@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 +1,184 @@
|
|
|
1
|
+
// Pure utilities: no DOM access, no shared state.
|
|
2
|
+
// Ported from Rev-032v2 prototype (Sections 7-10, 16-17, 21-22, 35).
|
|
3
|
+
import { PII_VALUE_PATTERNS, PII_KEY_PATTERNS, SOCIAL_DOMAIN_RE } from './constants.js';
|
|
4
|
+
export const sleep = (ms) => new Promise((r) => setTimeout(r, ms));
|
|
5
|
+
// Race a Promise against a hard wall-clock timeout. On timeout resolves to
|
|
6
|
+
// `onTimeout` (default null) — never rejects, so callers can treat "no capture"
|
|
7
|
+
// and "timed out" uniformly. Does NOT abort the racing work: if `p` is a slow
|
|
8
|
+
// synchronous computation (e.g. renderNode on a big DOM), it keeps running to
|
|
9
|
+
// completion in the background but its result is discarded. Guards only real
|
|
10
|
+
// async waits (network fetches, MutationObserver settles inside clickAndWait).
|
|
11
|
+
export const withHardTimeout = (p, ms, onTimeout = null) => {
|
|
12
|
+
let timer = null;
|
|
13
|
+
const timeout = new Promise((resolve) => {
|
|
14
|
+
timer = setTimeout(() => resolve(onTimeout), ms);
|
|
15
|
+
});
|
|
16
|
+
return Promise.race([
|
|
17
|
+
p.then((v) => {
|
|
18
|
+
if (timer)
|
|
19
|
+
clearTimeout(timer);
|
|
20
|
+
return v;
|
|
21
|
+
}),
|
|
22
|
+
timeout,
|
|
23
|
+
]);
|
|
24
|
+
};
|
|
25
|
+
export const slugify = (t) => (t || 'page')
|
|
26
|
+
.toLowerCase()
|
|
27
|
+
.replace(/https?:\/\//g, '')
|
|
28
|
+
.replace(/[^a-z0-9а-яё]+/gi, '-')
|
|
29
|
+
.replace(/^-+|-+$/g, '')
|
|
30
|
+
.slice(0, 80) || 'page';
|
|
31
|
+
export const absUrl = (u, base) => {
|
|
32
|
+
const resolvedBase = base ?? globalThis.location?.href;
|
|
33
|
+
try {
|
|
34
|
+
return new URL(u || '', resolvedBase).href;
|
|
35
|
+
}
|
|
36
|
+
catch {
|
|
37
|
+
return u || '';
|
|
38
|
+
}
|
|
39
|
+
};
|
|
40
|
+
export const cleanInline = (t) => (t || '').replace(/\u00a0/g, ' ').replace(/[ \t\r\n]+/g, ' ').trim();
|
|
41
|
+
export const cleanBlock = (t) => (t || '').replace(/\u00a0/g, ' ').replace(/[ \t]+/g, ' ').replace(/\n{3,}/g, '\n\n').trim();
|
|
42
|
+
export const stripNoisyHtmlAttributes = (t) => {
|
|
43
|
+
if (!t)
|
|
44
|
+
return t || '';
|
|
45
|
+
return String(t)
|
|
46
|
+
.replace(/\s*data-od-[a-z0-9_-]*\s*=\s*"[^"]*"/gi, '')
|
|
47
|
+
.replace(/\s*data-od-[a-z0-9_-]*\s*=\s*'[^']*'/gi, '')
|
|
48
|
+
.replace(/\s*data-(?:dominant-color|has-transparency)\s*=\s*"[^"]*"/gi, '')
|
|
49
|
+
.replace(/\s*data-(?:dominant-color|has-transparency)\s*=\s*'[^']*'/gi, '')
|
|
50
|
+
.replace(/\s*style\s*=\s*"[^"]*"/gi, '')
|
|
51
|
+
.replace(/\s*style\s*=\s*'[^']*'/gi, '')
|
|
52
|
+
.replace(/\s*(?:loading|decoding|fetchpriority)\s*=\s*"[^"]*"/gi, '')
|
|
53
|
+
.replace(/\s+/g, ' ')
|
|
54
|
+
.trim();
|
|
55
|
+
};
|
|
56
|
+
export const escapeMd = (t, stripNoisy = false) => {
|
|
57
|
+
if (!t)
|
|
58
|
+
return '';
|
|
59
|
+
let c = cleanInline(t);
|
|
60
|
+
if (stripNoisy)
|
|
61
|
+
c = stripNoisyHtmlAttributes(c);
