@lobu/connector-sdk 6.0.1

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 (93) hide show
  1. package/dist/api-paginated.d.ts +79 -0
  2. package/dist/api-paginated.d.ts.map +1 -0
  3. package/dist/api-paginated.js +120 -0
  4. package/dist/api-paginated.js.map +1 -0
  5. package/dist/base.d.ts +66 -0
  6. package/dist/base.d.ts.map +1 -0
  7. package/dist/base.js +122 -0
  8. package/dist/base.js.map +1 -0
  9. package/dist/browser/acquire.d.ts +66 -0
  10. package/dist/browser/acquire.d.ts.map +1 -0
  11. package/dist/browser/acquire.js +109 -0
  12. package/dist/browser/acquire.js.map +1 -0
  13. package/dist/browser/cdp-page.d.ts +48 -0
  14. package/dist/browser/cdp-page.d.ts.map +1 -0
  15. package/dist/browser/cdp-page.js +165 -0
  16. package/dist/browser/cdp-page.js.map +1 -0
  17. package/dist/browser/cdp.d.ts +44 -0
  18. package/dist/browser/cdp.d.ts.map +1 -0
  19. package/dist/browser/cdp.js +252 -0
  20. package/dist/browser/cdp.js.map +1 -0
  21. package/dist/browser/launcher.d.ts +29 -0
  22. package/dist/browser/launcher.d.ts.map +1 -0
  23. package/dist/browser/launcher.js +157 -0
  24. package/dist/browser/launcher.js.map +1 -0
  25. package/dist/browser/stealth.d.ts +55 -0
  26. package/dist/browser/stealth.d.ts.map +1 -0
  27. package/dist/browser/stealth.js +170 -0
  28. package/dist/browser/stealth.js.map +1 -0
  29. package/dist/browser-network.d.ts +51 -0
  30. package/dist/browser-network.d.ts.map +1 -0
  31. package/dist/browser-network.js +175 -0
  32. package/dist/browser-network.js.map +1 -0
  33. package/dist/browser-paginated.d.ts +141 -0
  34. package/dist/browser-paginated.d.ts.map +1 -0
  35. package/dist/browser-paginated.js +269 -0
  36. package/dist/browser-paginated.js.map +1 -0
  37. package/dist/connector-runtime.d.ts +70 -0
  38. package/dist/connector-runtime.d.ts.map +1 -0
  39. package/dist/connector-runtime.js +48 -0
  40. package/dist/connector-runtime.js.map +1 -0
  41. package/dist/connector-types.d.ts +613 -0
  42. package/dist/connector-types.d.ts.map +1 -0
  43. package/dist/connector-types.js +27 -0
  44. package/dist/connector-types.js.map +1 -0
  45. package/dist/event-taxonomy.d.ts +3 -0
  46. package/dist/event-taxonomy.d.ts.map +1 -0
  47. package/dist/event-taxonomy.js +30 -0
  48. package/dist/event-taxonomy.js.map +1 -0
  49. package/dist/http.d.ts +18 -0
  50. package/dist/http.d.ts.map +1 -0
  51. package/dist/http.js +74 -0
  52. package/dist/http.js.map +1 -0
  53. package/dist/identity-normalize.d.ts +53 -0
  54. package/dist/identity-normalize.d.ts.map +1 -0
  55. package/dist/identity-normalize.js +146 -0
  56. package/dist/identity-normalize.js.map +1 -0
  57. package/dist/identity-types.d.ts +214 -0
  58. package/dist/identity-types.d.ts.map +1 -0
  59. package/dist/identity-types.js +217 -0
  60. package/dist/identity-types.js.map +1 -0
  61. package/dist/index.d.ts +37 -0
  62. package/dist/index.d.ts.map +1 -0
  63. package/dist/index.js +33 -0
  64. package/dist/index.js.map +1 -0
  65. package/dist/logger.d.ts +7 -0
  66. package/dist/logger.d.ts.map +1 -0
  67. package/dist/logger.js +10 -0
  68. package/dist/logger.js.map +1 -0
  69. package/dist/paginated.d.ts +93 -0
  70. package/dist/paginated.d.ts.map +1 -0
  71. package/dist/paginated.js +167 -0
  72. package/dist/paginated.js.map +1 -0
  73. package/dist/reaction-sdk.d.ts +43 -0
  74. package/dist/reaction-sdk.d.ts.map +1 -0
  75. package/dist/reaction-sdk.js +9 -0
  76. package/dist/reaction-sdk.js.map +1 -0
  77. package/dist/retry.d.ts +19 -0
  78. package/dist/retry.d.ts.map +1 -0
  79. package/dist/retry.js +131 -0
  80. package/dist/retry.js.map +1 -0
  81. package/dist/scoring.d.ts +17 -0
  82. package/dist/scoring.d.ts.map +1 -0
  83. package/dist/scoring.js +28 -0
  84. package/dist/scoring.js.map +1 -0
  85. package/dist/types.d.ts +280 -0
  86. package/dist/types.d.ts.map +1 -0
  87. package/dist/types.js +2 -0
  88. package/dist/types.js.map +1 -0
  89. package/dist/watcher-time.d.ts +14 -0
  90. package/dist/watcher-time.d.ts.map +1 -0
  91. package/dist/watcher-time.js +112 -0
  92. package/dist/watcher-time.js.map +1 -0
  93. package/package.json +59 -0
@@ -0,0 +1,79 @@
1
+ /**
2
+ * API Paginated Feed Base Class
3
+ *
4
+ * Extends PaginatedFeed for HTTP/REST API-based feeds.
5
+ * Provides HTTP client setup and common API patterns.
