@autorag/librarian 2.2.0 → 2.4.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (106) hide show
  1. package/README.md +194 -47
  2. package/dist/agent/agent.d.ts +34 -47
  3. package/dist/agent/bash-tool.d.ts +2 -11
  4. package/dist/agent/dupey-tool.d.ts +15 -0
  5. package/dist/agent/emit-results-tool.d.ts +14 -0
  6. package/dist/agent/index.d.ts +1 -0
  7. package/dist/agent/local-model.d.ts +12 -0
  8. package/dist/agent/search-bm25-tool.d.ts +11 -3
  9. package/dist/agent/search-documents.d.ts +2 -2
  10. package/dist/agent/search-minsync-tool.d.ts +2 -2
  11. package/dist/agent/system-prompt.d.ts +1 -2
  12. package/dist/cli/commands/duplicates.d.ts +4 -0
  13. package/dist/cli/commands/health.d.ts +9 -65
  14. package/dist/cli/commands/index.d.ts +3 -3
  15. package/dist/cli/commands/init.d.ts +1 -1
  16. package/dist/cli/commands/search.d.ts +2 -2
  17. package/dist/cli/commands/tui.d.ts +36 -0
  18. package/dist/cli/commands/ui.d.ts +13 -0
  19. package/dist/cli/config.d.ts +27 -22
  20. package/dist/cli/index.js +14965 -12563
  21. package/dist/cli/tui-commands.d.ts +16 -0
  22. package/dist/cli/tui-session-store.d.ts +16 -0
  23. package/dist/datasource/access-context.d.ts +6 -12
  24. package/dist/datasource/aliased-skill.d.ts +28 -0
  25. package/dist/datasource/chunk-store.d.ts +2 -0
  26. package/dist/datasource/connector-skill.d.ts +3 -0
  27. package/dist/datasource/connector.d.ts +4 -0
  28. package/dist/datasource/crawler-skill.d.ts +6 -0
  29. package/dist/datasource/crawler-types.d.ts +2 -0
  30. package/dist/datasource/described-skill.d.ts +17 -0
  31. package/dist/datasource/index.d.ts +4 -2
  32. package/dist/datasource/result-filter.d.ts +2 -1
  33. package/dist/datasource/skills/clawgallery/client.d.ts +9 -0
  34. package/dist/datasource/skills/clawgallery/index.d.ts +5 -0
  35. package/dist/datasource/skills/clawgallery/methods.d.ts +19 -0
  36. package/dist/datasource/skills/clawgallery/paths.d.ts +7 -0
  37. package/dist/datasource/skills/clawgallery/skill.d.ts +35 -0
  38. package/dist/datasource/skills/clawgallery/types.d.ts +58 -0
  39. package/dist/datasource/skills/cloud-drive/index.d.ts +2 -0
  40. package/dist/datasource/skills/cloud-drive/rclone-connector.d.ts +47 -0
  41. package/dist/datasource/skills/cloud-drive/skill.d.ts +14 -0
  42. package/dist/datasource/skills/discrawl/skill.d.ts +3 -0
  43. package/dist/datasource/skills/discrawl/types.d.ts +4 -8
  44. package/dist/datasource/skills/factory.d.ts +9 -0
  45. package/dist/datasource/skills/gmail/himalaya-connector.d.ts +12 -1
  46. package/dist/datasource/skills/katok/methods.d.ts +6 -4
  47. package/dist/datasource/skills/katok/types.d.ts +8 -26
  48. package/dist/datasource/skills/mailcrawl/client.d.ts +15 -0
  49. package/dist/datasource/skills/mailcrawl/config.d.ts +2 -0
  50. package/dist/datasource/skills/mailcrawl/index.d.ts +4 -0
  51. package/dist/datasource/skills/mailcrawl/methods.d.ts +15 -0
  52. package/dist/datasource/skills/mailcrawl/skill.d.ts +34 -0
  53. package/dist/datasource/skills/mailcrawl/types.d.ts +71 -0
  54. package/dist/datasource/skills/obsidian/skill.d.ts +2 -0
