@page2ai/core 0.1.0
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.
- package/LICENSE +21 -0
- package/README.md +128 -0
- package/dist/browser/browser-adapter.d.ts +15 -0
- package/dist/browser/browser-adapter.d.ts.map +1 -0
- package/dist/browser/browser-adapter.js +74 -0
- package/dist/browser/browser-adapter.js.map +1 -0
- package/dist/browser/index.d.ts +3 -0
- package/dist/browser/index.d.ts.map +1 -0
- package/dist/browser/index.js +7 -0
- package/dist/browser/index.js.map +1 -0
- package/dist/node/basic-renderer.d.ts +9 -0
- package/dist/node/basic-renderer.d.ts.map +1 -0
- package/dist/node/basic-renderer.js +347 -0
- package/dist/node/basic-renderer.js.map +1 -0
- package/dist/node/fetch-protected.d.ts +19 -0
- package/dist/node/fetch-protected.d.ts.map +1 -0
- package/dist/node/fetch-protected.js +92 -0
- package/dist/node/fetch-protected.js.map +1 -0
- package/dist/node/index.d.ts +28 -0
- package/dist/node/index.d.ts.map +1 -0
- package/dist/node/index.js +136 -0
- package/dist/node/index.js.map +1 -0
- package/dist/node/linkedom-adapter.d.ts +17 -0
- package/dist/node/linkedom-adapter.d.ts.map +1 -0
- package/dist/node/linkedom-adapter.js +112 -0
- package/dist/node/linkedom-adapter.js.map +1 -0
- package/dist/node/ssrf-guard.d.ts +6 -0
- package/dist/node/ssrf-guard.d.ts.map +1 -0
- package/dist/node/ssrf-guard.js +73 -0
- package/dist/node/ssrf-guard.js.map +1 -0
- package/dist/shared/constants.d.ts +12 -0
- package/dist/shared/constants.d.ts.map +1 -0
- package/dist/shared/constants.js +164 -0
- package/dist/shared/constants.js.map +1 -0
- package/dist/shared/md-postprocess.d.ts +18 -0
- package/dist/shared/md-postprocess.d.ts.map +1 -0
- package/dist/shared/md-postprocess.js +538 -0
- package/dist/shared/md-postprocess.js.map +1 -0
- package/dist/shared/page-adapter.d.ts +26 -0
- package/dist/shared/page-adapter.d.ts.map +1 -0
- package/dist/shared/page-adapter.js +15 -0
- package/dist/shared/page-adapter.js.map +1 -0
- package/dist/shared/profiles.d.ts +6 -0
- package/dist/shared/profiles.d.ts.map +1 -0
- package/dist/shared/profiles.js +185 -0
- package/dist/shared/profiles.js.map +1 -0
- package/dist/shared/types.d.ts +265 -0
- package/dist/shared/types.d.ts.map +1 -0
- package/dist/shared/types.js +4 -0
- package/dist/shared/types.js.map +1 -0
- package/dist/shared/utils.d.ts +21 -0
- package/dist/shared/utils.d.ts.map +1 -0
- package/dist/shared/utils.js +184 -0
- package/dist/shared/utils.js.map +1 -0
- package/package.json +81 -0
|
@@ -0,0 +1,185 @@
|
|
|
1
|
+
// Profiles, auto-detection and config resolution.
|
|
2
|
+
// Ported from Rev-032v2 prototype (Sections 2, 2.1, 4).
|
|
3
|
+
import { DEFAULTS } from './constants.js';
|
|
4
|
+
export const PROFILES = {
|
|
5
|
+
'dashboard': {
|
|
6
|
+
interactionMode: 'none',
|
|
7
|
+
lazyLoadMode: 'none',
|
|
8
|
+
tabPanelStrategy: 'safe',
|
|
9
|
+
visualMarkersMode: 'none',
|
|
10
|
+
officialMarkdownMode: 'never',
|
|
11
|
+
structuredDataPosition: 'never-emit',
|
|
12
|
+
extractInternalState: false,
|
|
13
|
+
extractMicrodata: false,
|
|
14
|
+
extractJsonLd: false,
|
|
15
|
+
extractOpenGraph: false,
|
|
16
|
+
collapseShortAdjacentLines: false,
|
|
17
|
+
stripNoisyAttributesInLinkText: false,
|
|
18
|
+
typeLinkedImages: false,
|
|
19
|
+
filterDecorativeImages: false,
|
|
20
|
+
filterDecorativeAlt: false,
|
|
21
|
+
suppressRepeatedImages: false,
|
|
22
|
+
outputMode: 'clean',
|
|
23
|
+
},
|
|
24
|
+
'docs': {
|
|
25
|
+
interactionMode: 'safe-tabs-and-details',
|
|
26
|
+
lazyLoadMode: 'full',
|
|
27
|
+
lazyScrollSteps: 12,
|
|
28
|
+
lazyScrollWaitMs: 250,
|
|
29
|
+
lazyScrollExtraIdleMs: 600,
|
|
30
|
+
tabPanelStrategy: 'aggressive',
|
|
31
|
+
visualMarkersMode: 'important-only',
|
|
32
|
+
officialMarkdownMode: 'trusted-docs-only',
|
|
33
|
+
structuredDataPosition: 'before-content',
|
|
34
|
+
extractInternalState: false,
|
|
35
|
+
extractMicrodata: false,
|
|
36
|
+
extractJsonLd: false,
|
|
37
|
+
extractOpenGraph: false,
|
|
38
|
+
collapseShortAdjacentLines: false,
|
|
39
|
+
stripNoisyAttributesInLinkText: false,
|
|
40
|
+
typeLinkedImages: false,
|
|
41
|
+
filterDecorativeImages: true,
|
|
42
|
+
filterDecorativeAlt: true,
|
|
43
|
+
suppressRepeatedImages: true,
|
|
44
|
+
outputMode: 'clean',
|
|
45
|
+
},
|
|
46
|
+
'marketing': {
|
|
47
|
+
interactionMode: 'safe-tabs-and-details',
|
|
48
|
+
lazyLoadMode: 'safe',
|
|
49
|
+
tabPanelStrategy: 'balanced',
|
|
50
|
+
visualMarkersMode: 'important-only',
|
|
51
|
+
officialMarkdownMode: 'never',
|
|
52
|
+
structuredDataPosition: 'after-content',
|
|
53
|
+
extractInternalState: false,
|
|
54
|
+
extractMicrodata: false,
|
|
55
|
+
collapseShortAdjacentLines: true,
|
|
56
|
+
stripNoisyAttributesInLinkText: false,
|
|
57
|
+
typeLinkedImages: false,
|
|
58
|
+
filterDecorativeImages: true,
|
|
59
|
+
filterDecorativeAlt: true,
|
|
60
|
+
suppressRepeatedImages: true,
|
|
61
|
+
outputMode: 'clean',
|
|
62
|
+
},
|
|
63
|
+
'wordpress-marketing': {
|
|
64
|
+
interactionMode: 'safe-tabs-and-details',
|
|
65
|
+
lazyLoadMode: 'safe',
|
|
66
|
+
lazyScrollWaitMs: 350,
|
|
67
|
+
lazyScrollExtraIdleMs: 500,
|
|
68
|
+
tabPanelStrategy: 'balanced',
|
|
69
|
+
visualMarkersMode: 'important-only',
|
|
70
|
+
officialMarkdownMode: 'never',
|
|
71
|
+
structuredDataPosition: 'after-content',
|
|
72
|
+
extractInternalState: false,
|
|
73
|
+
extractMicrodata: false,
|
|
74
|
+
collapseShortAdjacentLines: true,
|
|
75
|
+
stripNoisyAttributesInLinkText: true,
|
|
76
|
+
typeLinkedImages: true,
|
|
77
|
+
filterDecorativeImages: true,
|
|
78
|
+
filterDecorativeAlt: true,
|
|
79
|
+
suppressRepeatedImages: true,
|
|
80
|
+
filterWordPressNoiseLinks: true,
|
|
81
|
+
outputMode: 'clean',
|
|
82
|
+
},
|
|
83
|
+
'research': {
|
|
84
|
+
interactionMode: 'aggressive',
|
|
85
|
+
lazyLoadMode: 'full',
|
|
86
|
+
tabPanelStrategy: 'aggressive',
|
|
87
|
+
visualMarkersMode: 'all',
|
|
88
|
+
officialMarkdownMode: 'page-specific',
|
|
89
|
+
structuredDataPosition: 'before-content',
|
|
90
|
+
extractInternalState: true,
|
|
91
|
+
extractMicrodata: true,
|
|
92
|
+
extractJsonLd: true,
|
|
93
|
+
extractOpenGraph: true,
|
|
94
|
+
emitImageSrcsetCandidatesComment: true,
|
|
95
|
+
collapseShortAdjacentLines: false,
|
|
96
|
+
stripNoisyAttributesInLinkText: false,
|
|
97
|
+
typeLinkedImages: false,
|
|
98
|
+
filterDecorativeImages: false,
|
|
99
|
+
filterDecorativeAlt: false,
|
|
100
|
+
suppressRepeatedImages: false,
|
|
101
|
+
outputMode: 'debug',
|
|
102
|
+
},
|
|
103
|
+
};
|
|
104
|
+
// Auto-detects docs/wordpress/marketing/dashboard from page characteristics.