6
+ */
7
+ import type { KyInstance } from 'ky';
8
+ import { HTTPError } from 'ky';
9
+ import type { PageFetchResult, PaginatedCheckpoint } from './paginated.js';
10
+ import { PaginatedFeed } from './paginated.js';
11
+ import type { Env, FeedOptions, SessionState } from './types.js';
12
+ /**
13
+ * API session state for OAuth/token-based feeds
14
+ */
15
+ export interface ApiSessionState extends SessionState {
16
+ /** OAuth/API access token */
17
+ access_token?: string;
18
+ /** OAuth refresh token (for token refresh flows) */
19
+ refresh_token?: string;
20
+ /** Token type (e.g., 'Bearer') */
21
+ token_type?: string;
22
+ /** Token expiration time (ISO string) */
23
+ expires_at?: string;
24
+ /** Additional headers to include in requests */
25
+ headers?: Record<string, string>;
26
+ /** API key (alternative to OAuth tokens) */
27
+ api_key?: string;
28
+ }
29
+ /**
30
+ * Base class for API-based feeds with HTTP pagination
31
+ */
32
+ export declare abstract class ApiPaginatedFeed<TItem, TResponse = unknown, TCheckpoint extends PaginatedCheckpoint = PaginatedCheckpoint> extends PaginatedFeed<TItem, TCheckpoint> {
33
+ readonly apiType: "api";
34
+ /**
35
+ * Session state for this sync session
36
+ */
37
+ protected _sessionState: ApiSessionState | null;
38
+ /**
39
+ * Set session state for this sync session
40
+ */
41
+ protected setSessionState(sessionState: SessionState | null | undefined): void;
42
+ /**
43
+ * Get current session state
44
+ */
45
+ protected getSessionState(): ApiSessionState | null;
46
+ /**
47
+ * ABSTRACT: Build URL for fetching a specific page
48
+ */
49
+ protected abstract buildPageUrl(cursor: string | null, options: FeedOptions): string;
50
+ /**
51
+ * ABSTRACT: Parse API response into items and next token
52
+ */
53
+ protected abstract parseResponse(response: TResponse, options: FeedOptions): PageFetchResult<TItem>;
54
+ /**
55
+ * Get configured HTTP client for this feed
56
+ */
57
+ protected getHttpClient(_env: Env): KyInstance;
58
+ /**
59
+ * Create HTTP client with Bearer token authentication
60
+ */
61
+ protected createBearerClient(token: string, additionalHeaders?: Record<string, string>): KyInstance;
62
+ /**
63
+ * Create HTTP client with custom headers
64
+ */
65
+ protected createClientWithHeaders(headers: Record<string, string>): KyInstance;
66
+ /**
67
+ * Create HTTP client from session state
68
+ */
69
+ protected createClientFromSessionState(additionalHeaders?: Record<string, string>): KyInstance;
70
+ /**
71
+ * Handle HTTP errors with platform-specific messages
72
+ */
73
+ protected handleHttpError(error: HTTPError, url: string): never;
74
+ /**
75
+ * Default fetchPage implementation using HTTP client with retry
76
+ */
77
+ protected fetchPage(cursor: string | null, options: FeedOptions, env: Env): Promise<PageFetchResult<TItem>>;
78
+ }
79
+ //# sourceMappingURL=api-paginated.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"api-paginated.d.ts","sourceRoot":"","sources":["../src/api-paginated.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,IAAI,CAAC;AACrC,OAAO,EAAE,SAAS,EAAE,MAAM,IAAI,CAAC;AAI/B,OAAO,KAAK,EAAE,eAAe,EAAE,mBAAmB,EAAE,MAAM,gBAAgB,CAAC;AAC3E,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,KAAK,EAAE,GAAG,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAEjE;;GAEG;AACH,MAAM,WAAW,eAAgB,SAAQ,YAAY;IACnD,6BAA6B;IAC7B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,oDAAoD;IACpD,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,kCAAkC;IAClC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,yCAAyC;IACzC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,gDAAgD;IAChD,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,4CAA4C;IAC5C,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;GAEG;AACH,8BAAsB,gBAAgB,CACpC,KAAK,EACL,SAAS,GAAG,OAAO,EACnB,WAAW,SAAS,mBAAmB,GAAG,mBAAmB,CAC7D,SAAQ,aAAa,CAAC,KAAK,EAAE,WAAW,CAAC;IACzC,QAAQ,CAAC,OAAO,EAAG,KAAK,CAAU;IAElC;;OAEG;IACH,SAAS,CAAC,aAAa,EAAE,eAAe,GAAG,IAAI,CAAQ;IAEvD;;OAEG;IACH,SAAS,CAAC,eAAe,CAAC,YAAY,EAAE,YAAY,GAAG,IAAI,GAAG,SAAS,GAAG,IAAI;IAU9E;;OAEG;IACH,SAAS,CAAC,eAAe,IAAI,eAAe,GAAG,IAAI;IAInD;;OAEG;IACH,SAAS,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,EAAE,MAAM,GAAG,IAAI,EAAE,OAAO,EAAE,WAAW,GAAG,MAAM;IAEpF;;OAEG;IACH,SAAS,CAAC,QAAQ,CAAC,aAAa,CAC9B,QAAQ,EAAE,SAAS,EACnB,OAAO,EAAE,WAAW,GACnB,eAAe,CAAC,KAAK,CAAC;IAEzB;;OAEG;IACH,SAAS,CAAC,aAAa,CAAC,IAAI,EAAE,GAAG,GAAG,UAAU;IAI9C;;OAEG;IACH,SAAS,CAAC,kBAAkB,CAC1B,KAAK,EAAE,MAAM,EACb,iBAAiB,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,GACzC,UAAU;IAIb;;OAEG;IACH,SAAS,CAAC,uBAAuB,CAAC,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,UAAU;IAI9E;;OAEG;IACH,SAAS,CAAC,4BAA4B,CAAC,iBAAiB,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,UAAU;IAsB9F;;OAEG;IACH,SAAS,CAAC,eAAe,CAAC,KAAK,EAAE,SAAS,EAAE,GAAG,EAAE,MAAM,GAAG,KAAK;IAI/D;;OAEG;cACa,SAAS,CACvB,MAAM,EAAE,MAAM,GAAG,IAAI,EACrB,OAAO,EAAE,WAAW,EACpB,GAAG,EAAE,GAAG,GACP,OAAO,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC;CAsCnC"}
@@ -0,0 +1,120 @@
1
+ /**
2
+ * API Paginated Feed Base Class
3
+ *
4
+ * Extends PaginatedFeed for HTTP/REST API-based feeds.
5
+ * Provides HTTP client setup and common API patterns.