  55. package/dist/datasource/skills/obsidian/types.d.ts +2 -0
  56. package/dist/datasource/types.d.ts +5 -2
  57. package/dist/dupey/cli.d.ts +40 -0
  58. package/dist/dupey/filter.d.ts +8 -0
  59. package/dist/dupey/index.d.ts +2 -0
  60. package/dist/filesystem/cloud-placeholder.d.ts +26 -0
  61. package/dist/index.d.ts +1 -0
  62. package/dist/index.js +8889 -9292
  63. package/dist/jikji/client.d.ts +1 -0
  64. package/dist/memory/check-memory-tool.d.ts +1 -0
  65. package/dist/memory/index.d.ts +2 -2
  66. package/dist/memory/memory.d.ts +33 -3
  67. package/dist/memory/renderer.d.ts +2 -1
  68. package/dist/minsync/client.d.ts +5 -1
  69. package/dist/minsync/embedder-config.d.ts +4 -1
  70. package/dist/minsync/index.d.ts +4 -4
  71. package/dist/minsync/installer.d.ts +1 -0
  72. package/dist/minsync/method.d.ts +22 -1
  73. package/dist/minsync/types.d.ts +5 -0
  74. package/dist/mirror/sync.d.ts +3 -1
  75. package/dist/observability/run-log.d.ts +1 -26
  76. package/dist/parser/defaults.d.ts +2 -0
  77. package/dist/parser/index.d.ts +2 -2
  78. package/dist/parser/opendataloader-pdf.d.ts +10 -0
  79. package/dist/parser/types.d.ts +7 -0
  80. package/dist/process/portable-spawn.d.ts +10 -0
  81. package/dist/retrieval/index.d.ts +2 -2
  82. package/dist/retrieval/methods/bm25.d.ts +7 -36
  83. package/dist/ui/browse.d.ts +11 -0
  84. package/dist/ui/catalog.d.ts +59 -0
  85. package/dist/ui/config-store.d.ts +43 -0
  86. package/dist/ui/discover.d.ts +19 -0
  87. package/dist/ui/html.d.ts +2 -0
  88. package/dist/ui/probe.d.ts +24 -0
  89. package/dist/ui/prompt.d.ts +21 -0
  90. package/dist/ui/server.d.ts +28 -0
  91. package/package.json +17 -15
  92. package/skills/autorag/SKILL.md +79 -108
  93. package/skills/autorag-setup/SKILL.md +170 -395
  94. package/dist/datasource/skills/discrawl/config.d.ts +0 -9
  95. package/dist/datasource/skills/gdrive/connector.d.ts +0 -30
  96. package/dist/datasource/skills/gdrive/index.d.ts +0 -3
  97. package/dist/datasource/skills/gdrive/rclone-connector.d.ts +0 -51
  98. package/dist/datasource/skills/gdrive/skill.d.ts +0 -24
  99. package/dist/subagents/contracts.d.ts +0 -90
  100. package/dist/subagents/dispatch-templates.d.ts +0 -104
  101. package/dist/subagents/dispatch-validation.d.ts +0 -169
  102. package/dist/subagents/explorer-tools-extension.d.ts +0 -4
  103. package/dist/subagents/explorer-tools-extension.js +0 -130
  104. package/dist/subagents/local-models.d.ts +0 -14
  105. package/dist/subagents/model-policy.d.ts +0 -31
  106. package/dist/subagents/runtime.d.ts +0 -52
@@ -0,0 +1,15 @@
1
+ import type { MailcrawlFailure, MailcrawlIndexInfo, MailcrawlOk, MailcrawlOptions, MailcrawlSearchMode, MailcrawlSearchOptions, MailcrawlSearchResult, MailcrawlSyncResult } from "./types.js";
2
+ export interface MailcrawlSearchClient {
3
+ search(mode: MailcrawlSearchMode, query: string, options?: MailcrawlSearchOptions): Promise<MailcrawlSearchResult>;
4
+ }
5
+ export interface MailcrawlIndexClient extends MailcrawlSearchClient {
6
+ index(signal?: AbortSignal): Promise<MailcrawlOk<MailcrawlIndexInfo> | MailcrawlFailure>;
7
+ }
8
+ export declare class MailcrawlClient implements MailcrawlSearchClient {
