orm";
+import { and as and20, eq as eq31, sql as sql12 } from "drizzle-orm";
 var log6 = createLogger("CommonCrawlSync");
 var INSERT_CHUNK_SIZE = 1e4;
 function defaultDeps() {
@@ -24850,7 +24880,7 @@ function computeSummary(rows) {
 // src/backlink-extract.ts
 import crypto29 from "crypto";
 import fs8 from "fs";
-import { and as and20, desc as desc15, eq as eq32 } from "drizzle-orm";
+import { and as and21, desc as desc15, eq as eq32 } from "drizzle-orm";
 var log7 = createLogger("BacklinkExtract");
 function defaultDeps2() {
   return {
@@ -24896,7 +24926,7 @@ async function executeBacklinkExtract(db, runId, projectId, opts = {}) {
     const targetDomain = project.canonicalDomain;
     db.transaction((tx) => {
       tx.delete(backlinkDomains).where(
-        and20(eq32(backlinkDomains.projectId, projectId), eq32(backlinkDomains.release, release))
+        and21(eq32(backlinkDomains.projectId, projectId), eq32(backlinkDomains.release, release))
       ).run();
       if (rows.length > 0) {
         const values = rows.map((r) => ({
@@ -25313,7 +25343,7 @@ var ProviderRegistry = class {
 
 // src/scheduler.ts
 import cron from "node-cron";
-import { and as and21, eq as eq34 } from "drizzle-orm";
+import { and as and22, eq as eq34 } from "drizzle-orm";
 var log9 = createLogger("Scheduler");
 function taskKey(projectId, kind) {
   return `${projectId}::${kind}`;
@@ -25358,7 +25388,7 @@ var Scheduler = class {
       this.stopTask(key, existing, "Stopped");
       this.tasks.delete(key);
     }
-    const schedule = this.db.select().from(schedules).where(and21(eq34(schedules.projectId, projectId), eq34(schedules.kind, kind))).get();
+    const schedule = this.db.select().from(schedules).where(and22(eq34(schedules.projectId, projectId), eq34(schedules.kind, kind))).get();
     if (schedule && schedule.enabled === 1) {
       this.registerCronTask(schedule);
     }
@@ -25482,7 +25512,7 @@ var Scheduler = class {
 };
 
 // src/notifier.ts
-import { eq as eq35, desc as desc16, and as and22, inArray as inArray9, or as or4 } from "drizzle-orm";
+import { eq as eq35, desc as desc16, and as and23, inArray as inArray10, or as or4 } from "drizzle-orm";
 import crypto31 from "crypto";
 var log10 = createLogger("Notifier");
 var Notifier = class {
@@ -25589,7 +25619,7 @@ var Notifier = class {
   computeTransitions(runId, projectId) {
     const thisRun = this.db.select().from(runs).where(eq35(runs.id, runId)).get();
     if (!thisRun) return [];
-    const groupSiblings = this.db.select().from(runs).where(and22(
+    const groupSiblings = this.db.select().from(runs).where(and23(
       eq35(runs.projectId, projectId),
       eq35(runs.kind, thisRun.kind),
       eq35(runs.createdAt, thisRun.createdAt)
@@ -25615,7 +25645,7 @@ var Notifier = class {
     );
     const RECENT_FETCH_LIMIT = Math.max(8, locationCount * 4);
     const recentRuns = this.db.select().from(runs).where(
-      and22(
+      and23(
         eq35(runs.projectId, projectId),
         eq35(runs.kind, thisRun.kind),
         or4(eq35(runs.status, "completed"), eq35(runs.status, "partial"))
@@ -25635,13 +25665,13 @@ var Notifier = class {
       provider: querySnapshots.provider,
       location: querySnapshots.location,
       citationState: querySnapshots.citationState
-    }).from(querySnapshots).leftJoin(queries, eq35(querySnapshots.queryId, queries.id)).where(inArray9(querySnapshots.runId, currentRunIds)).all();
+    }).from(querySnapshots).leftJoin(queries, eq35(querySnapshots.queryId, queries.id)).where(inArray10(querySnapshots.runId, currentRunIds)).all();
     const previousSnapshots = this.db.select({
       queryId: querySnapshots.queryId,
       provider: querySnapshots.provider,
       location: querySnapshots.location,
       citationState: querySnapshots.citationState
-    }).from(querySnapshots).where(inArray9(querySnapshots.runId, previousRunIds)).all();
+    }).from(querySnapshots).where(inArray10(querySnapshots.runId, previousRunIds)).all();
     const prevMap = /* @__PURE__ */ new Map();
     for (const s of previousSnapshots) {
       prevMap.set(`${s.queryId}:${s.provider}:${s.location ?? ""}`, s.citationState);
@@ -26152,7 +26182,7 @@ function resolveSessionProviderAndModel(config, opts) {
 