|
|
105
|
+
export const detectProfile = (doc = document, loc = location) => {
|
|
106
|
+
const hostname = loc.hostname.toLowerCase();
|
|
107
|
+
const pathname = loc.pathname.toLowerCase();
|
|
108
|
+
const generator = (doc.querySelector('meta[name="generator"]')?.content || '').toLowerCase();
|
|
109
|
+
// 1) Known docs engines by meta[name="generator"]
|
|
110
|
+
if (generator.includes('mintlify') || generator.includes('docusaurus') ||
|
|
111
|
+
generator.includes('gitbook') || generator.includes('readthedocs') ||
|
|
112
|
+
generator.includes('mkdocs') || generator.includes('nextra') ||
|
|
113
|
+
generator.includes('vitepress') || generator.includes('hugo')) {
|
|
114
|
+
return 'docs';
|
|
115
|
+
}
|
|
116
|
+
// 2) Known docs DOM signatures
|
|
117
|
+
if (doc.querySelector('[data-mintlify], .docusaurus-wrapper, .gitbook-root, ' +
|
|
118
|
+
'.nextra-container, .vp-doc, [data-docs-component]')) {
|
|
119
|
+
return 'docs';
|
|
120
|
+
}
|
|
121
|
+
// 3) Known documentation hostname patterns
|
|
122
|
+
const docsHostPatterns = [
|
|
123
|
+
/^docs?\./, /\.docs?\./, /^(api|reference|sdk)\./,
|
|
124
|
+
/^developer[s]?\./, /^learn\./, /^guide[s]?\./,
|
|
125
|
+
];
|
|
126
|
+
if (docsHostPatterns.some((p) => p.test(hostname)))
|
|
127
|
+
return 'docs';
|
|
128
|
+
// 4) URL path patterns for documentation
|
|
129
|
+
if (/^\/(docs|api|reference|guide|sdk|manual|handbook|tutorial|learn)\b/i.test(pathname)) {
|
|
130
|
+
return 'docs';
|
|
131
|
+
}
|
|
132
|
+
// 5) Multiple code blocks + tablists = likely dev docs
|
|
133
|
+
const preCount = doc.querySelectorAll('pre code, pre > code').length;
|
|
134
|
+
const tablistCount = doc.querySelectorAll('[role="tablist"]').length;
|
|
135
|
+
if (preCount >= 4 && tablistCount >= 2)
|
|
136
|
+
return 'docs';
|
|
137
|
+
// 6) WordPress detection
|
|
138
|
+
if (generator.includes('wordpress') ||
|
|
139
|
+
doc.querySelector('meta[name="wp-parsely_version"], link[href*="wp-content"], link[href*="wp-includes"]')) {
|
|
140
|
+
return 'wordpress-marketing';
|
|
141
|
+
}
|
|
142
|
+
// 7) Dashboard / app detection
|
|
143
|
+
if (doc.querySelector('[role="application"], [data-reactroot]') &&
|
|
144
|
+
!doc.querySelector('article, .post, .entry-content')) {
|
|
145
|
+
if (pathname.includes('/app') || pathname.includes('/dashboard') || pathname.includes('/settings')) {
|
|
146
|
+
return 'dashboard';
|
|
147
|
+
}
|
|
148
|
+
}
|
|
149
|
+
return 'marketing';
|
|
150
|
+
};
|
|
151
|
+
export const resolveConfig = (options = {}, progress) => {
|
|
152
|
+
let profileName = options.profile || 'auto';
|
|
153
|
+
if (profileName === 'auto') {
|
|
154
|
+
profileName = detectProfile();
|
|
155
|
+
progress?.('profile', `auto-detected profile="${profileName}"`);
|
|
156
|
+
}
|
|
157
|
+
const activeProfile = profileName in PROFILES ? profileName : 'marketing';
|
|
158
|
+
const profileSettings = PROFILES[activeProfile];
|
|
159
|
+
return {
|
|
160
|
+
...DEFAULTS,
|
|
161
|
+
...profileSettings,
|
|
162
|
+
...(options.overrides || {}),
|
|
163
|
+
activeProfile,
|
|
164
|
+
};
|
|
165
|
+
};
|
|
166
|
+
export const deriveFlags = (config) => {
|
|
167
|
+
const allowClickTabs = config.interactionMode !== 'none';
|
|
168
|
+
const allowOpenDetails = config.interactionMode === 'safe-tabs-and-details' || config.interactionMode === 'aggressive';
|
|
169
|
+
const allowDropdownClicks = config.interactionMode === 'aggressive';
|
|
170
|
+
const visualImportanceEnabled = config.visualMarkersMode === 'all';
|
|
171
|
+
const stepMarkersEnabled = config.visualMarkersMode === 'important-only' || config.visualMarkersMode === 'all';
|
|
172
|
+
const ariaHiddenMarkersEnabled = config.visualMarkersMode === 'important-only' || config.visualMarkersMode === 'all';
|
|
173
|
+
const badgesEnabled = config.detectBadges &&
|
|
174
|
+
(config.visualMarkersMode === 'important-only' || config.visualMarkersMode === 'all');
|
|
175
|
+
return {
|
|
176
|
+
allowClickTabs,
|
|
177
|
+
allowOpenDetails,
|
|
178
|
+
allowDropdownClicks,
|
|
179
|
+
visualImportanceEnabled,
|
|
180
|
+
stepMarkersEnabled,
|
|
181
|
+
ariaHiddenMarkersEnabled,
|
|
182
|
+
badgesEnabled,
|
|
183
|
+
};
|
|
184
|
+
};
|
|
185
|
+