9
+ private readonly options;
10
+ constructor(options?: MailcrawlOptions);
11
+ sync(signal?: AbortSignal): Promise<MailcrawlSyncResult>;
12
+ search(mode: MailcrawlSearchMode, query: string, options?: MailcrawlSearchOptions): Promise<MailcrawlSearchResult>;
13
+ private run;
14
+ index(signal?: AbortSignal): Promise<MailcrawlOk<MailcrawlIndexInfo> | MailcrawlFailure>;
15
+ }
@@ -0,0 +1,2 @@
1
+ export declare function validateMailcrawlInstanceId(instanceId: string): void;
2
+ export declare function ensurePrivateMailcrawlDataDir(path: string): void;
@@ -0,0 +1,4 @@
1
+ export { MailcrawlClient, type MailcrawlIndexClient, type MailcrawlSearchClient } from "./client.js";
2
+ export { MAILCRAWL_MODES, MailcrawlMethod } from "./methods.js";
3
+ export { MailcrawlSkill, type MailcrawlSkillClient, type MailcrawlSkillOptions } from "./skill.js";
4
+ export type { MailcrawlFailure, MailcrawlFailureReason, MailcrawlIndexInfo, MailcrawlOptions, MailcrawlSearchHit, MailcrawlSearchMode, MailcrawlSearchOptions, MailcrawlSearchResult, MailcrawlSyncInfo, MailcrawlSyncResult, } from "./types.js";
@@ -0,0 +1,15 @@
1
+ import type { RetrievalMethod, RetrievalMethodDescriptor, RetrievalOptions, RetrievalResult } from "../../../retrieval/types.js";
2
+ import type { MailcrawlSearchClient } from "./client.js";
3
+ import type { MailcrawlSearchMode } from "./types.js";
4
+ export declare class MailcrawlMethod implements RetrievalMethod {
5
+ private readonly client;
6
+ private readonly mode;
7
+ private readonly instanceId;
8
+ private readonly tags;
9
+ private readonly account;
10
+ private readonly mailbox;
11
+ constructor(client: MailcrawlSearchClient, mode: MailcrawlSearchMode, instanceId: string, tags: readonly string[], account?: string, mailbox?: string);
12
+ describe(): RetrievalMethodDescriptor;
13
+ retrieve(query: string, options: RetrievalOptions): Promise<RetrievalResult[]>;
14
+ }
15
+ export declare const MAILCRAWL_MODES: readonly MailcrawlSearchMode[];
@@ -0,0 +1,34 @@
1
+ import type { RetrievalMethod } from "../../../retrieval/types.js";
2
+ import type { DatasourceIndexResult, DatasourceSkill, DatasourceSkillDescriptor, DatasourceSkillManifest, PollingMetadata, SourceDescription } from "../../types.js";
3
+ import { type MailcrawlIndexClient } from "./client.js";
4
+ import type { MailcrawlOptions, MailcrawlSyncResult } from "./types.js";
5
+ export interface MailcrawlSkillClient extends MailcrawlIndexClient {
6
+ sync(signal?: AbortSignal): Promise<MailcrawlSyncResult>;
7
+ }
8
+ export interface MailcrawlSkillOptions extends MailcrawlOptions {
9
+ readonly client?: MailcrawlSkillClient;
10
+ readonly datasourceId?: string;
11
+ readonly instanceId?: string;
12
+ readonly instances?: readonly string[];
13
+ readonly pollingIntervalMs?: number;
14
+ readonly tags?: readonly string[];
15
+ readonly connectorOptions?: MailcrawlOptions;
16
+ }
17
+ export declare class MailcrawlSkill implements DatasourceSkill {
18
+ private readonly client;
19
+ private readonly instanceId;
20
+ private readonly instances;
21
+ private readonly interval;
22
+ private readonly tags;
23
+ private readonly account;
24
