@equationalapplications/core-llm-wiki 4.23.0 → 5.0.0

package/dist/{chunk-MYZJLVX4.mjs → chunk-W53E3I44.mjs}

@@ -1569,6 +1569,8 @@ var ONTOLOGY_BACKFILL_RECHECK_MS = 7 * 24 * 60 * 60 * 1e3;
 var HEAL_MAX_ANCHORS = 50;
 var HEAL_ANCHOR_SEARCH_OVERFETCH = 4;
 var HEAL_MAX_PROMPT_CHARS = 4e4;
+var HEAL_BATCH_SIZE = 25;
+var HEAL_RECHECK_MS = 7 * 24 * 60 * 60 * 1e3;
 var MaintenanceService = class {
   constructor(db, prefix, options, entryRepo, taskRepo, eventRepo, metadataRepo, searchService, jobManager, embeddingService, promptService, ontologyService) {
     this.db = db;
@@ -1669,7 +1671,7 @@ var MaintenanceService = class {
   async runHeal(entityId, options) {
     this.jobManager.acquireLock("heal", entityId);
     try {
-      await this.doRunHeal(entityId, options?.promptOverride);
+      return await this.doRunHeal(entityId, options);
     } finally {
       this.jobManager.releaseLock("heal", entityId);
     }
@@ -1921,9 +1923,21 @@ var MaintenanceService = class {
     }
     this.searchService.evictCache(entityId);
   }
-  /** Core heal pass (locks handled by {@link runHeal}). Package-internal orchestration hook. */
-  async doRunHeal(entityId, promptOverride) {
+  /**
+   * Core heal pass (locks handled by {@link runHeal}). Package-internal orchestration hook.
+   *
+   * Bounded: at most `batchSize` candidates per pass (#67). Loop on
+   * `result.remaining > 0` for convergence — see {@link HealResult.remaining}
+   * for what convergence means here.
+   */
+  async doRunHeal(entityId, options) {
+    const promptOverride = options?.promptOverride;
+    const batchSize = options?.batchSize ?? HEAL_BATCH_SIZE;
+    if (!Number.isInteger(batchSize) || batchSize < 1) {
+      throw new Error("Invalid batchSize: must be an integer >= 1");
+    }
     const now = Date.now();
+    const recheckCutoff = now - HEAL_RECHECK_MS;
     const orphanAfterDays = this.options.config?.orphanAfterDays !== void 0 ? this.options.config?.orphanAfterDays : 30;
     const staleInferredAfterDays = this.options.config?.staleInferredAfterDays !== void 0 ? this.options.config?.staleInferredAfterDays : 60;
     const MS_PER_DAY = 24 * 60 * 60 * 1e3;
@@ -1934,6 +1948,7 @@ var MaintenanceService = class {
       throw new Error("Invalid staleInferredAfterDays: must be a finite number >= 0 or null");
     }
     const orphanedIds = [];
+    const staleDowngradedIds = [];
     await this.db.withTransactionAsync(async (tx) => {
       if (orphanAfterDays !== null) {
         const orphanThreshold = now - orphanAfterDays * MS_PER_DAY;
@@ -1941,7 +1956,7 @@ var MaintenanceService = class {
       }
       if (staleInferredAfterDays !== null) {
         const staleThreshold = now - staleInferredAfterDays * MS_PER_DAY;
-        await this.entryRepo.downgradeStaleInferred(entityId, staleThreshold, tx);
+        staleDowngradedIds.push(...await this.entryRepo.downgradeStaleInferred(entityId, staleThreshold, tx));
       }
     });
     for (const factId of orphanedIds) {
@@ -1951,7 +1966,21 @@ var MaintenanceService = class {
         console.warn(`[WikiMemory] onEmbeddingPersisted hook failed during heal orphan pass for ${factId}:`, hookErr);
       }
     }
-    const healCandidates = await this.entryRepo.findHealCandidatesByEntityId(entityId);
+    const healCandidates = await this.entryRepo.findHealCandidatesByEntityId(entityId, batchSize, recheckCutoff);
+    if (healCandidates.length === 0) {
+      await this.searchService.sync(entityId);
+      this.searchService.evictCache(entityId);
+      const counts2 = await this.entryRepo.countHealCandidatesByEntityId(entityId, recheckCutoff);
+      return {
+        scanned: 0,
+        downgraded: staleDowngradedIds.length,
+        deleted: orphanedIds.length,
+        newFactsCreated: 0,
+        skipped: 0,
+        remaining: counts2.eligible,
+        deferred: counts2.deferred
+      };
+    }
     const allTasks = await this.taskRepo.findAllPending([entityId]);
     const recentEvents = await this.eventRepo.getRecent(entityId, 20);
     const toPromptShape = (f) => {
@@ -2004,7 +2033,7 @@ var MaintenanceService = class {
     const validNewFacts = newFacts.map(validateFact).filter((f) => f !== null);