6
+ */
7
+ import { HTTPError } from 'ky';
8
+ import { RateLimitError } from './base.js';
9
+ import { createAuthenticatedClient, createHttpClient, httpClient } from './http.js';
10
+ import { sdkLogger } from './logger.js';
11
+ import { PaginatedFeed } from './paginated.js';
12
+ import { withHttpRetry } from './retry.js';
13
+ /**
14
+ * Base class for API-based feeds with HTTP pagination
15
+ */
16
+ export class ApiPaginatedFeed extends PaginatedFeed {
17
+ apiType = 'api';
18
+ /**
19
+ * Session state for this sync session
20
+ */
21
+ _sessionState = null;
22
+ /**
23
+ * Set session state for this sync session
24
+ */
25
+ setSessionState(sessionState) {
26
+ this._sessionState = sessionState || null;
27
+ if (this._sessionState) {
28
+ sdkLogger.debug({ hasToken: !!this._sessionState.access_token, hasApiKey: !!this._sessionState.api_key }, `[${this.type}] Session state set`);
29
+ }
30
+ }
31
+ /**
32
+ * Get current session state
33
+ */
34
+ getSessionState() {
35
+ return this._sessionState;
36
+ }
37
+ /**
38
+ * Get configured HTTP client for this feed
39
+ */
40
+ getHttpClient(_env) {
41
+ return httpClient;
42
+ }
43
+ /**
44
+ * Create HTTP client with Bearer token authentication
45
+ */
46
+ createBearerClient(token, additionalHeaders) {
47
+ return createAuthenticatedClient(`Bearer ${token}`, additionalHeaders);
48
+ }
49
+ /**
50
+ * Create HTTP client with custom headers
51
+ */
52
+ createClientWithHeaders(headers) {
53
+ return createHttpClient({ headers });
54
+ }
55
+ /**
56
+ * Create HTTP client from session state
57
+ */
58
+ createClientFromSessionState(additionalHeaders) {
59
+ if (!this._sessionState) {
60
+ return additionalHeaders ? createHttpClient({ headers: additionalHeaders }) : httpClient;
61
+ }
62
+ const headers = {
63
+ ...this._sessionState.headers,
64
+ ...additionalHeaders,
65
+ };
66
+ if (this._sessionState.access_token) {
67
+ const tokenType = this._sessionState.token_type || 'Bearer';
68
+ return createAuthenticatedClient(`${tokenType} ${this._sessionState.access_token}`, headers);
69
+ }
70
+ if (this._sessionState.api_key) {
71
+ return createAuthenticatedClient(this._sessionState.api_key, headers);
72
+ }
73
+ return Object.keys(headers).length > 0 ? createHttpClient({ headers }) : httpClient;
74
+ }
75
+ /**
76
+ * Handle HTTP errors with platform-specific messages
77
+ */
78
+ handleHttpError(error, url) {
79
+ this.handleHTTPError(error.response.status, url);
80
+ }
81
+ /**
82
+ * Default fetchPage implementation using HTTP client with retry
83
+ */
84
+ async fetchPage(cursor, options, env) {
85
+ const client = this.getHttpClient(env);
86
+ const url = this.buildPageUrl(cursor, options);
87
+ try {
88
+ const response = await withHttpRetry(async () => client.get(url).json(), {
89
+ operation: `${this.type} API fetch`,
90
+ context: { url, cursor },
91
+ });
92
+ return this.parseResponse(response, options);
93
+ }
94
+ catch (error) {
95
+ if (error instanceof HTTPError) {
96
+ if (error.response.status === 429) {
97
+ const retryAfter = error.response.headers.get('retry-after');
98
+ let retryAfterMs;
99
+ if (retryAfter) {
100
+ const numericRetry = Number(retryAfter);
101
+ if (!Number.isNaN(numericRetry)) {
102
+ retryAfterMs = numericRetry * 1000;
103
+ }
104
+ else {
105
+ const retryDate = Date.parse(retryAfter);
106
+ if (!Number.isNaN(retryDate)) {
107
+ retryAfterMs = retryDate - Date.now();
108
+ }
109
+ }
110
+ }
111
+ throw new RateLimitError(`${this.displayName} rate limit exceeded. Please wait before retrying.`, retryAfterMs && retryAfterMs > 0 ? retryAfterMs : undefined);
112
+ }
113
+ this.handleHttpError(error, url);
114
+ }
115
+ sdkLogger.error({ error, url }, `[${this.type}] API fetch failed`);
116
+ throw error;
117
+ }
118
+ }
119
+ }
120
+ //# sourceMappingURL=api-paginated.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"api-paginated.js","sourceRoot":"","sources":["../src/api-paginated.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAGH,OAAO,EAAE,SAAS,EAAE,MAAM,IAAI,CAAC;AAC/B,OAAO,EAAE,cAAc,EAAE,MAAM,WAAW,CAAC;AAC3C,OAAO,EAAE,yBAAyB,EAAE,gBAAgB,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACpF,OAAO,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAExC,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAqB3C;;GAEG;AACH,MAAM,OAAgB,gBAIpB,SAAQ,aAAiC;IAChC,OAAO,GAAG,KAAc,CAAC;IAElC;;OAEG;IACO,aAAa,GAA2B,IAAI,CAAC;IAEvD;;OAEG;IACO,eAAe,CAAC,YAA6C;QACrE,IAAI,CAAC,aAAa,GAAI,YAAgC,IAAI,IAAI,CAAC;QAC/D,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;YACvB,SAAS,CAAC,KAAK,CACb,EAAE,QAAQ,EAAE,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,YAAY,EAAE,SAAS,EAAE,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,OAAO,EAAE,EACxF,IAAI,IAAI,CAAC,IAAI,qBAAqB,CACnC,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACO,eAAe;QACvB,OAAO,IAAI,CAAC,aAAa,CAAC;IAC5B,CAAC;IAeD;;OAEG;IACO,aAAa,CAAC,IAAS;QAC/B,OAAO,UAAU,CAAC;IACpB,CAAC;IAED;;OAEG;IACO,kBAAkB,CAC1B,KAAa,EACb,iBAA0C;QAE1C,OAAO,yBAAyB,CAAC,UAAU,KAAK,EAAE,EAAE,iBAAiB,CAAC,CAAC;IACzE,CAAC;IAED;;OAEG;IACO,uBAAuB,CAAC,OAA+B;QAC/D,OAAO,gBAAgB,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC;IACvC,CAAC;IAED;;OAEG;IACO,4BAA4B,CAAC,iBAA0C;QAC/E,IAAI,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC;YACxB,OAAO,iBAAiB,CAAC,CAAC,CAAC,gBAAgB,CAAC,EAAE,OAAO,EAAE,iBAAiB,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;QAC3F,CAAC;QAED,MAAM,OAAO,GAA2B;YACtC,GAAG,IAAI,CAAC,aAAa,CAAC,OAAO;YAC7B,GAAG,iBAAiB;SACrB,CAAC;QAEF,IAAI,IAAI,CAAC,aAAa,CAAC,YAAY,EAAE,CAAC;YACpC,MAAM,SAAS,GAAG,IAAI,CAAC,aAAa,CAAC,UAAU,IAAI,QAAQ,CAAC;YAC5D,OAAO,yBAAyB,CAAC,GAAG,SAAS,IAAI,IAAI,CAAC,aAAa,CAAC,YAAY,EAAE,EAAE,OAAO,CAAC,CAAC;QAC/F,CAAC;QAED,IAAI,IAAI,CAAC,aAAa,CAAC,OAAO,EAAE,CAAC;YAC/B,OAAO,yBAAyB,CAAC,IAAI,CAAC,aAAa,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