//# sourceMappingURL=profiles.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"profiles.js","sourceRoot":"","sources":["../../src/shared/profiles.ts"],"names":[],"mappings":"AAAA,kDAAkD;AAClD,wDAAwD;AAExD,OAAO,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAU1C,MAAM,CAAC,MAAM,QAAQ,GAAyC;IAC5D,WAAW,EAAE;QACX,eAAe,EAAE,MAAM;QACvB,YAAY,EAAE,MAAM;QACpB,gBAAgB,EAAE,MAAM;QACxB,iBAAiB,EAAE,MAAM;QACzB,oBAAoB,EAAE,OAAO;QAC7B,sBAAsB,EAAE,YAAY;QACpC,oBAAoB,EAAE,KAAK;QAC3B,gBAAgB,EAAE,KAAK;QACvB,aAAa,EAAE,KAAK;QACpB,gBAAgB,EAAE,KAAK;QACvB,0BAA0B,EAAE,KAAK;QACjC,8BAA8B,EAAE,KAAK;QACrC,gBAAgB,EAAE,KAAK;QACvB,sBAAsB,EAAE,KAAK;QAC7B,mBAAmB,EAAE,KAAK;QAC1B,sBAAsB,EAAE,KAAK;QAC7B,UAAU,EAAE,OAAO;KACpB;IACD,MAAM,EAAE;QACN,eAAe,EAAE,uBAAuB;QACxC,YAAY,EAAE,MAAM;QACpB,eAAe,EAAE,EAAE;QACnB,gBAAgB,EAAE,GAAG;QACrB,qBAAqB,EAAE,GAAG;QAC1B,gBAAgB,EAAE,YAAY;QAC9B,iBAAiB,EAAE,gBAAgB;QACnC,oBAAoB,EAAE,mBAAmB;QACzC,sBAAsB,EAAE,gBAAgB;QACxC,oBAAoB,EAAE,KAAK;QAC3B,gBAAgB,EAAE,KAAK;QACvB,aAAa,EAAE,KAAK;QACpB,gBAAgB,EAAE,KAAK;QACvB,0BAA0B,EAAE,KAAK;QACjC,8BAA8B,EAAE,KAAK;QACrC,gBAAgB,EAAE,KAAK;QACvB,sBAAsB,EAAE,IAAI;QAC5B,mBAAmB,EAAE,IAAI;QACzB,sBAAsB,EAAE,IAAI;QAC5B,UAAU,EAAE,OAAO;KACpB;IACD,WAAW,EAAE;QACX,eAAe,EAAE,uBAAuB;QACxC,YAAY,EAAE,MAAM;QACpB,gBAAgB,EAAE,UAAU;QAC5B,iBAAiB,EAAE,gBAAgB;QACnC,oBAAoB,EAAE,OAAO;QAC7B,sBAAsB,EAAE,eAAe;QACvC,oBAAoB,EAAE,KAAK;QAC3B,gBAAgB,EAAE,KAAK;QACvB,0BAA0B,EAAE,IAAI;QAChC,8BAA8B,EAAE,KAAK;QACrC,gBAAgB,EAAE,KAAK;QACvB,sBAAsB,EAAE,IAAI;QAC5B,mBAAmB,EAAE,IAAI;QACzB,sBAAsB,EAAE,IAAI;QAC5B,UAAU,EAAE,OAAO;KACpB;IACD,qBAAqB,EAAE;QACrB,eAAe,EAAE,uBAAuB;QACxC,YAAY,EAAE,MAAM;QACpB,gBAAgB,EAAE,GAAG;QACrB,qBAAqB,EAAE,GAAG;QAC1B,gBAAgB,EAAE,UAAU;QAC5B,iBAAiB,EAAE,gBAAgB;QACnC,oBAAoB,EAAE,OAAO;QAC7B,sBAAsB,EAAE,eAAe;QACvC,oBAAoB,EAAE,KAAK;QAC3B,gBAAgB,EAAE,KAAK;QACvB,0BAA0B,EAAE,IAAI;QAChC,8BAA8B,EAAE,IAAI;QACpC,gBAAgB,EAAE,IAAI;QACtB,sBAAsB,EAAE,IAAI;QAC5B,mBAAmB,EAAE,IAAI;QACzB,sBAAsB,EAAE,IAAI;QAC5B,yBAAyB,EAAE,IAAI;QAC/B,UAAU,EAAE,OAAO;KACpB;IACD,UAAU,EAAE;QACV,eAAe,EAAE,YAAY;QAC7B,YAAY,EAAE,MAAM;QACpB,gBAAgB,EAAE,YAAY;QAC9B,iBAAiB,EAAE,KAAK;QACxB,oBAAoB,EAAE,eAAe;QACrC,sBAAsB,EAAE,gBAAgB;QACxC,oBAAoB,EAAE,IAAI;QAC1B,gBAAgB,EAAE,IAAI;QACtB,aAAa,EAAE,IAAI;QACnB,gBAAgB,EAAE,IAAI;QACtB,gCAAgC,EAAE,IAAI;QACtC,0BAA0B,EAAE,KAAK;QACjC,8BAA8B,EAAE,KAAK;QACrC,gBAAgB,EAAE,KAAK;QACvB,sBAAsB,EAAE,KAAK;QAC7B,mBAAmB,EAAE,KAAK;QAC1B,sBAAsB,EAAE,KAAK;QAC7B,UAAU,EAAE,OAAO;KACpB;CACF,CAAC;AAEF,6EAA6E;AAC7E,MAAM,CAAC,MAAM,aAAa,GAAG,CAC3B,MAAgB,QAAQ,EACxB,MAAgB,QAAQ,EACX,EAAE;IACf,MAAM,QAAQ,GAAG,GAAG,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,QAAQ,GAAG,GAAG,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,SAAS,GAAG,CAChB,GAAG,CAAC,aAAa,CAAkB,wBAAwB,CAAC,EAAE,OAAO,IAAI,EAAE,CAC5E,CAAC,WAAW,EAAE,CAAC;IAEhB,kDAAkD;IAClD,IACE,SAAS,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,YAAY,CAAC;QAClE,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,aAAa,CAAC;QAClE,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC;QAC5D,SAAS,CAAC,QAAQ,CAAC,WAAW,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,EAC7D,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,+BAA+B;IAC/B,IACE,GAAG,CAAC,aAAa,CACf,uDAAuD;QACrD,mDAAmD,CACtD,EACD,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,2CAA2C;IAC3C,MAAM,gBAAgB,GAAG;QACvB,UAAU,EAAE,WAAW,EAAE,wBAAwB;QACjD,kBAAkB,EAAE,UAAU,EAAE,cAAc;KAC/C,CAAC;IACF,IAAI,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAAE,OAAO,MAAM,CAAC;IAElE,yCAAyC;IACzC,IAAI,qEAAqE,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;QACzF,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,uDAAuD;IACvD,MAAM,QAAQ,GAAG,GAAG,CAAC,gBAAgB,CAAC,sBAAsB,CAAC,CAAC,MAAM,CAAC;IACrE,MAAM,YAAY,GAAG,GAAG,CAAC,gBAAgB,CAAC,kBAAkB,CAAC,CAAC,MAAM,CAAC;IACrE,IAAI,QAAQ,IAAI,CAAC,IAAI,YAAY,IAAI,CAAC;QAAE,OAAO,MAAM,CAAC;IAEtD,yBAAyB;IACzB,IACE,SAAS,CAAC,QAAQ,CAAC,WAAW,CAAC;QAC/B,GAAG,CAAC,aAAa,CACf,sFAAsF,CACvF,EACD,CAAC;QACD,OAAO,qBAAqB,CAAC;IAC/B,CAAC;IAED,+BAA+B;IAC/B,IACE,GAAG,CAAC,aAAa,CAAC,wCAAwC,CAAC;QAC3D,CAAC,GAAG,CAAC,aAAa,CAAC,gCAAgC,CAAC,EACpD,CAAC;QACD,IACE,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,WAAW,CAAC,EAC9F,CAAC;YACD,OAAO,WAAW,CAAC;QACrB,CAAC;IACH,CAAC;IAED,OAAO,WAAW,CAAC;AACrB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,aAAa,GAAG,CAC3B,UAA0B,EAAE,EAC5B,QAA2B,EACX,EAAE;IAClB,IAAI,WAAW,GAAG,OAAO,CAAC,OAAO,IAAI,MAAM,CAAC;IAC5C,IAAI,WAAW,KAAK,MAAM,EAAE,CAAC;QAC3B,WAAW,GAAG,aAAa,EAAE,CAAC;QAC9B,QAAQ,EAAE,CAAC,SAAS,EAAE,0BAA0B,WAAW,GAAG,CAAC,CAAC;IAClE,CAAC;IACD,MAAM,aAAa,GAAgB,WAAW,IAAI,QAAQ,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,WAAW,CAAC;IACvF,MAAM,eAAe,GAAG,QAAQ,CAAC,aAAa,CAAC,CAAC;IAChD,OAAO;QACL,GAAG,QAAQ;QACX,GAAG,eAAe;QAClB,GAAG,CAAC,OAAO,CAAC,SAAS,IAAI,EAAE,CAAC;QAC5B,aAAa;KACd,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,MAAsB,EAAgB,EAAE;IAClE,MAAM,cAAc,GAAG,MAAM,CAAC,eAAe,KAAK,MAAM,CAAC;IACzD,MAAM,gBAAgB,GACpB,MAAM,CAAC,eAAe,KAAK,uBAAuB,IAAI,MAAM,CAAC,eAAe,KAAK,YAAY,CAAC;IAChG,MAAM,mBAAmB,GAAG,MAAM,CAAC,eAAe,KAAK,YAAY,CAAC;IACpE,MAAM,uBAAuB,GAAG,MAAM,CAAC,iBAAiB,KAAK,KAAK,CAAC;IACnE,MAAM,kBAAkB,GACtB,MAAM,CAAC,iBAAiB,KAAK,gBAAgB,IAAI,MAAM,CAAC,iBAAiB,KAAK,KAAK,CAAC;IACtF,MAAM,wBAAwB,GAC5B,MAAM,CAAC,iBAAiB,KAAK,gBAAgB,IAAI,MAAM,CAAC,iBAAiB,KAAK,KAAK,CAAC;IACtF,MAAM,aAAa,GACjB,MAAM,CAAC,YAAY;QACnB,CAAC,MAAM,CAAC,iBAAiB,KAAK,gBAAgB,IAAI,MAAM,CAAC,iBAAiB,KAAK,KAAK,CAAC,CAAC;IACxF,OAAO;QACL,cAAc;QACd,gBAAgB;QAChB,mBAAmB;QACnB,uBAAuB;QACvB,kBAAkB;QAClB,wBAAwB;QACxB,aAAa;KACd,CAAC;AACJ,CAAC,CAAC"}