 // src/agent/memory-store.ts
 import crypto32 from "crypto";
-import { and as and23, desc as desc17, eq as eq37, like as like2, sql as sql13 } from "drizzle-orm";
+import { and as and24, desc as desc17, eq as eq37, like as like2, sql as sql13 } from "drizzle-orm";
 var COMPACTION_KEY_PREFIX = "compaction:";
 var COMPACTION_NOTES_PER_SESSION = 3;
 function rowToDto2(row) {
@@ -26197,12 +26227,12 @@ function upsertMemoryEntry(db, args) {
       updatedAt: now
     }
   }).run();
-  const row = db.select().from(agentMemory).where(and23(eq37(agentMemory.projectId, args.projectId), eq37(agentMemory.key, args.key))).get();
+  const row = db.select().from(agentMemory).where(and24(eq37(agentMemory.projectId, args.projectId), eq37(agentMemory.key, args.key))).get();
   if (!row) throw new Error("memory upsert produced no row");
   return rowToDto2(row);
 }
 function deleteMemoryEntry(db, projectId, key) {
-  const result = db.delete(agentMemory).where(and23(eq37(agentMemory.projectId, projectId), eq37(agentMemory.key, key))).run();
+  const result = db.delete(agentMemory).where(and24(eq37(agentMemory.projectId, projectId), eq37(agentMemory.key, key))).run();
   const changes = result.changes ?? 0;
   return changes > 0;
 }
@@ -26231,7 +26261,7 @@ function writeCompactionNote(db, args) {
     }).run();
     const sessionPrefix = `${COMPACTION_KEY_PREFIX}${args.sessionId}:`;
     const existing = tx.select({ id: agentMemory.id, updatedAt: agentMemory.updatedAt }).from(agentMemory).where(
-      and23(
+      and24(
         eq37(agentMemory.projectId, args.projectId),
         like2(agentMemory.key, `${sessionPrefix}%`)
       )
@@ -26240,7 +26270,7 @@ function writeCompactionNote(db, args) {
     if (stale.length > 0) {
       tx.delete(agentMemory).where(sql13`${agentMemory.id} IN (${sql13.join(stale.map((s) => sql13`${s}`), sql13`, `)})`).run();
     }
-    const row = tx.select().from(agentMemory).where(and23(eq37(agentMemory.projectId, args.projectId), eq37(agentMemory.key, key))).get();
+    const row = tx.select().from(agentMemory).where(and24(eq37(agentMemory.projectId, args.projectId), eq37(agentMemory.key, key))).get();
     if (row) inserted = rowToDto2(row);
   });
   if (!inserted) throw new Error("compaction note write produced no row");

package/dist/cli.js

@@ -20,7 +20,7 @@ import {
   setTelemetrySource,
   showFirstRunNotice,
   trackEvent
-} from "./chunk-JJHBPITI.js";
+} from "./chunk-DZENHID5.js";
 import {
   CliError,
   EXIT_SYSTEM_ERROR,
@@ -2020,10 +2020,17 @@ Sessions already started (recover with \`canonry discover show ${project} <id>\`
     }
     for (const { angle, start } of runs2) {
       if (angle) console.log(`[${angle}]`);
-      console.log(`Discovery run started: ${start.runId}`);
-      console.log(`  Session: ${start.sessionId}`);
-      console.log(`  Status:  ${start.status}`);
-      console.log(`  Tail:    canonry discover show ${project} ${start.sessionId}`);
+      if (start.consolidated) {
+        console.log(`Reusing in-flight discovery session: ${start.sessionId}`);
+        console.log(`  Run:     ${start.runId}`);
+        console.log(`  Status:  ${start.status}`);
+        console.log(`  Tail:    canonry discover show ${project} ${start.sessionId}`);
+      } else {
+        console.log(`Discovery run started: ${start.runId}`);
+        console.log(`  Session: ${start.sessionId}`);
+        console.log(`  Status:  ${start.status}`);
+        console.log(`  Tail:    canonry discover show ${project} ${start.sessionId}`);
+      }
       if (runs2.length > 1) console.log();
     }
     return;

package/dist/index.js

@@ -1,6 +1,6 @@
 import {
   createServer
-} from "./chunk-JJHBPITI.js";
+} from "./chunk-DZENHID5.js";
 import {
   loadConfig
 } from "./chunk-5EBN7736.js";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ainyc/canonry",
-  "version": "4.33.0",
+  "version": "4.33.1",
   "type": "module",
   "description": "Agent-first open-source AEO operating platform - track how answer engines cite your domain",
   "license": "FSL-1.1-ALv2",
@@ -61,21 +61,21 @@
     "tsup": "^8.5.1",
     "tsx": "^4.19.0",
     "@ainyc/canonry-api-routes": "0.0.0",
+    "@ainyc/canonry-config": "0.0.0",
     "@ainyc/canonry-contracts": "0.0.0",
-    "@ainyc/canonry-intelligence": "0.0.0",
+    "@ainyc/canonry-integration-bing": "0.0.0",
     "@ainyc/canonry-db": "0.0.0",
+    "@ainyc/canonry-intelligence": "0.0.0",
     "@ainyc/canonry-integration-cloud-run": "0.0.0",
-    "@ainyc/canonry-config": "0.0.0",
-    "@ainyc/canonry-integration-bing": "0.0.0",
-    "@ainyc/canonry-integration-google": "0.0.0",
+    "@ainyc/canonry-integration-traffic": "0.0.0",
+    "@ainyc/canonry-provider-cdp": "0.0.0",
     "@ainyc/canonry-integration-commoncrawl": "0.0.0",
+    "@ainyc/canonry-integration-google": "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-integration-traffic": "0.0.0",
+    "@ainyc/canonry-provider-openai": "0.0.0",
     "@ainyc/canonry-provider-claude": "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": {