@autorag/librarian 2.13.5 → 2.13.7

package/dist/agent/agent.d.ts

@@ -50,12 +50,12 @@ export interface AutoRAGRefreshLimits {
     /** Retries for a transient per-file I/O error (EMFILE/EBUSY/EIO/ETIMEDOUT) before the file is skipped. Default 3. */
     readonly transientRetries?: number;
     /**
-     * Most cloud-file placeholders (files whose data is not on disk) one refresh may download.
-     * Reading a placeholder downloads it, so this bounds disk and network use per refresh; the rest wait
-     * for the next one. Default 200.
+     * Most cloud-file placeholders (files whose data is not on disk) one refresh batch may download.
+     * Reading a placeholder downloads it; the next one over the budget starts a new batch, so every
+     * placeholder is still read in the same refresh. Default 200.
      */
     readonly cloudMaxFiles?: number;
-    /** Most placeholder bytes one refresh may download; the first file is always allowed. Default 536870912 (512 MiB). */
+    /** Most placeholder bytes one refresh batch may download; a batch's first file is always allowed. Default 536870912 (512 MiB). */
     readonly cloudMaxBytes?: number;
     /** Release a downloaded cloud file again once it is parsed and indexed. Default true. */
     readonly cloudEvictAfterParse?: boolean;
@@ -952,7 +952,7 @@ export declare class AutoRAGAgent {
      */
     startWatchRefresh(options?: AutoRAGWatchRefreshOptions): AutoRAGWatchRefreshHandle;
     private defaultWatcherFactory;
-    syncParsedMirrors(force?: boolean, options?: Pick<ParsedMirrorSyncOptions, "signal" | "onProgress" | "retrySkipped">): Promise<ParsedMirrorSyncResult>;
+    syncParsedMirrors(force?: boolean, options?: Pick<ParsedMirrorSyncOptions, "signal" | "onProgress" | "retrySkipped" | "onBatch" | "batchSize">): Promise<ParsedMirrorSyncResult>;
     /**
      * Lightweight stat-only staleness scan used when `refresh` is called with
      * methods that exclude parsed mirrors (e.g. only `datasources` or `jikji`).

package/dist/agent/dupey-tool.d.ts

@@ -4,8 +4,6 @@ import type { DupeyScanResult } from "../dupey/index.js";
 export declare const SCAN_DUPLICATE_DOCUMENTS_TOOL_NAME = "scan_duplicate_documents";
 export interface ScanDuplicateDocumentsDetails {
     readonly scans: readonly DupeyScanResult[];
-    /** Roots inside a cloud or network folder: `dupey scan` opens file contents, so they were not scanned. */
-    readonly skippedRoots: readonly string[];
     readonly familyCount: number;
     readonly exactDuplicateCount: number;
 }

package/dist/agent/pipelined-sync.d.ts

@@ -0,0 +1,34 @@
+import type { ParsedMirrorBatch, ParsedMirrorBatchOutcome } from "../mirror/sync.js";
+/** One sync run: the 1-based run number and the sources of every batch committed since the previous run began. */
+export interface PipelinedSyncRun {
+    readonly index: number;
+    readonly virtualPaths: readonly string[];
+}
+export interface PipelinedSyncOptions {
+    /**
+     * Index everything committed so far. Resolving with `stop` ends the pipeline (a held lock means another
+     * process owns the index, so every further run would fail the same way); a rejection is raised from the
+     * next {@link PipelinedSync.batch} call.
+     */
+    readonly sync: (run: PipelinedSyncRun) => Promise<ParsedMirrorBatchOutcome | undefined>;
+}
+/**
+ * Runs index syncs alongside the parse loop instead of between its batches.
+ *
+ * A sync only sees what is committed when it starts, and the indexer batches and commits on its own, so the
+ * first batch starts a sync right away while parsing carries on. At most one sync is in flight; when it ends,
+ * the next one covers every batch committed meanwhile. A slow sync therefore turns into few large ones
+ * instead of one per parse batch, and each sync pays its start-up and end-of-run maintenance once.
+ */
+export declare class PipelinedSync {
+    #private;
+    constructor(options: PipelinedSyncOptions);
+    /**
+     * Hand over a committed batch. Returns at once, except for the final batch: it waits until a sync that
+     * started after every batch was committed has finished, and always runs one, so a sync that failed
+     * earlier is retried even when nothing changed.
+     */
+    batch(batch: Pick<ParsedMirrorBatch, "virtualPaths" | "final">): Promise<ParsedMirrorBatchOutcome | undefined>;
+    /** Wait for the sync in flight, if any, and start no further one. Never rejects. */
+    settle(): Promise<void>;
+}