|
|
62
|
+
return c
|
|
63
|
+
.replace(/\\/g, '\\\\')
|
|
64
|
+
.replace(/`/g, '\\`')
|
|
65
|
+
.replace(/\*/g, '\\*')
|
|
66
|
+
.replace(/_/g, '\\_')
|
|
67
|
+
.replace(/\[/g, '\\[')
|
|
68
|
+
.replace(/\]/g, '\\]');
|
|
69
|
+
};
|
|
70
|
+
export const escapeMdTableCell = (t) => {
|
|
71
|
+
if (!t)
|
|
72
|
+
return '';
|
|
73
|
+
return String(t)
|
|
74
|
+
.replace(/\u00a0/g, ' ')
|
|
75
|
+
.replace(/\\/g, '\\\\')
|
|
76
|
+
.replace(/`/g, '\\`')
|
|
77
|
+
.replace(/\|/g, '\\|')
|
|
78
|
+
.replace(/\r?\n/g, '<br>')
|
|
79
|
+
.replace(/[ \t]+/g, ' ')
|
|
80
|
+
.trim();
|
|
81
|
+
};
|
|
82
|
+
export const stripOuterBlankLines = (t) => (t || '').replace(/\u00a0/g, ' ').replace(/^\s*\n/, '').replace(/\n\s*$/, '');
|
|
83
|
+
export const longestRun = (t, c) => {
|
|
84
|
+
const e = c === '`' ? '\\`' : '\\~';
|
|
85
|
+
const m = String(t || '').match(new RegExp(`${e}+`, 'g')) || [];
|
|
86
|
+
return m.reduce((mx, r) => Math.max(mx, r.length), 0);
|
|
87
|
+
};
|
|
88
|
+
export const chooseCodeFence = (t) => {
|
|
89
|
+
const b = longestRun(t, '`');
|
|
90
|
+
const tl = longestRun(t, '~');
|
|
91
|
+
if (b < 3)
|
|
92
|
+
return '```';
|
|
93
|
+
if (tl < b)
|
|
94
|
+
return '~'.repeat(Math.max(3, tl + 1));
|
|
95
|
+
return '`'.repeat(Math.max(3, b + 1));
|
|
96
|
+
};
|
|
97
|
+
// --- PII masking (Section 16) ---
|
|
98
|
+
export const maskString = (s) => {
|
|
99
|
+
let r = s;
|
|
100
|
+
for (const { name, re } of PII_VALUE_PATTERNS)
|
|
101
|
+
r = r.replace(re, `<MASKED:${name}>`);
|
|
102
|
+
return r;
|
|
103
|
+
};
|
|
104
|
+
export const maskPii = (v, enabled, d = 0) => {
|
|
105
|
+
if (!enabled)
|
|
106
|
+
return v;
|
|
107
|
+
if (d > 12)
|
|
108
|
+
return '<MASKED:depth-limit>';
|
|
109
|
+
if (v === null || v === undefined)
|
|
110
|
+
return v;
|
|
111
|
+
if (typeof v === 'string')
|
|
112
|
+
return maskString(v);
|
|
113
|
+
if (typeof v === 'number' || typeof v === 'boolean')
|
|
114
|
+
return v;
|
|
115
|
+
if (Array.isArray(v))
|
|
116
|
+
return v.map((x) => maskPii(x, enabled, d + 1));
|
|
117
|
+
if (typeof v === 'object') {
|
|
118
|
+
const o = {};
|
|
119
|
+
for (const [k, x] of Object.entries(v)) {
|
|
120
|
+
if (PII_KEY_PATTERNS.some((re) => re.test(k)))
|
|
121
|
+
o[k] = '<MASKED:by-key>';
|
|
122
|
+
else
|
|
123
|
+
o[k] = maskPii(x, enabled, d + 1);
|
|
124
|
+
}
|
|
125
|
+
return o;
|
|
126
|
+
}
|
|
127
|
+
return v;
|
|
128
|
+
};
|
|
129
|
+
// --- Form field sanitization (Section 17) ---
|
|
130
|
+
export const sanitizeFormFieldValue = (name, value, type, config) => {
|
|
131
|
+
if (!value)
|
|
132
|
+
return value || '';
|
|
133
|
+
const sn = String(name || '');
|
|
134
|
+
const sv = String(value);
|
|
135
|
+
if (config.maskFormFieldNamePatterns.some((re) => re.test(sn))) {
|
|
136
|
+
return `<MASKED:by-name, len=${sv.length}>`;
|
|
137
|
+
}
|
|
138
|
+
if (type === 'hidden' && sv.length > config.maxHiddenFormValueLength) {
|
|
139
|
+