|
|
@@ -0,0 +1,265 @@
|
|
|
1
|
+
export type ProfileName = 'dashboard' | 'docs' | 'marketing' | 'wordpress-marketing' | 'research';
|
|
2
|
+
export type AutoProfile = ProfileName | 'auto';
|
|
3
|
+
export type InteractionMode = 'none' | 'safe-tabs-and-details' | 'aggressive';
|
|
4
|
+
export type LazyLoadMode = 'none' | 'safe' | 'full';
|
|
5
|
+
export type TabPanelStrategy = 'safe' | 'balanced' | 'aggressive';
|
|
6
|
+
export type VisualMarkersMode = 'none' | 'important-only' | 'all';
|
|
7
|
+
export type OfficialMarkdownMode = 'never' | 'always' | 'page-specific' | 'trusted-docs-only';
|
|
8
|
+
export type StructuredDataPosition = 'never-emit' | 'before-content' | 'after-content';
|
|
9
|
+
export type OutputMode = 'clean' | 'debug';
|
|
10
|
+
export interface ExtractorConfig {
|
|
11
|
+
fileExtension: string;
|
|
12
|
+
interactionMode: InteractionMode;
|
|
13
|
+
lazyLoadMode: LazyLoadMode;
|
|
14
|
+
lazyScrollSteps: number;
|
|
15
|
+
lazySafeViewports: number;
|
|
16
|
+
lazyScrollWaitMs: number;
|
|
17
|
+
lazyScrollExtraIdleMs: number;
|
|
18
|
+
restoreScrollAfterExport: boolean;
|
|
19
|
+
tabPanelStrategy: TabPanelStrategy;
|
|
20
|
+
visualMarkersMode: VisualMarkersMode;
|
|
21
|
+
minLargeFontPx: number;
|
|
22
|
+
minBoldWeight: number;
|
|
23
|
+
outputMode: OutputMode;
|
|
24
|
+
officialMarkdownMode: OfficialMarkdownMode;
|
|
25
|
+
trustedOfficialMarkdownHosts: string[];
|
|
26
|
+
officialMarkdownMinRatio: number;
|
|
27
|
+
extractJsonLd: boolean;
|
|
28
|
+
extractOpenGraph: boolean;
|
|
29
|
+
extractMicrodata: boolean;
|
|
30
|
+
structuredDataPosition: StructuredDataPosition;
|
|
31
|
+
extractInternalState: boolean;
|
|
32
|
+
internalStateMaskingEnabled: boolean;
|
|
33
|
+
traverseShadowDom: boolean;
|
|
34
|
+
extractIframeSources: boolean;
|
|
35
|
+
emitImageSrcsetCandidatesComment: boolean;
|
|
36
|
+
typeLinkedImages: boolean;
|
|
37
|
+
includeHiddenMeaningfulText: boolean;
|
|
38
|
+
maxTextNodeLength: number;
|
|
39
|
+
includeVisibleAriaHiddenText: boolean;
|
|
40
|
+
minAriaHiddenTextLength: number;
|
|
41
|
+
longAriaHiddenThreshold: number;
|
|
42
|
+
preserveCodeWhitespace: boolean;
|
|
43
|
+
includePseudoBeforeText: boolean;
|
|
44
|
+
includePseudoAfterText: boolean;
|
|
45
|
+
extractTabs: boolean;
|
|
46
|
+
tabClickWaitMs: number;
|
|
47
|
+
tabSettleMs: number;
|
|
48
|
+
maxTabGroups: number;
|
|
49
|
+
maxButtonsPerTabGroup: number;
|
|
50
|
+
minTabButtonTextLength: number;
|
|
51
|
+
maxTabButtonTextLength: number;
|
|
52
|
+
maxTabPanelChars: number;
|
|
53
|
+
skipCapturedTabPanelsInMainRender: boolean;
|
|
54
|
+
abortTabCaptureOnUrlDrift: boolean;
|
|
55
|
+
tabFallbackMaxBodyRatio: number;
|
|
56
|
+
perTabHardTimeoutMs: number;
|
|
57
|
+
tabPhaseBudgetMs: number;
|
|
58
|
+
dropdownClickWaitMs: number;
|
|
59
|
+
dropdownSettleMs: number;
|
|
60
|
+
maxDropdownButtons: number;
|
|
61
|
+
maxDropdownPanelChars: number;
|
|
62
|
+
skipCapturedDropdownPanelsInMainRender: boolean;
|
|
63
|
+
restoreDetailsAfterExport: boolean;
|
|
64
|
+
enableFailSafeDownload: boolean;
|
|
65
|
+
qualityCheckEnabled: boolean;
|
|
66
|
+
minMainRenderToBodyRatio: number;
|
|
67
|
+
maxMainRenderToBodyRatio: number;
|
|
68
|
+
dedupeConsecutiveDuplicates: boolean;
|
|
69
|
+
minConsecutiveDuplicatesToCollapse: number;
|
|
70
|
+
collapseShortAdjacentLines: boolean;
|
|
71
|
+
shortLineCollapseMaxChars: number;
|
|
72
|
+
cleanupPunctuation: boolean;
|
|
73
|
+
dedupeAdjacentLinks: boolean;
|
|
74
|
+
aggressiveCleanup: boolean;
|
|
75
|
+
aggressiveCleanupMaxLinkTextLength: number;
|
|
76
|
+
aggressiveCleanupHtmlTagPattern: RegExp;
|
|
77
|
+
stripNoisyAttributesInLinkText: boolean;
|
|
78
|
+
filterDecorativeImages: boolean;
|
|
79
|
+
decorativeImageFilenamePatterns: RegExp[];
|
|
80
|
+
filterDecorativeAlt: boolean;
|
|
81
|
+
decorativeAltPattern: RegExp;
|
|
82
|
+
decorativeAltMaxWords: number;
|
|
83
|
+
suppressRepeatedImages: boolean;
|
|
84
|
+
suppressRepeatedImageThreshold: number;
|
|
85
|
+
filterWordPressNoiseLinks: boolean;
|
|
86
|
+