+ private readonly mailbox;
25
+ private lastIndexedAt;
26
+ constructor(options?: MailcrawlSkillOptions);
27
+ describe(): DatasourceSkillDescriptor;
28
+ polling(): PollingMetadata;
29
+ index(): Promise<DatasourceIndexResult>;
30
+ retrievalMethods(): readonly RetrievalMethod[];
31
+ describeSources(): readonly SourceDescription[];
32
+ skillManifest(): DatasourceSkillManifest;
33
+ private fail;
34
+ }
@@ -0,0 +1,71 @@
1
+ export type MailcrawlSearchMode = "keyword" | "bm25" | "semantic" | "hybrid";
2
+ export interface MailcrawlSearchOptions {
3
+ readonly topK?: number;
4
+ readonly signal?: AbortSignal;
5
+ }
6
+ export interface MailcrawlOptions {
7
+ readonly binaryPath?: string;
8
+ readonly instanceId?: string;
9
+ readonly dataDir?: string;
10
+ readonly account?: string;
11
+ readonly mailbox?: string;
12
+ readonly backend?: string;
13
+ readonly source?: string;
14
+ readonly fixture?: string;
15
+ readonly himalayaConfig?: string;
16
+ readonly timeoutMs?: number;
17
+ readonly maxBufferBytes?: number;
18
+ readonly env?: Readonly<Record<string, string | undefined>>;
19
+ }
20
+ export type MailcrawlFailureReason = "binary-missing" | "not-configured" | "nonzero-exit" | "spawn-error" | "timeout" | "aborted" | "stdout-too-large" | "stderr-too-large" | "invalid-output" | "remote-embedding-rejected";
21
+ export interface MailcrawlFailure {
22
+ readonly ok: false;
23
+ readonly reason: MailcrawlFailureReason;
24
+ readonly stdout: string;
25
+ readonly stderr: string;
26
+ readonly code: number | null;
27
+ }
28
+ export interface MailcrawlSyncInfo {
29
+ readonly added?: number;
30
+ readonly updated?: number;
31
+ readonly deleted?: number;
32
+ readonly unchanged?: number;
33
+ readonly chunksAdded?: number;
34
+ readonly archiveRevision?: string;
35
+ readonly messages: number;
36
+ }
37
+ export interface MailcrawlIndexInfo {
38
+ readonly embedded?: number;
39
+ readonly reused?: number;
40
+ readonly generation?: string;
41
+ }
42
+ export interface MailcrawlSearchHit {
43
+ readonly chunkId: string;
44
+ readonly messageId: string;
45
+ readonly threadId: string;
46
+ readonly accountId: string;
47
+ readonly mailbox: string;
48
+ readonly subject: string;
49
+ readonly from: string;
50
+ readonly to: readonly string[];
51
+ readonly date: string;
52
+ readonly snippet: string;
53
+ readonly score: number;
54
+ readonly mode: MailcrawlSearchMode;
55
+ }
56
+ export interface MailcrawlOk<T> {
57
+ readonly ok: true;
58
+ readonly data: T;
59
+ readonly stdout: string;
60
+ readonly stderr: string;
61
+ readonly code: number;
62
+ }
63
+ export interface MailcrawlSearchOk {
64
+ readonly ok: true;
65
+ readonly hits: readonly MailcrawlSearchHit[];
66
+ readonly stdout: string;
67
+ readonly stderr: string;
68
+ readonly code: number;
69
+ }
70
+ export type MailcrawlSyncResult = MailcrawlOk<MailcrawlSyncInfo> | MailcrawlFailure;
71
+ export type MailcrawlSearchResult = MailcrawlSearchOk | MailcrawlFailure;
@@ -8,6 +8,7 @@ export interface ObsidianSkillClient extends ObsidianSearchClient {
8
8
  embed(): Promise<QmdEmbedResult>;
9
9
  }
10
10
  export interface ObsidianSkillOptions {
11
+ readonly datasourceId?: string;