package/dist/agent/search-documents.d.ts

@@ -9,7 +9,7 @@ export type SearchDocumentDiagnosticSeverity = "info" | "warning" | "error";
  * (#21/#22) does not require another source-breaking type change; codes not yet
  * emitted are still valid members.
  */
-export type SearchDocumentDiagnosticCode = "empty-query" | "no-verified-results" | "unknown-warning" | "caller-tool-dropped" | "minsync-staging-excluded" | "minsync-file-skipped" | "minsync-sync-failed" | "minsync-lock-held" | "embedder-unavailable" | "embedding-identity-mismatch" | "parser-skipped" | "parser-failed" | "parser-crashed" | "source-changed" | "source-vanished" | "source-unreadable" | "source-read-failed" | "mirror-write-failed" | "parser-warning" | "duplicate-excluded" | "unsupported-file" | "stale-index" | "deleted-mirror" | "search-root-unavailable" | "mass-delete-suppressed" | "walk-truncated" | "directory-unreadable" | "cloud-download-deferred" | "cloud-placeholder-skipped" | "cloud-evict-failed" | "cloud-fingerprint-failed" | "pdf-extract-thin" | "retrieval-method-failed" | "rerank-failed" | "rerank-disabled" | "jikji-unavailable" | "jikji-prepare-failed" | "jikji-cloud-root-skipped" | "jikji-find-failed" | "jikji-discover-failed" | "jikji-unsupported-version" | "everything-index-failed" | "fsearch-index-failed" | "fsearch-binary-missing" | "refresh-failed" | "refresh-interrupted" | "watch-failed" | "watch-limited" | "unknown-datasource-skill" | "no-final-answer" | "model-request-failed" | "tool-call-burst" | "decision-model-checked" | "decision-model-fallback" | "query-routed" | "query-route-fallback" | "query-decomposition-failed" | "follow-up-skipped" | "follow-up-check-fallback" | "datasources-selected" | "datasource-selection-fallback" | "web-datasource-id-taken" | "web-search-failed" | "citation-without-result" | "line-reference-unverified" | "citation-irrelevant" | "citation-relevance-fallback" | "evidence-judged" | "evidence-judgment-fallback" | "memory-search-fallback" | "self-config-unavailable" | "self-config-rolled-back" | "organize-skill-unavailable" | "embedding-unavailable";
+export type SearchDocumentDiagnosticCode = "empty-query" | "no-verified-results" | "unknown-warning" | "caller-tool-dropped" | "minsync-staging-excluded" | "minsync-file-skipped" | "minsync-sync-failed" | "minsync-lock-held" | "embedder-unavailable" | "embedding-identity-mismatch" | "parser-skipped" | "parser-failed" | "parser-crashed" | "source-changed" | "source-vanished" | "source-unreadable" | "source-read-failed" | "mirror-write-failed" | "parser-warning" | "duplicate-excluded" | "unsupported-file" | "stale-index" | "deleted-mirror" | "search-root-unavailable" | "mass-delete-suppressed" | "walk-truncated" | "directory-unreadable" | "cloud-placeholder-skipped" | "cloud-evict-failed" | "cloud-fingerprint-failed" | "pdf-extract-thin" | "retrieval-method-failed" | "rerank-failed" | "rerank-disabled" | "jikji-unavailable" | "jikji-prepare-failed" | "jikji-cloud-root-skipped" | "jikji-find-failed" | "jikji-discover-failed" | "jikji-unsupported-version" | "everything-index-failed" | "fsearch-index-failed" | "fsearch-binary-missing" | "refresh-failed" | "refresh-interrupted" | "watch-failed" | "watch-limited" | "unknown-datasource-skill" | "retired-datasource-skill" | "no-final-answer" | "model-request-failed" | "tool-call-burst" | "decision-model-checked" | "decision-model-fallback" | "query-routed" | "query-route-fallback" | "query-decomposition-failed" | "follow-up-skipped" | "follow-up-check-fallback" | "datasources-selected" | "datasource-selection-fallback" | "web-datasource-id-taken" | "web-search-failed" | "citation-without-result" | "line-reference-unverified" | "citation-irrelevant" | "citation-relevance-fallback" | "evidence-judged" | "evidence-judgment-fallback" | "memory-search-fallback" | "self-config-unavailable" | "self-config-rolled-back" | "organize-skill-unavailable" | "embedding-unavailable";
 export interface SearchDocumentDiagnostic {
     readonly code: SearchDocumentDiagnosticCode;
     readonly severity: SearchDocumentDiagnosticSeverity;

package/dist/cli/commands/duplicates.d.ts

@@ -1,11 +1,10 @@
 import type { DupeyScanResult } from "../../dupey/index.js";
-import { type CloudFileProbe } from "../../filesystem/cloud-placeholder.js";
 import type { CommandContext } from "./types.js";
-export type DuplicateScanner = (path: string, excludeDirs: readonly string[]) => Promise<DupeyScanResult>;
+export type DuplicateScanner = (path: string) => Promise<DupeyScanResult>;
 /**
  * `autorag duplicates` — scan configured roots for exact and near duplicate
  * document families. Never deletes or moves source files.
  *
  * Exit codes: 0 on success, 2 on config error, 1 on runtime error.
  */
-export declare function runDuplicates(ctx: CommandContext, scanner?: DuplicateScanner, cloudProbe?: CloudFileProbe | undefined): Promise<number>;
+export declare function runDuplicates(ctx: CommandContext, scanner?: DuplicateScanner): Promise<number>;