     const insertedFacts = [];
     const uniqueDeletedFactIds = Array.from(new Set(safeDeleted));
-    const healFactsForDedupe = [...healCandidates];
+    const healFactsForDedupe = (await this.entryRepo.findInferredTitlesByEntityId(entityId)).filter((f) => !safeDeletedSet.has(f.id));
     await this.db.withTransactionAsync(async (tx) => {
       await this.entryRepo.downgradeByIds(safeDowngraded, entityId, tx);
       await this.entryRepo.softDeleteByIds(safeDeleted, entityId, tx);
@@ -2013,7 +2042,6 @@ var MaintenanceService = class {
         let skip = false;
         if (newTokens.size >= MIN_TOKENS_TO_QUALIFY) {
           for (const existing of healFactsForDedupe) {
-            if (existing.source_type !== "librarian_inferred") continue;
             const existingTokens = titleTokens(existing.title);
             if (existingTokens.size >= MIN_TOKENS_TO_QUALIFY) {
               if (jaccardScore(newTokens, existingTokens) >= FUZZY_THRESHOLD) {
@@ -2043,8 +2071,14 @@ var MaintenanceService = class {
         };
         await this.entryRepo.upsert(factObj, tx);
         insertedFacts.push({ id, entity_id: entityId, title: fact.title, body: fact.body, tags: JSON.stringify(fact.tags) });
-        healFactsForDedupe.push(factObj);
+        healFactsForDedupe.push({ id, title: fact.title });
       }
+      await this.entryRepo.markHealChecked(
+        [...healCandidates.map((f) => f.id), ...insertedFacts.map((f) => f.id)],
+        entityId,
+        now,
+        tx
+      );
     });
     await this.searchService.sync(entityId);
     for (const factId of uniqueDeletedFactIds) {
@@ -2058,6 +2092,20 @@ var MaintenanceService = class {
       await this.embeddingService.embedFact(fact);
     }
     this.searchService.evictCache(entityId);
+    let scanned = outcome.skipped.length;
+    for (const batchResult of outcome.results) scanned += batchResult.batch.length;
+    const allDowngraded = /* @__PURE__ */ new Set([...staleDowngradedIds, ...safeDowngraded]);
+    const allDeleted = /* @__PURE__ */ new Set([...orphanedIds, ...uniqueDeletedFactIds]);
+    const counts = await this.entryRepo.countHealCandidatesByEntityId(entityId, recheckCutoff);
+    return {
+      scanned,
+      downgraded: allDowngraded.size,
+      deleted: allDeleted.size,
+      newFactsCreated: insertedFacts.length,
+      skipped: outcome.skipped.length,
+      remaining: counts.eligible,
+      deferred: counts.deferred
+    };
   }
   /** Core ontology backfill pass (locks handled by {@link runOntologyBackfill}). Package-internal orchestration hook. */
   async doRunOntologyBackfill(entityId, options) {
@@ -3433,6 +3481,7 @@ var WriteService = class {
     let shouldRunLibrarian = false;
     let librarianCount = 0;
     let prevMemoryCheckpoint = 0;
+    let eventCount = 0;
     await this.db.withTransactionAsync(async (tx) => {
       await this.eventRepo.add(newEvent, tx);
       const threshold = this.options.config?.autoLibrarianThreshold || 20;
@@ -3440,6 +3489,7 @@ var WriteService = class {
         this.eventRepo.count(entityId, tx),
         this.metadataRepo.getCheckpoint(entityId, tx)
       ]);
+      eventCount = count;
       let memoryCheckpoint = cp.memory ?? 0;
       if (memoryCheckpoint > count) memoryCheckpoint = 0;
       if (count - memoryCheckpoint >= threshold) {
@@ -3461,6 +3511,8 @@ var WriteService = class {
         if (!(e instanceof WikiBusyError)) throw e;
         await this.metadataRepo.updateCheckpoint(entityId, { memory: prevMemoryCheckpoint }, this.db);
       }
+    } else if (!this.jobManager.isBlocked("librarian", entityId)) {
+      this.maybeRunHeal(entityId, eventCount).catch(console.error);
     }
   }
   async runLibrarianThenMaybeHeal(entityId, currentEventCount, prevCheckpoint) {
@@ -3471,6 +3523,14 @@ var WriteService = class {
       await this.metadataRepo.updateCheckpoint(entityId, { memory: prevCheckpoint }, this.db);
       throw e;
     }
+    await this.maybeRunHeal(entityId, currentEventCount);
+  }
+  /**
+   * Run one bounded auto-heal pass if the heal checkpoint has fallen
+   * `autoHealThreshold` events behind. Called after every write (see
+   * {@link write}) so a partial pass retries on the next write.