wordpressNoiseLinkPatterns: RegExp[];
|
|
87
|
+
compactLinkLabels: boolean;
|
|
88
|
+
maxLinkLabelChars: number;
|
|
89
|
+
maxHiddenFormValueLength: number;
|
|
90
|
+
maskFormFieldNamePatterns: RegExp[];
|
|
91
|
+
emitAppendix: boolean;
|
|
92
|
+
emitNoscriptFallback: boolean;
|
|
93
|
+
noscriptOnlyIfMainRenderShort: boolean;
|
|
94
|
+
detectBadges: boolean;
|
|
95
|
+
maxBadgeTextLength: number;
|
|
96
|
+
badgeClassPattern: RegExp;
|
|
97
|
+
unsafeTabButtonTextPatterns: RegExp[];
|
|
98
|
+
skipSelectors: string[];
|
|
99
|
+
}
|
|
100
|
+
export type ProfileSettings = Partial<ExtractorConfig>;
|
|
101
|
+
export interface ResolvedConfig extends ExtractorConfig {
|
|
102
|
+
activeProfile: ProfileName;
|
|
103
|
+
}
|
|
104
|
+
export interface ExtractOptions {
|
|
105
|
+
profile?: AutoProfile;
|
|
106
|
+
overrides?: ProfileSettings;
|
|
107
|
+
}
|
|
108
|
+
export type ProgressLevel = 'info' | 'warn' | 'error';
|
|
109
|
+
export type ProgressCallback = (step: string, message: string, level?: ProgressLevel) => void;
|
|
110
|
+
export interface DerivedFlags {
|
|
111
|
+
allowClickTabs: boolean;
|
|
112
|
+
allowOpenDetails: boolean;
|
|
113
|
+
allowDropdownClicks: boolean;
|
|
114
|
+
visualImportanceEnabled: boolean;
|
|
115
|
+
stepMarkersEnabled: boolean;
|
|
116
|
+
ariaHiddenMarkersEnabled: boolean;
|
|
117
|
+
badgesEnabled: boolean;
|
|
118
|
+
}
|
|
119
|
+
export type TabCaptureSource = 'aria-controls' | 'panel-after-group' | 'content-delta' | 'root-without-tab-buttons';
|
|
120
|
+
export interface TabPanelCapture {
|
|
121
|
+
label?: string;
|
|
122
|
+
lines?: string[];
|
|
123
|
+
text?: string;
|
|
124
|
+
source: TabCaptureSource;
|
|
125
|
+
}
|
|
126
|
+
export interface DropdownPanelCapture {
|
|
127
|
+
label: string;
|
|
128
|
+
lines: string[];
|
|
129
|
+
source: 'dropdown-panel';
|
|
130
|
+
}
|
|
131
|
+
export interface FormFieldRecord {
|
|
132
|
+
tag: string;
|
|
133
|
+
type: string;
|
|
134
|
+
label: string;
|
|
135
|
+
name: string;
|
|
136
|
+
placeholder: string;
|
|
137
|
+
rawValue: string;
|
|
138
|
+
aria: string;
|
|
139
|
+
text: string;
|
|
140
|
+
required: boolean;
|
|
141
|
+
autocomplete: string;
|
|
142
|
+
sanitizedValue: string;
|
|
143
|
+
}
|
|
144
|
+
export interface FormRecord {
|
|
145
|
+
action: string;
|
|
146
|
+
method: string;
|
|
147
|
+
formName: string;
|
|
148
|
+
visibleFields: FormFieldRecord[];
|
|
149
|
+
hiddenFields: FormFieldRecord[];
|
|
150
|
+
}
|
|
151
|
+
export interface IframeRecord {
|
|
152
|
+
src: string;
|
|
153
|
+
title: string;
|
|
154
|
+
}
|
|
155
|
+
export interface FooterLinkRecord {
|
|
156
|
+
text: string;
|
|
157
|
+
href: string;
|
|
158
|
+
}
|
|
159
|
+
export interface CapturedTabRecord {
|
|
160
|
+
label: string;
|
|
161
|
+
source: TabCaptureSource;
|
|
162
|
+
}
|
|
163
|
+
export interface AppendixData {
|
|
164
|
+
forms: FormRecord[];
|
|
165
|
+
iframes: IframeRecord[];
|
|
166
|
+
footerLinks: FooterLinkRecord[];
|
|
167
|
+
capturedTabs: CapturedTabRecord[];
|
|
168
|
+
images: number;
|
|
169
|
+
headings: number;
|
|
170
|
+
links: number;
|
|
171
|
+
codeBlocks: number;
|
|
172
|
+
}
|
|
173
|
+
export interface TouchedDetailsRecord {
|
|
174
|
+
el: HTMLDetailsElement;
|
|
175
|
+
hadOpen: boolean;
|
|
176
|
+
originalName: string | null;
|
|
177
|
+
hadName: boolean;
|
|
178
|
+
}
|
|
179
|
+
export interface ExtractorState {
|
|
180
|
+
tabPanelsByButtonId: Map<string, TabPanelCapture>;
|
|
181
|
+
dropdownPanelsByButtonId: Map<string, DropdownPanelCapture>;
|
|
182
|
+
capturedTabPanelElements: WeakSet<Element>;
|
|
183
|
+
capturedDropdownPanelElements: WeakSet<Element>;
|
|
184
|
+
capturedTabPanelTextSignatures: Set<string>;
|
|
185
|
+
knownTabButtonIds: Set<string>;
|
|
186
|
+
touchedDetailsElements: TouchedDetailsRecord[];
|
|
187
|
+
taggedExporterAttrElements: Set<HTMLElement>;
|
|
188
|
+
originalScrollPosition: {
|
|
189
|
+
x: number;
|
|
190
|
+
y: number;
|
|
191
|
+
} | null;
|
|
192
|
+
baselineBodyText: string;
|
|
193
|
+
initialUrl: string;
|
|
194
|
+
tabCaptureAborted: boolean;
|
|
195
|
+
tabPhaseStartMs: number | null;
|
|
196
|
+
}
|
|
197
|
+
export interface ExtractContext {
|
|
198
|
+
config: ResolvedConfig;
|
|
199
|
+
flags: DerivedFlags;
|
|
200
|
+
state: ExtractorState;
|
|
201
|
+
appendix: AppendixData;
|
|
202
|
+
progress: ProgressCallback;
|
|
203
|
+
}
|
|
204
|
+
export interface RenderOptions {
|
|
205
|
+
expandTabs?: boolean;
|
|
206
|
+
expandDropdowns?: boolean;
|
|
207
|
+
includeRegionMarkers?: boolean;
|
|
208
|
+
skipCapturedTabPanels?: boolean;