QACxE,CAAC;QAED,OAAO,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;IACtF,CAAC;IAED;;OAEG;IACO,eAAe,CAAC,KAAgB,EAAE,GAAW;QACrD,IAAI,CAAC,eAAe,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IACnD,CAAC;IAED;;OAEG;IACO,KAAK,CAAC,SAAS,CACvB,MAAqB,EACrB,OAAoB,EACpB,GAAQ;QAER,MAAM,MAAM,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC;QACvC,MAAM,GAAG,GAAG,IAAI,CAAC,YAAY,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;QAE/C,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,aAAa,CAAC,KAAK,IAAI,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,EAAa,EAAE;gBAClF,SAAS,EAAE,GAAG,IAAI,CAAC,IAAI,YAAY;gBACnC,OAAO,EAAE,EAAE,GAAG,EAAE,MAAM,EAAE;aACzB,CAAC,CAAC;YAEH,OAAO,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAC/C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,KAAK,YAAY,SAAS,EAAE,CAAC;gBAC/B,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;oBAClC,MAAM,UAAU,GAAG,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;oBAC7D,IAAI,YAAgC,CAAC;oBACrC,IAAI,UAAU,EAAE,CAAC;wBACf,MAAM,YAAY,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC;wBACxC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC;4BAChC,YAAY,GAAG,YAAY,GAAG,IAAI,CAAC;wBACrC,CAAC;6BAAM,CAAC;4BACN,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;4BACzC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;gCAC7B,YAAY,GAAG,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;4BACxC,CAAC;wBACH,CAAC;oBACH,CAAC;oBACD,MAAM,IAAI,cAAc,CACtB,GAAG,IAAI,CAAC,WAAW,oDAAoD,EACvE,YAAY,IAAI,YAAY,GAAG,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,SAAS,CAC5D,CAAC;gBACJ,CAAC;gBACD,IAAI,CAAC,eAAe,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;YACnC,CAAC;YACD,SAAS,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,IAAI,CAAC,IAAI,oBAAoB,CAAC,CAAC;YACnE,MAAM,KAAK,CAAC;QACd,CAAC;IACH,CAAC;CACF"}
package/dist/base.d.ts ADDED
@@ -0,0 +1,66 @@
1
+ import type { TObject } from '@sinclair/typebox';
2
+ import type { Checkpoint, Content, Env, FeedAuthSchema, FeedOptions, FeedSyncResult, IFeed, ParentFeedDefinition, ScoringConfig, SessionState } from './types.js';
3
+ export declare class RateLimitError extends Error {
4
+ readonly retryAfterMs?: number;
5
+ constructor(message: string, retryAfterMs?: number);
6
+ }
7
+ /**
8
+ * Base feed implementation with common functionality
9
+ * All platform-specific feeds should extend this class
10
+ */
11
+ export declare abstract class BaseFeed implements IFeed {
12
+ abstract readonly type: string;
13
+ abstract readonly displayName: string;
14
+ abstract readonly apiType: 'api' | 'browser';
15
+ abstract readonly feedMode: 'entity' | 'search';
16
+ abstract readonly optionsSchema: TObject;
17
+ abstract readonly defaultScoringConfig: ScoringConfig;
18
+ abstract readonly defaultScoringFormula: string;
19
+ readonly authSchema: FeedAuthSchema;
20
+ abstract pull(options: FeedOptions, checkpoint: Checkpoint | null, env: Env, sessionState?: SessionState | null, updateCheckpointFn?: (checkpoint: Checkpoint) => Promise<void>): Promise<FeedSyncResult>;
21
+ abstract urlFromOptions(options: FeedOptions): string;
22
+ abstract displayLabelFromOptions(options: FeedOptions): string;
23
+ abstract validateOptions(options: FeedOptions): string | null;
24
+ getParentFeedDefinitions(_options: FeedOptions): ParentFeedDefinition[];
25
+ /**
26
+ * Validate options using TypeBox schema
27
+ * Subclasses can call this for schema validation before adding custom business logic
28
+ */
29
+ protected validateWithSchema(options: FeedOptions): string | null;
30
+ /**
31
+ * Get rate limit information for this platform
32
+ * Override this method in platform-specific feeds
33
+ * Default is conservative: 10 requests per minute
34
+ */
35
+ getRateLimit(): {
36
+ requests_per_minute: number;
37
+ recommended_interval_ms: number;
38
+ };
39
+ /**
40
+ * Helper to check if content is newer than checkpoint
41
+ */
42
+ protected isNewerThan(contentDate: Date, checkpoint: Checkpoint | null): boolean;
43
+ /**
44
+ * Calculate lookback date from options
45
+ * @param options - Feed options with optional lookback_days
46
+ * @param defaultDays - Default lookback period (default: 365)
47
+ */
48
+ protected getLookbackDate(options: FeedOptions, defaultDays?: number): Date;
49
+ /**
50
+ * Sleep for specified milliseconds (for rate limiting)
51
+ */
52
+ protected sleep(ms: number): Promise<void>;
53
+ /**
54
+ * Check if feed is in incremental mode
55
+ */
56
+ protected isIncrementalMode(checkpoint: Checkpoint | null, paginationToken?: string | null): boolean;
57
+ /**
58
+ * Helper to deduplicate content by origin_id
59
+ */
60
+ protected deduplicate(contents: Content[], seenIds: Set<string>): Content[];
61
+ /**
62
+ * Handle HTTP errors with structured logging and platform-specific messages
63
+ */
64
+ protected handleHTTPError(status: number, context: string, platformName?: string): never;
65
+ }
66
+ //# sourceMappingURL=base.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"base.d.ts","sourceRoot":"","sources":["../src/base.