+   */
+  async maybeRunHeal(entityId, currentEventCount) {
     const autoHealThreshold = this.options.config?.autoHealThreshold || 100;
     const cp = await this.metadataRepo.getCheckpoint(entityId, this.db);
     let healCheckpoint = cp.heal ?? 0;
@@ -3478,8 +3538,10 @@ var WriteService = class {
     const shouldRunHeal = currentEventCount - healCheckpoint >= autoHealThreshold;
     if (shouldRunHeal && this.jobManager.tryAcquireAutoHealLock(entityId)) {
       try {
-        await this.maintenanceService.doRunHeal(entityId);
-        await this.metadataRepo.updateCheckpoint(entityId, { heal: currentEventCount }, this.db);
+        const result = await this.maintenanceService.doRunHeal(entityId);
+        if (result.remaining === 0) {
+          await this.metadataRepo.updateCheckpoint(entityId, { heal: currentEventCount }, this.db);
+        }
       } finally {
         this.jobManager.releaseLock("heal", entityId);
       }
@@ -3487,6 +3549,6 @@ var WriteService = class {
   }
 };
 
-export { BaseRepository, EmbeddingService, HOOK_TIMEOUT_MARKER, ImportExportService, IngestionService, JobManager, MaintenanceService, MetadataRepository, ONTOLOGY_BACKFILL_BATCH_SIZE, ONTOLOGY_BACKFILL_MAX_PROMPT_CHARS, ONTOLOGY_BACKFILL_RECHECK_MS, ONTOLOGY_BACKFILL_SYSTEM_PROMPT, PromptService, PrunePartialFailureError, RetrievalService, SearchService, WikiBusyError, WikiTransactionError, WriteService, __privateAdd, __privateGet, __privateSet, configureRandomSource, emptyManifest, entitySummaryMetaKey, extractSqliteCode, generateId, normalizeSourceHash, normalizeSourceRef, normalizeTitleKey, parseEmbedding, resolveEdgeDefinitions, resolveNodeType, validateInlineEdges, validateManifest };
-//# sourceMappingURL=chunk-MYZJLVX4.mjs.map
-//# sourceMappingURL=chunk-MYZJLVX4.mjs.map
+export { BaseRepository, EmbeddingService, HEAL_BATCH_SIZE, HEAL_RECHECK_MS, HOOK_TIMEOUT_MARKER, ImportExportService, IngestionService, JobManager, MaintenanceService, MetadataRepository, ONTOLOGY_BACKFILL_BATCH_SIZE, ONTOLOGY_BACKFILL_MAX_PROMPT_CHARS, ONTOLOGY_BACKFILL_RECHECK_MS, ONTOLOGY_BACKFILL_SYSTEM_PROMPT, PromptService, PrunePartialFailureError, RetrievalService, SearchService, WikiBusyError, WikiTransactionError, WriteService, __privateAdd, __privateGet, __privateSet, configureRandomSource, emptyManifest, entitySummaryMetaKey, extractSqliteCode, generateId, normalizeSourceHash, normalizeSourceRef, normalizeTitleKey, parseEmbedding, resolveEdgeDefinitions, resolveNodeType, validateInlineEdges, validateManifest };
+//# sourceMappingURL=chunk-W53E3I44.mjs.map
+//# sourceMappingURL=chunk-W53E3I44.mjs.map