|
|
209
|
+
skipCapturedDropdownPanels?: boolean;
|
|
210
|
+
listType?: 'ul' | 'ol' | null;
|
|
211
|
+
listIndex?: number | null;
|
|
212
|
+
}
|
|
213
|
+
export interface TabGroup {
|
|
214
|
+
container: Element;
|
|
215
|
+
buttons: HTMLElement[];
|
|
216
|
+
reason: string;
|
|
217
|
+
signature: string;
|
|
218
|
+
}
|
|
219
|
+
export type LinkType = 'mailto' | 'tel' | 'social' | 'linked-image' | 'link';
|
|
220
|
+
export interface ClassifiedLink {
|
|
221
|
+
type: LinkType;
|
|
222
|
+
label: string;
|
|
223
|
+
}
|
|
224
|
+
export interface SrcResolution {
|
|
225
|
+
src: string;
|
|
226
|
+
candidates: string[];
|
|
227
|
+
}
|
|
228
|
+
export interface OfficialMarkdownResult {
|
|
229
|
+
url: string;
|
|
230
|
+
markdown: string;
|
|
231
|
+
}
|
|
232
|
+
export type QualityRatioStatus = 'OK' | 'UNDER-EXTRACTED' | 'OVER-EXTRACTED-NOISE';
|
|
233
|
+
export interface QualityReport {
|
|
234
|
+
mainRenderChars: number;
|
|
235
|
+
baselineChars: number;
|
|
236
|
+
ratio: number;
|
|
237
|
+
ratioStatus: QualityRatioStatus;
|
|
238
|
+
headings: number;
|
|
239
|
+
links: number;
|
|
240
|
+
images: number;
|
|
241
|
+
forms: number;
|
|
242
|
+
iframes: number;
|
|
243
|
+
codeBlocks: number;
|
|
244
|
+
capturedTabs: number;
|
|
245
|
+
footerLinks: number;
|
|
246
|
+
domPreCount: number;
|
|
247
|
+
codeGroupTablists: number;
|
|
248
|
+
}
|
|
249
|
+
export type ExtractStatus = 'ok' | 'official-md' | 'fallback-after-crash';
|
|
250
|
+
export interface ExtractResult {
|
|
251
|
+
status: ExtractStatus;
|
|
252
|
+
markdown: string;
|
|
253
|
+
filename: string;
|
|
254
|
+
profile: ProfileName;
|
|
255
|
+
officialMarkdownUrl?: string;
|
|
256
|
+
officialMarkdownRatio?: number;
|
|
257
|
+
quality?: QualityReport;
|
|
258
|
+
tabsCaptured: number;
|
|
259
|
+
dropdownsCaptured: number;
|
|
260
|
+
error?: {
|
|
261
|
+
name: string;
|
|
262
|
+
message: string;
|
|
263
|
+
};
|
|
264
|
+
}
|
|
265
|
+
//# sourceMappingURL=types.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/shared/types.ts"],"names":[],"mappings":"AAGA,MAAM,MAAM,WAAW,GACnB,WAAW,GACX,MAAM,GACN,WAAW,GACX,qBAAqB,GACrB,UAAU,CAAC;AAEf,MAAM,MAAM,WAAW,GAAG,WAAW,GAAG,MAAM,CAAC;AAE/C,MAAM,MAAM,eAAe,GAAG,MAAM,GAAG,uBAAuB,GAAG,YAAY,CAAC;AAC9E,MAAM,MAAM,YAAY,GAAG,MAAM,GAAG,MAAM,GAAG,MAAM,CAAC;AACpD,MAAM,MAAM,gBAAgB,GAAG,MAAM,GAAG,UAAU,GAAG,YAAY,CAAC;AAClE,MAAM,MAAM,iBAAiB,GAAG,MAAM,GAAG,gBAAgB,GAAG,KAAK,CAAC;AAClE,MAAM,MAAM,oBAAoB,GAC5B,OAAO,GACP,QAAQ,GACR,eAAe,GACf,mBAAmB,CAAC;AACxB,MAAM,MAAM,sBAAsB,GAAG,YAAY,GAAG,gBAAgB,GAAG,eAAe,CAAC;AACvF,MAAM,MAAM,UAAU,GAAG,OAAO,GAAG,OAAO,CAAC;AAI3C,MAAM,WAAW,eAAe;IAC9B,aAAa,EAAE,MAAM,CAAC;IACtB,eAAe,EAAE,eAAe,CAAC;IACjC,YAAY,EAAE,YAAY,CAAC;IAC3B,eAAe,EAAE,MAAM,CAAC;IACxB,iBAAiB,EAAE,MAAM,CAAC;IAC1B,gBAAgB,EAAE,MAAM,CAAC;IACzB,qBAAqB,EAAE,MAAM,CAAC;IAC9B,wBAAwB,EAAE,OAAO,CAAC;IAClC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,iBAAiB,EAAE,iBAAiB,CAAC;IACrC,cAAc,EAAE,MAAM,CAAC;IACvB,aAAa,EAAE,MAAM,CAAC;IACtB,UAAU,EAAE,UAAU,CAAC;IACvB,oBAAoB,EAAE,oBAAoB,CAAC;IAC3C,4BAA4B,EAAE,MAAM,EAAE,CAAC;IACvC,wBAAwB,EAAE,MAAM,CAAC;IACjC,aAAa,EAAE,OAAO,CAAC;IACvB,gBAAgB,EAAE,OAAO,CAAC;IAC1B,gBAAgB,EAAE,OAAO,CAAC;IAC1B,sBAAsB,EAAE,sBAAsB,CAAC;IAC/C,oBAAoB,EAAE,OAAO,CAAC;IAC9B,2BAA2B,EAAE,OAAO,CAAC;IACrC,iBAAiB,EAAE,OAAO,CAAC;IAC3B,oBAAoB,EAAE,OAAO,CAAC;IAC9B,gCAAgC,EAAE,OAAO,CAAC;IAC1C,gBAAgB,EAAE,OAAO,CAAC;IAC1B,2BAA2B,EAAE,OAAO,CAAC;IACrC,iBAAiB,EAAE,MAAM,CAAC;IAC1B,4BAA4B,EAAE,OAAO,CAAC;IACtC,uBAAuB,EAAE,MAAM,CAAC;IAChC,uBAAuB,EAAE,MAAM,CAAC;IAChC,sBAAsB,EAAE,OAAO,CAAC;IAChC,uBAAuB,EAAE,OAAO,CAAC;IACjC,sBAAsB,EAAE,OAAO,CAAC;IAChC,WAAW,EAAE,OAAO,CAAC;IACrB,cAAc,EAAE,MAAM,CAAC;IACvB,WAAW,EAAE,MAAM,CAAC;IACpB,YAAY,EAAE,MAAM,CAAC;IACrB,qBAAqB,EAAE,MAAM,CAAC;IAC9B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,gBAAgB,EAAE,MAAM,CAAC;IACzB,iCAAiC,EAAE,OAAO,CAAC;IAC3C,yBAAyB,EAAE,OAAO,CAAC;IACnC,uBAAuB,EAAE,MAAM,CAAC;IAIhC,mBAAmB,EAAE,MAAM,CAAC;IAK5B,gBAAgB,EAAE,MAAM,CAAC;IACzB,mBAAmB,EAAE,MAAM,CAAC;IAC5B,gBAAgB,EAAE,MAAM,CAAC;IACzB,kBAAkB,EAAE,MAAM,CAAC;IAC3B,qBAAqB,EAAE,MAAM,CAAC;IAC9B,sCAAsC,EAAE,OAAO,CAAC;IAChD,yBAAyB,EAAE,OAAO,CAAC;IACnC,sBAAsB,EAAE,OAAO,CAAC;IAChC,mBAAmB,EAAE,OAAO,CAAC;IAC7B,wBAAwB,EAAE,MAAM,CAAC;IACjC,wBAAwB,EAAE,MAAM,CAAC;IACjC,2BAA2B,EAAE,OAAO,CAAC;IACrC,kCAAkC,EAAE,MAAM,CAAC;IAC3C,0BAA0B,EAAE,OAAO,CAAC;IACpC,yBAAyB,EAAE,MAAM,CAAC;IAClC,kBAAkB,EAAE,OAAO,CAAC;IAC5B,mBAAmB,EAAE,OAAO,CAAC;IAC7B,iBAAiB,EAAE,OAAO,CAAC;IAC3B,kCAAkC,EAAE,MAAM,CAAC;IAC3C,+BAA+B,EAAE,MAAM,CAAC;IACxC,8BAA8B,EAAE,OAAO,CAAC;IACxC,sBAAsB,EAAE,OAAO,CAAC;IAChC,+BAA+B,EAAE,MAAM,EAAE,CAAC;IAC1C,mBAAmB,EAAE,OAAO,CAAC;IAC7B,oBAAoB,EAAE,MAAM,CAAC;IAC7B,qBAAqB,EAAE,MAAM,CAAC;IAC9B,sBAAsB,EAAE,OAAO,CAAC;IAChC,8BAA8B,EAAE,MAAM,CAAC;IACvC,yBAAyB,EAAE,OAAO,CAAC;IACnC,0BAA0B,EAAE,MAAM,EAAE,CAAC;IACrC,iBAAiB,EAAE,OAAO,CAAC;IAC3B,iBAAiB,EAAE,MAAM,CAAC;IAC1B,wBAAwB,EAAE,MAAM,CAAC;IACjC,yBAAyB,EAAE,MAAM,EAAE,CAAC;IACpC,YAAY,EAAE,OAAO,CAAC;IACtB,oBAAoB,EAAE,OAAO,CAAC;IAC9B,6BAA6B,EAAE,OAAO,CAAC;IACvC,YAAY,EAAE,OAAO,CAAC;IACtB,kBAAkB,EAAE,MAAM,CAAC;IAC3B,iBAAiB,EAAE,MAAM,CAAC;IAC1B,2BAA2B,EAAE,MAAM,EAAE,CAAC;IACtC,aAAa,EAAE,MAAM,EAAE,CAAC;CACzB;AAED,MAAM,MAAM,eAAe,GAAG,OAAO,CAAC,eAAe,CAAC,CAAC;AAEvD,MAAM,WAAW,cAAe,SAAQ,eAAe;IACrD,aAAa,EAAE,WAAW,CAAC;CAC5B;AAED,MAAM,WAAW,cAAc;IAC7B,OAAO,CAAC,EAAE,WAAW,CAAC;IACtB,SAAS,CAAC,EAAE,eAAe,CAAC;CAC7B;AAED,MAAM,MAAM,aAAa,GAAG,MAAM,GAAG,MAAM,GAAG,OAAO,CAAC;AACtD,MAAM,MAAM,gBAAgB,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,CAAC,EAAE,aAAa,KAAK,IAAI,CAAC;AAG9F,MAAM,WAAW,YAAY;IAC3B,cAAc,EAAE,OAAO,CAAC;IACxB,gBAAgB,EAAE,OAAO,CAAC;IAC1B,mBAAmB,EAAE,OAAO,CAAC;IAC7B,uBAAuB,EAAE,OAAO,CAAC;IACjC,kBAAkB,EAAE,OAAO,CAAC;IAC5B,wBAAwB,EAAE,OAAO,CAAC;IAClC,aAAa,EAAE,OAAO,CAAC;CACxB;AAED,MAAM,MAAM,gBAAgB,GACxB,eAAe,GACf,mBAAmB,GACnB,eAAe,GACf,0BAA0B,CAAC;AAE/B,MAAM,WAAW,eAAe;IAC9B,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC;IACjB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,gBAAgB,CAAC;CAC1B;AAED,MAAM,WAAW,oBAAoB;IACnC,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,MAAM,EAAE,gBAAgB,CAAC;CAC1B;AAED,MAAM,WAAW,eAAe;IAC9B,GAAG,EAAE,MAAM,CAAC;IACZ,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,CAAC;IACpB,QAAQ,EAAE,MAAM,CAAC;IACjB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,OAAO,CAAC;IAClB,YAAY,EAAE,MAAM,CAAC;IACrB,cAAc,EAAE,MAAM,CAAC;CACxB;AAED,MAAM,WAAW,UAAU;IACzB,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,aAAa,EAAE,eAAe,EAAE,CAAC;IACjC,YAAY,EAAE,eAAe,EAAE,CAAC;CACjC;AAED,MAAM,WAAW,YAAY;IAC3B,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,iBAAiB;IAChC,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,gBAAgB,CAAC;CAC1B;AAED,MAAM,WAAW,YAAY;IAC3B,KAAK,EAAE,UAAU,EAAE,CAAC;IACpB,OAAO,EAAE,YAAY,EAAE,CAAC;IACxB,WAAW,EAAE,gBAAgB,EAAE,CAAC;IAChC,YAAY,EAAE,iBAAiB,EAAE,CAAC;IAClC,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,oBAAoB;IACnC,EAAE,EAAE,kBAAkB,CAAC;IACvB,OAAO,EAAE,OAAO,CAAC;IACjB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,OAAO,EAAE,OAAO,CAAC;CAClB;AAGD,MAAM,WAAW,cAAc;IAC7B,mBAAmB,EAAE,GAAG,CAAC,MAAM,EAAE,eAAe,CAAC,CAAC;IAClD,wBAAwB,EAAE,GAAG,CAAC,MAAM,EAAE,oBAAoB,CAAC,CAAC;IAC5D,wBAAwB,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC;IAC3C,6BAA6B,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC;IAChD,8BAA8B,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC5C,iBAAiB,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC/B,sBAAsB,EAAE,oBAAoB,EAAE,CAAC;IAC/C,0BAA0B,EAAE,GAAG,CAAC,WAAW,CAAC,CAAC;IAC7C,sBAAsB,EAAE;QAAE,CAAC,EAAE,MAAM,CAAC;QAAC,CAAC,EAAE,MAAM,CAAA;KAAE,GAAG,IAAI,CAAC;IACxD,gBAAgB,EAAE,MAAM,CAAC;IACzB,UAAU,EAAE,MAAM,CAAC;IAKnB,iBAAiB,EAAE,OAAO,CAAC;IAG3B,eAAe,EAAE,MAAM,GAAG,IAAI,CAAC;CAChC;AAGD,MAAM,WAAW,cAAc;IAC7B,MAAM,EAAE,cAAc,CAAC;IACvB,KAAK,EAAE,YAAY,CAAC;IACpB,KAAK,EAAE,cAAc,CAAC;IACtB,QAAQ,EAAE,YAAY,CAAC;IACvB,QAAQ,EAAE,gBAAgB,CAAC;CAC5B;AAED,MAAM,WAAW,aAAa;IAC5B,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,oBAAoB,CAAC,EAAE,OAAO,CAAC;IAC/B,qBAAqB,CAAC,EAAE,OAAO,CAAC;IAChC,0BAA0B,CAAC,EAAE,OAAO,CAAC;IACrC,QAAQ,CAAC,EAAE,IAAI,GAAG,IAAI,GAAG,IAAI,CAAC;IAC9B,SAAS,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;CAC3B;AAED,MAAM,WAAW,QAAQ;IACvB,SAAS,EAAE,OAAO,CAAC;IACnB,OAAO,EAAE,WAAW,EAAE,CAAC;IACvB,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,MAAM,QAAQ,GAAG,QAAQ,GAAG,KAAK,GAAG,QAAQ,GAAG,cAAc,GAAG,MAAM,CAAC;AAE7E,MAAM,WAAW,cAAc;IAC7B,IAAI,EAAE,QAAQ,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,aAAa;IAC5B,GAAG,EAAE,MAAM,CAAC;IACZ,UAAU,EAAE,MAAM,EAAE,CAAC;CACtB;AAED,MAAM,WAAW,sBAAsB;IACrC,GAAG,EAAE,MAAM,CAAC;IACZ,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,MAAM,kBAAkB,GAAG,IAAI,GAAG,iBAAiB,GAAG,sBAAsB,CAAC;AAEnF,MAAM,WAAW,aAAa;IAC5B,eAAe,EAAE,MAAM,CAAC;IACxB,aAAa,EAAE,MAAM,CAAC;IACtB,KAAK,EAAE,MAAM,CAAC;IACd,WAAW,EAAE,kBAAkB,CAAC;IAChC,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,EAAE,MAAM,CAAC;IACrB,WAAW,EAAE,MAAM,CAAC;IACpB,WAAW,EAAE,MAAM,CAAC;IACpB,iBAAiB,EAAE,MAAM,CAAC;CAC3B;AAED,MAAM,MAAM,aAAa,GAAG,IAAI,GAAG,aAAa,GAAG,sBAAsB,CAAC;AAE1E,MAAM,WAAW,aAAa;IAC5B,MAAM,EAAE,aAAa,CAAC;IACtB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,WAAW,CAAC;IACrB,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B,qBAAqB,CAAC,EAAE,MAAM,CAAC;IAC/B,OAAO,CAAC,EAAE,aAAa,CAAC;IACxB,YAAY,EAAE,MAAM,CAAC;IACrB,iBAAiB,EAAE,MAAM,CAAC;IAC1B,KAAK,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC;CAC3C"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types.js","sourceRoot":"","sources":["../../src/shared/types.ts"],"names":[],"mappings":"AAAA,sBAAsB;AACtB,qEAAqE"}