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;AAGjD,OAAO,KAAK,EACV,UAAU,EACV,OAAO,EACP,GAAG,EACH,cAAc,EACd,WAAW,EACX,cAAc,EACd,KAAK,EACL,oBAAoB,EACpB,aAAa,EACb,YAAY,EACb,MAAM,YAAY,CAAC;AAEpB,qBAAa,cAAe,SAAQ,KAAK;IACvC,QAAQ,CAAC,YAAY,CAAC,EAAE,MAAM,CAAC;gBAEnB,OAAO,EAAE,MAAM,EAAE,YAAY,CAAC,EAAE,MAAM;CAKnD;AAED;;;GAGG;AACH,8BAAsB,QAAS,YAAW,KAAK;IAC7C,QAAQ,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAC/B,QAAQ,CAAC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IACtC,QAAQ,CAAC,QAAQ,CAAC,OAAO,EAAE,KAAK,GAAG,SAAS,CAAC;IAC7C,QAAQ,CAAC,QAAQ,CAAC,QAAQ,EAAE,QAAQ,GAAG,QAAQ,CAAC;IAChD,QAAQ,CAAC,QAAQ,CAAC,aAAa,EAAE,OAAO,CAAC;IACzC,QAAQ,CAAC,QAAQ,CAAC,oBAAoB,EAAE,aAAa,CAAC;IACtD,QAAQ,CAAC,QAAQ,CAAC,qBAAqB,EAAE,MAAM,CAAC;IAEhD,QAAQ,CAAC,UAAU,EAAE,cAAc,CAAmC;IAEtE,QAAQ,CAAC,IAAI,CACX,OAAO,EAAE,WAAW,EACpB,UAAU,EAAE,UAAU,GAAG,IAAI,EAC7B,GAAG,EAAE,GAAG,EACR,YAAY,CAAC,EAAE,YAAY,GAAG,IAAI,EAClC,kBAAkB,CAAC,EAAE,CAAC,UAAU,EAAE,UAAU,KAAK,OAAO,CAAC,IAAI,CAAC,GAC7D,OAAO,CAAC,cAAc,CAAC;IAE1B,QAAQ,CAAC,cAAc,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM;IAErD,QAAQ,CAAC,uBAAuB,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM;IAE9D,QAAQ,CAAC,eAAe,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM,GAAG,IAAI;IAE7D,wBAAwB,CAAC,QAAQ,EAAE,WAAW,GAAG,oBAAoB,EAAE;IAIvE;;;OAGG;IACH,SAAS,CAAC,kBAAkB,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM,GAAG,IAAI;IAgBjE;;;;OAIG;IACH,YAAY;;;;IAOZ;;OAEG;IACH,SAAS,CAAC,WAAW,CAAC,WAAW,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,GAAG,IAAI,GAAG,OAAO;IAKhF;;;;OAIG;IACH,SAAS,CAAC,eAAe,CAAC,OAAO,EAAE,WAAW,EAAE,WAAW,GAAE,MAAY,GAAG,IAAI;IAKhF;;OAEG;IACH,SAAS,CAAC,KAAK,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAI1C;;OAEG;IACH,SAAS,CAAC,iBAAiB,CACzB,UAAU,EAAE,UAAU,GAAG,IAAI,EAC7B,eAAe,CAAC,EAAE,MAAM,GAAG,IAAI,GAC9B,OAAO;IAIV;;OAEG;IACH,SAAS,CAAC,WAAW,CAAC,QAAQ,EAAE,OAAO,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,MAAM,CAAC,GAAG,OAAO,EAAE;IAS3E;;OAEG;IACH,SAAS,CAAC,eAAe,CAAC,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,YAAY,CAAC,EAAE,MAAM,GAAG,KAAK;CA+BzF"}
package/dist/base.js ADDED
@@ -0,0 +1,122 @@
1
+ import { Value } from '@sinclair/typebox/value';
2
+ import { sdkLogger } from './logger.js';
3
+ export class RateLimitError extends Error {
4
+ retryAfterMs;
5
+ constructor(message, retryAfterMs) {
6
+ super(message);
7
+ this.name = 'RateLimitError';
8
+ this.retryAfterMs = retryAfterMs;
9
+ }
10
+ }
11
+ /**
12
+ * Base feed implementation with common functionality
13
+ * All platform-specific feeds should extend this class
14
+ */
15
+ export class BaseFeed {
16
+ authSchema = { methods: [{ type: 'none' }] };
17
+ getParentFeedDefinitions(_options) {
18
+ return [];
19
+ }
20
+ /**
21
+ * Validate options using TypeBox schema
22
+ * Subclasses can call this for schema validation before adding custom business logic
23
+ */
24
+ validateWithSchema(options) {
25
+ try {
26
+ const errors = [...Value.Errors(this.optionsSchema, options)];
27
+ if (errors.length > 0) {
28
+ // Format first error for user-friendly message
29
+ const firstError = errors[0];
30
+ const field = firstError.path.replace(/^\//, '');
31
+ return `Invalid option ${field ? `"${field}"` : ''}: ${firstError.message}`;
32
+ }
33
+ return null;
34
+ }
35
+ catch (error) {
36
+ sdkLogger.error({ error }, '[BaseFeed] Schema validation error:');
37
+ return 'Invalid feed options format';
38
+ }
39
+ }
40
+ /**
41
+ * Get rate limit information for this platform
42
+ * Override this method in platform-specific feeds
43
+ * Default is conservative: 10 requests per minute
44
+ */
45
+ getRateLimit() {
46
+ return {
47
+ requests_per_minute: 10,
48
+ recommended_interval_ms: 6000, // 6 seconds
49
+ };
50
+ }
51
+ /**
52
+ * Helper to check if content is newer than checkpoint
53
+ */
54
+ isNewerThan(contentDate, checkpoint) {
55
+ if (!checkpoint || !checkpoint.last_timestamp)
56
+ return true;
57
+ return contentDate > checkpoint.last_timestamp;
58
+ }
59
+ /**
60
+ * Calculate lookback date from options
61
+ * @param options - Feed options with optional lookback_days
62
+ * @param defaultDays - Default lookback period (default: 365)
63
+ */
64
+ getLookbackDate(options, defaultDays = 365) {
65
+ const lookbackDays = options.lookback_days || defaultDays;
66
+ return new Date(Date.now() - lookbackDays * 24 * 60 * 60 * 1000);
67
+ }
68
+ /**
69
+ * Sleep for specified milliseconds (for rate limiting)
70
+ */
71
+ sleep(ms) {
72
+ return new Promise((resolve) => setTimeout(resolve, ms));
73
+ }
74
+ /**
75
+ * Check if feed is in incremental mode
76
+ */
77
+ isIncrementalMode(checkpoint, paginationToken) {
78
+ return !!checkpoint?.last_timestamp && !paginationToken;
79
+ }
80
+ /**
81
+ * Helper to deduplicate content by origin_id
82
+ */
83
+ deduplicate(contents, seenIds) {
84
+ return contents.filter((c) => {
85
+ if (!c.origin_id)
86
+ return false;
87
+ if (seenIds.has(c.origin_id))
88
+ return false;
89
+ seenIds.add(c.origin_id);
90
+ return true;
91
+ });
92
+ }
93
+ /**
94
+ * Handle HTTP errors with structured logging and platform-specific messages
95
+ */
96
+ handleHTTPError(status, context, platformName) {
97
+ const platform = platformName || this.displayName;
98
+ sdkLogger.error({
99
+ status,
100
+ context,
101
+ platform: this.type,
102
+ timestamp: new Date().toISOString(),
103
+ }, `[${platform}Feed] HTTP ${status} error:`);
104
+ const errorMessages = {
105
+ 400: `Bad request to ${platform}: ${context}. Check your feed options.`,
106
+ 401: `Authentication failed for ${platform}. Check your API credentials.`,
107
+ 403: `Access forbidden to ${platform} resource: ${context}. The resource may be private or require authentication.`,
108
+ 404: `Resource not found on ${platform}: ${context}. Verify the resource exists.`,
109
+ 422: `Invalid request to ${platform}: ${context}. Check your parameters.`,
110
+ 429: `${platform} rate limit exceeded. Please wait before retrying.`,
111
+ 500: `${platform} server error (${status}). This is temporary, please retry later.`,
112
+ 502: `${platform} bad gateway (${status}). This is temporary, please retry later.`,
113
+ 503: `${platform} service unavailable (${status}). This is temporary, please retry later.`,
114
+ };
115
+ const message = errorMessages[status] || `${platform} API error: ${status}`;
116
+ if (status === 429) {
117
+ throw new RateLimitError(message);
118
+ }
119
+ throw new Error(message);
120
+ }
121
+ }
122