11
12
  readonly client?: ObsidianSkillClient;
12
13
  readonly vaultPath?: string;
13
14
  readonly binaryPath?: string;
@@ -19,6 +20,7 @@ export interface ObsidianSkillOptions {
19
20
  readonly workspaceRoot?: string;
20
21
  readonly collectionName?: string;
21
22
  readonly timeoutMs?: number;
23
+ readonly configPath?: string;
22
24
  readonly connectorOptions?: {
23
25
  readonly vaultPath?: string;
24
26
  };
@@ -27,6 +27,8 @@ export interface QmdOptions {
27
27
  readonly collectionName?: string;
28
28
  /** Environment overrides merged on top of a restricted process env. */
29
29
  readonly env?: Readonly<Record<string, string | undefined>>;
30
+ /** Explicit operator-owned qmd config directory, never rewritten. */
31
+ readonly configPath?: string;
30
32
  }
31
33
  export declare const DEFAULT_QMD_BINARY = "qmd";
32
34
  export declare const DEFAULT_QMD_TIMEOUT_MS = 60000;
@@ -14,8 +14,9 @@
14
14
  * `undefined`-as-deny.
15
15
  * - Model/tool arguments never grant access. Only the trusted server-supplied
16
16
  * {@link DatasourceAccessContext} can authorize a datasource.
17
- * - Datasource sources are slash-hierarchical opaque paths such as
18
- * `/kakao/<instance-id>/chunks/<chunk-id>`; `#` fragments are rejected.
17
+ * - Datasource sources are slash-hierarchical opaque paths when the skill
18
+ * supports the `scoped` capability; skills without that capability may
19
+ * expose another opaque identity format.
19
20
  */
20
21
  import type { RetrievalMethod } from "../retrieval/types.js";
21
22
  export type { RetrievalMethod } from "../retrieval/types.js";
@@ -62,6 +63,8 @@ export interface DatasourceSkillDescriptor {
62
63
  export interface DatasourceAccessible {
63
64
  readonly datasourceId?: string;
64
65
  readonly tags?: readonly string[];
66
+ /** Capability names such as `scoped`; absent means no source-scope support. */
67
+ readonly capabilities?: readonly string[];
65
68
  }
66
69
  /**
67
70
  * A pluggable datasource skill. Skills are server-bound: they are constructed
@@ -0,0 +1,40 @@
1
+ export interface DupeyScanFile {
2
+ readonly path: string;
3
+ readonly content_hash?: string;
4
+ readonly [key: string]: unknown;
5
+ }
6
+ export interface DupeyScanEdge {
7
+ readonly relation: string;
8
+ readonly a: string;
9
+ readonly b: string;
10
+ readonly [key: string]: unknown;
11
+ }
12
+ export interface DupeyScanFamily {
13
+ readonly id: number;
14
+ readonly relation: string;
15
+ readonly files: readonly string[];
16
+ readonly members: readonly Record<string, unknown>[];
17
+ readonly edges: readonly DupeyScanEdge[];
18
+ readonly pick?: Record<string, unknown>;
19
+ readonly [key: string]: unknown;
20
+ }
21
+ export interface DupeyScanResult {
22
+ readonly dir: string;
23
+ readonly threshold?: number;
24
+ readonly contains_threshold?: number;
25
+ readonly contains_min_jaccard?: number;
26
+ readonly files: readonly DupeyScanFile[];
27
+ readonly families: readonly DupeyScanFamily[];
28
+ readonly errors: readonly Record<string, unknown>[];
29
+ readonly [key: string]: unknown;
30
+ }
31
+ export interface DupeyCliOptions {
32
+ readonly executable?: string;
33