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
import type { ClassifiedLink, ResolvedConfig } from './types.js';
|
|
2
|
+
export declare const sleep: (ms: number) => Promise<void>;
|
|
3
|
+
export declare const withHardTimeout: <T>(p: Promise<T>, ms: number, onTimeout?: T | null) => Promise<T | null>;
|
|
4
|
+
export declare const slugify: (t: string | null | undefined) => string;
|
|
5
|
+
export declare const absUrl: (u: string | null | undefined, base?: string) => string;
|
|
6
|
+
export declare const cleanInline: (t: string | null | undefined) => string;
|
|
7
|
+
export declare const cleanBlock: (t: string | null | undefined) => string;
|
|
8
|
+
export declare const stripNoisyHtmlAttributes: (t: string | null | undefined) => string;
|
|
9
|
+
export declare const escapeMd: (t: string | null | undefined, stripNoisy?: boolean) => string;
|
|
10
|
+
export declare const escapeMdTableCell: (t: string | null | undefined) => string;
|
|
11
|
+
export declare const stripOuterBlankLines: (t: string | null | undefined) => string;
|
|
12
|
+
export declare const longestRun: (t: string | null | undefined, c: "`" | "~") => number;
|
|
13
|
+
export declare const chooseCodeFence: (t: string | null | undefined) => string;
|
|
14
|
+
export declare const maskString: (s: string) => string;
|
|
15
|
+
export declare const maskPii: (v: unknown, enabled: boolean, d?: number) => unknown;
|
|
16
|
+
export declare const sanitizeFormFieldValue: (name: string | null | undefined, value: string | null | undefined, type: string, config: Pick<ResolvedConfig, "maskFormFieldNamePatterns" | "maxHiddenFormValueLength">) => string;
|
|
17
|
+
export declare const isWordPressNoiseLink: (href: string | null | undefined, config: Pick<ResolvedConfig, "filterWordPressNoiseLinks" | "wordpressNoiseLinkPatterns">) => boolean;
|
|
18
|
+
export declare const cleanMailtoTelLabel: (href: string, rawLabel: string) => string;
|
|
19
|
+
export declare const classifyLink: (href: string | null | undefined, alt: string | null | undefined, hasImg: boolean) => ClassifiedLink;
|
|
20
|
+
export declare const isStepNumberLike: (t: string | null | undefined) => boolean;
|
|
21
|
+
//# sourceMappingURL=utils.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"utils.d.ts","sourceRoot":"","sources":["../../src/shared/utils.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAEjE,eAAO,MAAM,KAAK,GAAI,IAAI,MAAM,KAAG,OAAO,CAAC,IAAI,CAA0C,CAAC;AAQ1F,eAAO,MAAM,eAAe,GAAI,CAAC,EAC/B,GAAG,OAAO,CAAC,CAAC,CAAC,EACb,IAAI,MAAM,EACV,YAAW,CAAC,GAAG,IAAW,KACzB,OAAO,CAAC,CAAC,GAAG,IAAI,CAYlB,CAAC;AAEF,eAAO,MAAM,OAAO,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MAM5B,CAAC;AAE5B,eAAO,MAAM,MAAM,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,EAAE,OAAO,MAAM,KAAG,MAOpE,CAAC;AAEF,eAAO,MAAM,WAAW,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MACW,CAAC;AAEvE,eAAO,MAAM,UAAU,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MACmC,CAAC;AAE9F,eAAO,MAAM,wBAAwB,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MAYvE,CAAC;AAEF,eAAO,MAAM,QAAQ,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,EAAE,oBAAkB,KAAG,MAW3E,CAAC;AAEF,eAAO,MAAM,iBAAiB,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MAUhE,CAAC;AAEF,eAAO,MAAM,oBAAoB,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MACW,CAAC;AAEhF,eAAO,MAAM,UAAU,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,EAAE,GAAG,GAAG,GAAG,GAAG,KAAG,MAIvE,CAAC;AAEF,eAAO,MAAM,eAAe,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,MAM9D,CAAC;AAIF,eAAO,MAAM,UAAU,GAAI,GAAG,MAAM,KAAG,MAItC,CAAC;AAEF,eAAO,MAAM,OAAO,GAAI,GAAG,OAAO,EAAE,SAAS,OAAO,EAAE,UAAK,KAAG,OAgB7D,CAAC;AAIF,eAAO,MAAM,sBAAsB,GACjC,MAAM,MAAM,GAAG,IAAI,GAAG,SAAS,EAC/B,OAAO,MAAM,GAAG,IAAI,GAAG,SAAS,EAChC,MAAM,MAAM,EACZ,QAAQ,IAAI,CAAC,cAAc,EAAE,2BAA2B,GAAG,0BAA0B,CAAC,KACrF,MAWF,CAAC;AAIF,eAAO,MAAM,oBAAoB,GAC/B,MAAM,MAAM,GAAG,IAAI,GAAG,SAAS,EAC/B,QAAQ,IAAI,CAAC,cAAc,EAAE,2BAA2B,GAAG,4BAA4B,CAAC,KACvF,OAGF,CAAC;AAEF,eAAO,MAAM,mBAAmB,GAAI,MAAM,MAAM,EAAE,UAAU,MAAM,KAAG,MAWpE,CAAC;AAEF,eAAO,MAAM,YAAY,GACvB,MAAM,MAAM,GAAG,IAAI,GAAG,SAAS,EAC/B,KAAK,MAAM,GAAG,IAAI,GAAG,SAAS,EAC9B,QAAQ,OAAO,KACd,cAcF,CAAC;AAEF,eAAO,MAAM,gBAAgB,GAAI,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,KAAG,OACE,CAAC"}
|