+ //# sourceMappingURL=base.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"base.js","sourceRoot":"","sources":["../src/base.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,KAAK,EAAE,MAAM,yBAAyB,CAAC;AAChD,OAAO,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAcxC,MAAM,OAAO,cAAe,SAAQ,KAAK;IAC9B,YAAY,CAAU;IAE/B,YAAY,OAAe,EAAE,YAAqB;QAChD,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,gBAAgB,CAAC;QAC7B,IAAI,CAAC,YAAY,GAAG,YAAY,CAAC;IACnC,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,OAAgB,QAAQ;IASnB,UAAU,GAAmB,EAAE,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAgBtE,wBAAwB,CAAC,QAAqB;QAC5C,OAAO,EAAE,CAAC;IACZ,CAAC;IAED;;;OAGG;IACO,kBAAkB,CAAC,OAAoB;QAC/C,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,aAAa,EAAE,OAAO,CAAC,CAAC,CAAC;YAC9D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACtB,+CAA+C;gBAC/C,MAAM,UAAU,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;gBAC7B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;gBACjD,OAAO,kBAAkB,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,UAAU,CAAC,OAAO,EAAE,CAAC;YAC9E,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,SAAS,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,EAAE,qCAAqC,CAAC,CAAC;YAClE,OAAO,6BAA6B,CAAC;QACvC,CAAC;IACH,CAAC;IAED;;;;OAIG;IACH,YAAY;QACV,OAAO;YACL,mBAAmB,EAAE,EAAE;YACvB,uBAAuB,EAAE,IAAI,EAAE,YAAY;SAC5C,CAAC;IACJ,CAAC;IAED;;OAEG;IACO,WAAW,CAAC,WAAiB,EAAE,UAA6B;QACpE,IAAI,CAAC,UAAU,IAAI,CAAC,UAAU,CAAC,cAAc;YAAE,OAAO,IAAI,CAAC;QAC3D,OAAO,WAAW,GAAG,UAAU,CAAC,cAAc,CAAC;IACjD,CAAC;IAED;;;;OAIG;IACO,eAAe,CAAC,OAAoB,EAAE,cAAsB,GAAG;QACvE,MAAM,YAAY,GAAG,OAAO,CAAC,aAAa,IAAI,WAAW,CAAC;QAC1D,OAAO,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,YAAY,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,CAAC;IACnE,CAAC;IAED;;OAEG;IACO,KAAK,CAAC,EAAU;QACxB,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;IAC3D,CAAC;IAED;;OAEG;IACO,iBAAiB,CACzB,UAA6B,EAC7B,eAA+B;QAE/B,OAAO,CAAC,CAAC,UAAU,EAAE,cAAc,IAAI,CAAC,eAAe,CAAC;IAC1D,CAAC;IAED;;OAEG;IACO,WAAW,CAAC,QAAmB,EAAE,OAAoB;QAC7D,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YAC3B,IAAI,CAAC,CAAC,CAAC,SAAS;gBAAE,OAAO,KAAK,CAAC;YAC/B,IAAI,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC;gBAAE,OAAO,KAAK,CAAC;YAC3C,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;YACzB,OAAO,IAAI,CAAC;QACd,CAAC,CAAC,CAAC;IACL,CAAC;IAED;;OAEG;IACO,eAAe,CAAC,MAAc,EAAE,OAAe,EAAE,YAAqB;QAC9E,MAAM,QAAQ,GAAG,YAAY,IAAI,IAAI,CAAC,WAAW,CAAC;QAElD,SAAS,CAAC,KAAK,CACb;YACE,MAAM;YACN,OAAO;YACP,QAAQ,EAAE,IAAI,CAAC,IAAI;YACnB,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;SACpC,EACD,IAAI,QAAQ,cAAc,MAAM,SAAS,CAC1C,CAAC;QAEF,MAAM,aAAa,GAA2B;YAC5C,GAAG,EAAE,kBAAkB,QAAQ,KAAK,OAAO,4BAA4B;YACvE,GAAG,EAAE,6BAA6B,QAAQ,+BAA+B;YACzE,GAAG,EAAE,uBAAuB,QAAQ,cAAc,OAAO,0DAA0D;YACnH,GAAG,EAAE,yBAAyB,QAAQ,KAAK,OAAO,+BAA+B;YACjF,GAAG,EAAE,sBAAsB,QAAQ,KAAK,OAAO,0BAA0B;YACzE,GAAG,EAAE,GAAG,QAAQ,oDAAoD;YACpE,GAAG,EAAE,GAAG,QAAQ,kBAAkB,MAAM,2CAA2C;YACnF,GAAG,EAAE,GAAG,QAAQ,iBAAiB,MAAM,2CAA2C;YAClF,GAAG,EAAE,GAAG,QAAQ,yBAAyB,MAAM,2CAA2C;SAC3F,CAAC;QAEF,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,GAAG,QAAQ,eAAe,MAAM,EAAE,CAAC;QAC5E,IAAI,MAAM,KAAK,GAAG,EAAE,CAAC;YACnB,MAAM,IAAI,cAAc,CAAC,OAAO,CAAC,CAAC;QACpC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC;IAC3B,CAAC;CACF"}
@@ -0,0 +1,66 @@
1
+ /**
2
+ * Browser Acquisition
3
+ *
4
+ * Single entry point for all browser-based connectors. Implements a two-layer
5
+ * cascade:
6
+ *
7
+ * 1. CDP — connect to user's real Chrome via raw CDP protocol.
8
+ * Uses CdpPage for DOM scraping (avoids Playwright's connectOverCDP crash
9
+ * on browsers with many tabs). For network interception, callers use
10
+ * Playwright's connectOverCDP on the resolved wsUrl directly.
11
+ *
12
+ * 2. Playwright — launch headless browser, inject stored cookies.
13
+ * Cookies may come from a previous CDP session (freshest) or CLI capture.
14
+ *
15
+ * Both paths share the same caller API. Fresh cookies are always captured
16
+ * from the resulting context so the caller can persist them for future fallback.
17
+ */
18
+ import type { Browser, BrowserContext, Cookie, Page } from 'playwright';
19
+ import { CdpPage } from './cdp-page.js';
20
+ export interface AcquireBrowserOptions {
21
+ /** CDP endpoint URL, 'auto' to auto-discover, or null to skip CDP entirely. */
22
+ cdpUrl?: string | 'auto' | null;
23
+ /** Stored cookies for Playwright fallback. May be empty. */
24
+ cookies: Cookie[];
25
+ /** Cookie domains to capture after sync (e.g., ['x.com', '.x.com']). */
26
+ authDomains: string[];
27
+ /** Use stealth/anti-detection mode for Playwright launch (default: false). */
28
+ stealth?: boolean;
29
+ }
30
+ export interface AcquiredBrowser {
31
+ /** Playwright Browser (null when using raw CDP). */
32
+ browser: Browser | null;
33
+ /** Playwright BrowserContext (null when using raw CDP). */
34
+ context: BrowserContext | null;
35
+ /** Playwright Page (null when using raw CDP — use cdpPage instead). */
36
+ page: Page | null;
37
+ /** Raw CDP page handle (null when using Playwright). */
38
+ cdpPage: CdpPage | null;
39
+ /** Resolved CDP WebSocket URL (available when backend is 'cdp'). */
40
+ cdpWsUrl: string | null;
41
+ /** Which backend was used. */
42
+ backend: 'cdp' | 'playwright';
43
+ /** If false, don't close the browser on cleanup (CDP — it's the user's Chrome). */
44
+ ownsBrowser: boolean;
45
+ screenshotDir: string;
46
+ }
47
+ /**
48
+ * Thrown when all cascade layers fail. Includes diagnostic info about each
49
+ * attempted layer so the user gets an actionable error message.
50
+ */
51
+ export declare class BrowserAuthCascadeError extends Error {
52
+ readonly attempts: Array<{
53
+ layer: string;
54
+ error: string;
55
+ }>;
56
+ constructor(attempts: Array<{
57
+ layer: string;
58
+ error: string;
59
+ }>);
60
+ }
61
+ /**
62
+ * Acquire a browser session using a two-layer cascade:
63
+ * CDP first (if available), then Playwright with stored cookies.