+ readonly cwd?: string;
34
+ readonly timeoutMs?: number;
35
+ readonly run?: (args: readonly string[]) => Promise<string>;
36
+ }
37
+ export declare class DupeyCliError extends Error {
38
+ constructor(message: string);
39
+ }
40
+ export declare function scanWithDupey(dir: string, options?: DupeyCliOptions): Promise<DupeyScanResult>;
@@ -0,0 +1,8 @@
1
+ import type { DupeyScanResult } from "./cli.js";
2
+ export interface ExactDuplicateFilterResult {
3
+ readonly excluded: ReadonlySet<string>;
4
+ readonly keepers: ReadonlySet<string>;
5
+ readonly errors: readonly string[];
6
+ }
7
+ /** Selects the newest filesystem copy for each exact canonical-text hash. */
8
+ export declare function selectExactDuplicateExclusions(root: string, scan: DupeyScanResult): Promise<ExactDuplicateFilterResult>;
@@ -0,0 +1,2 @@
1
+ export * from "./cli.js";
2
+ export * from "./filter.js";
@@ -0,0 +1,26 @@
1
+ /** Darwin st_flags bit used by iCloud/File Provider data-less placeholders. */
2
+ export declare const UF_DATALESS = 32768;
3
+ export declare const FILE_PROVIDER_XATTR = "com.apple.file-provider-domain-id";
4
+ /** Windows FILE_ATTRIBUTE_* bits commonly used by Cloud Files placeholders. */
5
+ export declare const WINDOWS_FILE_ATTRIBUTE_OFFLINE = 4096;
6
+ export declare const WINDOWS_FILE_ATTRIBUTE_RECALL_ON_OPEN = 262144;
7
+ export declare const WINDOWS_FILE_ATTRIBUTE_RECALL_ON_DATA_ACCESS = 4194304;
8
+ export type FilesystemKind = "local" | "file-provider";
9
+ export interface FilesystemClassification {
10
+ readonly kind: FilesystemKind;
11
+ readonly provider?: string;
12
+ readonly reason: "path-marker" | "file-provider-xattr" | "cloud-volume" | "placeholder-attribute" | "local";
13
+ }
14
+ export declare function pathLooksLikeCloudRoot(target: string): boolean;
15
+ export declare function homeCloudRoots(home?: string): readonly string[];
16
+ export declare function classifyFilesystemRoot(target: string): Promise<FilesystemClassification>;
17
+ export declare function isDatalessPlaceholder(target: string): Promise<boolean>;
18
+ export interface MaterializedWalk {
19
+ readonly materialized: readonly string[];
20
+ readonly skippedDataless: number;
21
+ }
22
+ export declare function listMaterializedFiles(root: string, options?: {
23
+ readonly limit?: number;
24
+ readonly timeoutMs?: number;
25
+ }): Promise<MaterializedWalk>;
26
+ export declare function relativeCloudHint(root: string, filePath: string): string;
package/dist/index.d.ts CHANGED
@@ -2,6 +2,7 @@ export { AutoRAGAgent, type AutoRAGAgentOptions, type AutoRAGRefreshComponentSta
2
2
  export { type AutoRAGResultsDetails, createEmitResultsTool, EMIT_AUTORAG_RESULTS_TOOL_NAME, type SearchDocumentDiagnostic, type SearchDocumentDiagnosticCode, type SearchDocumentDiagnosticSeverity, type SearchDocumentsResponse, type SearchDocumentWarning, } from "./agent/index.js";
3
3
  export { buildSystemPrompt, type SystemPromptConfig } from "./agent/system-prompt.js";
4
4
  export * from "./datasource/index.js";
5
+ export * from "./dupey/index.js";
5
6
  export * from "./jikji/index.js";
6
7
  export * from "./manifest/index.js";
7
8
  export * from "./memory/index.js";