64
+ */
65
+ export declare function acquireBrowser(opts: AcquireBrowserOptions): Promise<AcquiredBrowser>;
66
+ //# sourceMappingURL=acquire.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"acquire.d.ts","sourceRoot":"","sources":["../../src/browser/acquire.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,KAAK,EAAE,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,YAAY,CAAC;AAGxE,OAAO,EAAE,OAAO,EAAE,MAAM,eAAe,CAAC;AAOxC,MAAM,WAAW,qBAAqB;IACpC,+EAA+E;IAC/E,MAAM,CAAC,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAAC;IAChC,4DAA4D;IAC5D,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,wEAAwE;IACxE,WAAW,EAAE,MAAM,EAAE,CAAC;IACtB,8EAA8E;IAC9E,OAAO,CAAC,EAAE,OAAO,CAAC;CACnB;AAED,MAAM,WAAW,eAAe;IAC9B,oDAAoD;IACpD,OAAO,EAAE,OAAO,GAAG,IAAI,CAAC;IACxB,2DAA2D;IAC3D,OAAO,EAAE,cAAc,GAAG,IAAI,CAAC;IAC/B,uEAAuE;IACvE,IAAI,EAAE,IAAI,GAAG,IAAI,CAAC;IAClB,wDAAwD;IACxD,OAAO,EAAE,OAAO,GAAG,IAAI,CAAC;IACxB,oEAAoE;IACpE,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,8BAA8B;IAC9B,OAAO,EAAE,KAAK,GAAG,YAAY,CAAC;IAC9B,mFAAmF;IACnF,WAAW,EAAE,OAAO,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;CACvB;AAED;;;GAGG;AACH,qBAAa,uBAAwB,SAAQ,KAAK;IAChD,QAAQ,CAAC,QAAQ,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;gBAE/C,QAAQ,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAW9D;AAMD;;;GAGG;AACH,wBAAsB,cAAc,CAAC,IAAI,EAAE,qBAAqB,GAAG,OAAO,CAAC,eAAe,CAAC,CAyB1F"}
@@ -0,0 +1,109 @@
1
+ /**
2
+ * Browser Acquisition
3
+ *
4
+ * Single entry point for all browser-based connectors. Implements a two-layer
5
+ * cascade:
6
+ *
7
+ * 1. CDP — connect to user's real Chrome via raw CDP protocol.
8
+ * Uses CdpPage for DOM scraping (avoids Playwright's connectOverCDP crash
9
+ * on browsers with many tabs). For network interception, callers use
10
+ * Playwright's connectOverCDP on the resolved wsUrl directly.
11
+ *
12
+ * 2. Playwright — launch headless browser, inject stored cookies.
13
+ * Cookies may come from a previous CDP session (freshest) or CLI capture.
14
+ *
15
+ * Both paths share the same caller API. Fresh cookies are always captured
16
+ * from the resulting context so the caller can persist them for future fallback.
17
+ */
18
+ import { sdkLogger } from '../logger.js';
19
+ import { resolveCdpUrl } from './cdp.js';
20
+ import { CdpPage } from './cdp-page.js';
21
+ import { launchBrowser } from './launcher.js';
22
+ /**
23
+ * Thrown when all cascade layers fail. Includes diagnostic info about each
24
+ * attempted layer so the user gets an actionable error message.
25
+ */
26
+ export class BrowserAuthCascadeError extends Error {
27
+ attempts;
28
+ constructor(attempts) {
29
+ const lines = attempts.map((a, i) => ` ${i + 1}. ${a.layer}: ${a.error}`);
30
+ super('Browser authentication failed. Tried:\n' +
31
+ lines.join('\n') +
32
+ '\n\nFix: Enable remote debugging in Chrome (chrome://inspect/#remote-debugging)\n' +
33
+ 'Or run: lobu memory browser-auth --connector <key> --auth-profile-slug <slug>');
34
+ this.name = 'BrowserAuthCascadeError';
35
+ this.attempts = attempts;
36
+ }
37
+ }
38
+ // ---------------------------------------------------------------------------
39
+ // Cascade
40
+ // ---------------------------------------------------------------------------
41
+ /**
42
+ * Acquire a browser session using a two-layer cascade:
43
+ * CDP first (if available), then Playwright with stored cookies.
44
+ */
45
+ export async function acquireBrowser(opts) {
46
+ const attempts = [];
47
+ // --- Layer 1: CDP ---
48
+ if (opts.cdpUrl !== null && opts.cdpUrl !== undefined) {
49
+ try {
50
+ return await acquireViaCdp(opts);
51
+ }
52
+ catch (err) {
53
+ attempts.push({ layer: 'CDP', error: err.message });
54
+ sdkLogger.info({ error: err.message }, '[BrowserAcquire] CDP not available, trying Playwright');
55
+ }
56
+ }
57
+ // --- Layer 2: Playwright (with stored cookies if available) ---
58
+ try {
59
+ return await acquireViaPlaywright(opts);
60
+ }
61
+ catch (err) {
62
+ attempts.push({ layer: 'Playwright', error: err.message });
63
+ }
64
+ // --- All layers failed ---
65
+ throw new BrowserAuthCascadeError(attempts);
66
+ }
67
+ // ---------------------------------------------------------------------------
68
+ // Layer implementations
69
+ // ---------------------------------------------------------------------------
70
+ async function acquireViaCdp(opts) {
71
+ const wsUrl = await resolveCdpUrl(opts.cdpUrl === 'auto' ? 'auto' : opts.cdpUrl, {
72
+ loggerLabel: 'BrowserAcquire',
73
+ preferRealBrowser: true,
74
+ });
75
+ const cdpPage = await CdpPage.create(wsUrl);
76
+ sdkLogger.info({ wsUrl }, '[BrowserAcquire] Connected via raw CDP');
77
+ return {
78
+ browser: null,
79
+ context: null,
80
+ page: null,
81
+ cdpPage,
82
+ cdpWsUrl: wsUrl,
83
+ backend: 'cdp',
84
+ ownsBrowser: false,
85
+ screenshotDir: '/tmp/feed-screenshots',
86
+ };
87
+ }
88
+ async function acquireViaPlaywright(opts) {
89
+ const { browser, screenshotDir } = await launchBrowser({}, {
90
+ stealth: opts.stealth ?? false,
91
+ });
92
+ const context = (await browser.newContext());
93
+ if (opts.cookies.length > 0) {
94
+ await context.addCookies(opts.cookies);
95
+ }
96
+ sdkLogger.info({ cookies: opts.cookies.length }, '[BrowserAcquire] Launched Playwright with stored cookies');
97
+ const page = await context.newPage();
98
+ return {
99
+ browser: browser,
100
+ context,
101
+ page,
102
+ cdpPage: null,
103
+ cdpWsUrl: null,
104
+ backend: 'playwright',
105
+ ownsBrowser: true,
106
+ screenshotDir,
107
+ };
108
+ }
109
+ //# sourceMappingURL=acquire.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"acquire.js","sourceRoot":"","sources":["../../src/browser/acquire.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAGH,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,MAAM,eAAe,CAAC;AACxC,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAmC9C;;;GAGG;AACH,MAAM,OAAO,uBAAwB,SAAQ,KAAK;IACvC,QAAQ,CAA0C;IAE3D,YAAY,QAAiD;QAC3D,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC;QAC3E,KAAK,CACH,yCAAyC;YACvC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;YAChB,mFAAmF;YACnF,+EAA+E,CAClF,CAAC;QACF,IAAI,CAAC,IAAI,GAAG,yBAAyB,CAAC;QACtC,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;IAC3B,CAAC;CACF;AAED,8EAA8E;AAC9E,UAAU;AACV,8EAA8E;AAE9E;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAAC,IAA2B;IAC9D,MAAM,QAAQ,GAA4C,EAAE,CAAC;IAE7D,uBAAuB;IACvB,IAAI,IAAI,CAAC,MAAM,KAAK,IAAI,IAAI,IAAI,CAAC,MAAM,KAAK,SAAS,EAAE,CAAC;QACtD,IAAI,CAAC;YACH,OAAO,MAAM,aAAa,CAAC,IAAI,CAAC,CAAC;QACnC,CAAC;QAAC,OAAO,GAAQ,EAAE,CAAC;YAClB,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;YACpD,SAAS,CAAC,IAAI,CACZ,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,EAAE,EACtB,uDAAuD,CACxD,CAAC;QACJ,CAAC;IACH,CAAC;IAED,iEAAiE;IACjE,IAAI,CAAC;QACH,OAAO,MAAM,oBAAoB,CAAC,IAAI,CAAC,CAAC;IAC1C,CAAC;IAAC,OAAO,GAAQ,EAAE,CAAC;QAClB,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;IAC7D,CAAC;IAED,4BAA4B;IAC5B,MAAM,IAAI,uBAAuB,CAAC,QAAQ,CAAC,CAAC;AAC9C,CAAC;AAED,8EAA8E;AAC9E,wBAAwB;AACxB,8EAA8E;AAE9E,KAAK,UAAU,aAAa,CAAC,IAA2B;IACtD,MAAM,KAAK,GAAG,MAAM,aAAa,CAAC,IAAI,CAAC,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE;QAC/E,WAAW,EAAE,gBAAgB;QAC7B,iBAAiB,EAAE,IAAI;KACxB,CAAC,CAAC;IAEH,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAE5C,SAAS,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,wCAAwC,CAAC,CAAC;IAEpE,OAAO;QACL,OAAO,EAAE,IAAI;QACb,OAAO,EAAE,IAAI;QACb,IAAI,EAAE,IAAI;QACV,OAAO;QACP,QAAQ,EAAE,KAAK;QACf,OAAO,EAAE,KAAK;QACd,WAAW,EAAE,KAAK;QAClB,aAAa,EAAE,uBAAuB;KACvC,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,oBAAoB,CAAC,IAA2B;IAC7D,MAAM,EAAE,OAAO,EAAE,aAAa,EAAE,GAAG,MAAM,aAAa,CAAC,EAAW,EAAE;QAClE,OAAO,EAAE,IAAI,CAAC,OAAO,IAAI,KAAK;KAC/B,CAAC,CAAC;IAEH,MAAM,OAAO,GAAG,CAAC,MAAO,OAAmB,CAAC,UAAU,EAAE,CAAmB,CAAC;IAC5E,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,MAAM,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACzC,CAAC;IAED,SAAS,CAAC,IAAI,CACZ,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,EAChC,0DAA0D,CAC3D,CAAC;IAEF,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;IAErC,OAAO;QACL,OAAO,EAAE,OAAkB;QAC3B,OAAO;QACP,IAAI;QACJ,OAAO,EAAE,IAAI;QACb,QAAQ,EAAE,IAAI;QACd,OAAO,EAAE,YAAY;QACrB,WAAW,EAAE,IAAI;QACjB,aAAa;KACd,CAAC;AACJ,CAAC"}
@@ -0,0 +1,48 @@
1
+ /**
2
+ * Lightweight CDP Page
3
+ *
4
+ * A minimal Page-like interface over raw Chrome DevTools Protocol.
5
+ * Used when Playwright's connectOverCDP crashes on browsers with many tabs
6
+ * (it tries to enumerate all targets, which fails on real user browsers).
7
+ *
8
+ * Supports the subset of Playwright's Page API used by DOM scraper connectors:
9
+ * goto, evaluate, waitForSelector, waitForTimeout, click, close.
10
+ */
11
+ export declare class CdpPage {
12
+ private ws;
13
+ private sessionId;
14
+ private targetId;
15
+ private msgId;
16
+ private constructor();
17
+ /**
18
+ * Create a new tab in the browser and return a CdpPage handle.
19
+ * Does NOT enumerate existing tabs — avoids the Playwright crash.
20
+ */
21
+ static create(browserWsUrl: string): Promise<CdpPage>;
22
+ private send;
23
+ /** Navigate to a URL and wait for the page to load. */
24
+ goto(url: string, options?: {
25
+ waitUntil?: string;
26
+ timeout?: number;
27
+ }): Promise<void>;
28
+ /** Evaluate a JavaScript expression in the page and return the result. */
29
+ evaluate<T = unknown>(expression: string | (() => T)): Promise<T>;
30
+ /**
31
+ * Wait for a CSS selector to appear in the DOM.
32
+ * Returns true if found, false on timeout.
33
+ */
34
+ waitForSelector(selector: string, options?: {
35
+ timeout?: number;
36
+ }): Promise<boolean>;
37
+ /** Click the first element matching a CSS selector. */
38
+ click(selector: string): Promise<void>;
39
+ /** Wait for a fixed duration. */
40
+ waitForTimeout(ms: number): Promise<void>;
41
+ /** Get the current page URL. */
42
+ url(): Promise<string>;
43
+ /** Get the page title. */
44
+ title(): Promise<string>;
45
+ /** Close the tab and disconnect. */
46
+ close(): Promise<void>;
47
+ }
48
+ //# sourceMappingURL=cdp-page.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"cdp-page.d.ts","sourceRoot":"","sources":["../../src/browser/cdp-page.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAIH,qBAAa,OAAO;IAClB,OAAO,CAAC,EAAE,CAAY;IACtB,OAAO,CAAC,SAAS,CAAS;IAC1B,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,KAAK,CAAK;IAElB,OAAO;IAMP;;;OAGG;WACU,MAAM,CAAC,YAAY,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC;IA8C3D,OAAO,CAAC,IAAI;IAiBZ,uDAAuD;IACjD,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE;QAAE,SAAS,CAAC,EAAE,MAAM,CAAC;QAAC,OAAO,CAAC,EAAE,MAAM,CAAA;KAAE,GAAG,OAAO,CAAC,IAAI,CAAC;IAmB1F,0EAA0E;IACpE,QAAQ,CAAC,CAAC,GAAG,OAAO,EAAE,UAAU,EAAE,MAAM,GAAG,CAAC,MAAM,CAAC,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC;IAevE;;;OAGG;IACG,eAAe,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE;QAAE,OAAO,CAAC,EAAE,MAAM,CAAA;KAAE,GAAG,OAAO,CAAC,OAAO,CAAC;IAazF,uDAAuD;IACjD,KAAK,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAI5C,iCAAiC;IAC3B,cAAc,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAI/C,gCAAgC;IAC1B,GAAG,IAAI,OAAO,CAAC,MAAM,CAAC;IAI5B,0BAA0B;IACpB,KAAK,IAAI,OAAO,CAAC,MAAM,CAAC;IAI9B,oCAAoC;IAC9B,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;CAmB7B"}