return `${sv.slice(0, 30)}<TRUNCATED:total=${sv.length}-chars>`;
|
|
140
|
+
}
|
|
141
|
+
return sv;
|
|
142
|
+
};
|
|
143
|
+
// --- Link helpers (Sections 21-22) ---
|
|
144
|
+
export const isWordPressNoiseLink = (href, config) => {
|
|
145
|
+
if (!config.filterWordPressNoiseLinks || !href)
|
|
146
|
+
return false;
|
|
147
|
+
return config.wordpressNoiseLinkPatterns.some((p) => p.test(href));
|
|
148
|
+
};
|
|
149
|
+
export const cleanMailtoTelLabel = (href, rawLabel) => {
|
|
150
|
+
if (!href)
|
|
151
|
+
return rawLabel;
|
|
152
|
+
const isMailto = href.startsWith('mailto:');
|
|
153
|
+
const isTel = href.startsWith('tel:');
|
|
154
|
+
if (!isMailto && !isTel)
|
|
155
|
+
return rawLabel;
|
|
156
|
+
const address = href.replace(/^(?:mailto:|tel:)/, '').trim();
|
|
157
|
+
if (!rawLabel || rawLabel.length === 0)
|
|
158
|
+
return address;
|
|
159
|
+
if (rawLabel.includes(address))
|
|
160
|
+
return address;
|
|
161
|
+
if (/^(icon|иконка|логотип|значок|logo|figure|picture)\b/i.test(rawLabel))
|
|
162
|
+
return address;
|
|
163
|
+
return rawLabel;
|
|
164
|
+
};
|
|
165
|
+
export const classifyLink = (href, alt, hasImg) => {
|
|
166
|
+
const url = String(href || '');
|
|
167
|
+
const as = String(alt || '');
|
|
168
|
+
if (url.startsWith('mailto:'))
|
|
169
|
+
return { type: 'mailto', label: as || url.replace(/^mailto:/, '') };
|
|
170
|
+
if (url.startsWith('tel:'))
|
|
171
|
+
return { type: 'tel', label: as || url.replace(/^tel:/, '') };
|
|
172
|
+
const sm = url.match(SOCIAL_DOMAIN_RE);
|
|
173
|
+
if (sm) {
|
|
174
|
+
const d = sm[1].replace(/^www\./, '').replace(/^m\./, '');
|
|
175
|
+
return { type: 'social', label: d };
|
|
176
|
+
}
|
|
177
|
+
if (hasImg && /\.(jpe?g|png|webp|gif|svg|bmp|ico|avif)(?:\?|#|$)/i.test(url)) {
|
|
178
|
+
if (as.length >= 10)
|
|
179
|
+
return { type: 'linked-image', label: as };
|
|
180
|
+
}
|
|
181
|
+
return { type: 'link', label: as };
|
|
182
|
+
};
|
|
183
|
+
export const isStepNumberLike = (t) => /^(0?\d{1,3}|step\s*\d{1,3}|№\s*\d{1,3})$/i.test(cleanInline(t));
|
|
184
|
+
//# sourceMappingURL=utils.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../src/shared/utils.ts"],"names":[],"mappings":"AAAA,kDAAkD;AAClD,qEAAqE;AAErE,OAAO,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAC;AAGxF,MAAM,CAAC,MAAM,KAAK,GAAG,CAAC,EAAU,EAAiB,EAAE,CAAC,IAAI,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;AAE1F,2EAA2E;AAC3E,gFAAgF;AAChF,8EAA8E;AAC9E,8EAA8E;AAC9E,6EAA6E;AAC7E,+EAA+E;AAC/E,MAAM,CAAC,MAAM,eAAe,GAAG,CAC7B,CAAa,EACb,EAAU,EACV,YAAsB,IAAI,EACP,EAAE;IACrB,IAAI,KAAK,GAAyC,IAAI,CAAC;IACvD,MAAM,OAAO,GAAG,IAAI,OAAO,CAAW,CAAC,OAAO,EAAE,EAAE;QAChD,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,EAAE,CAAC,CAAC;IACnD,CAAC,CAAC,CAAC;IACH,OAAO,OAAO,CAAC,IAAI,CAAC;QAClB,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;YACX,IAAI,KAAK;gBAAE,YAAY,CAAC,KAAK,CAAC,CAAC;YAC/B,OAAO,CAAC,CAAC;QACX,CAAC,CAAC;QACF,OAAO;KACR,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,CAAC,CAA4B,EAAU,EAAE,CAC9D,CAAC,CAAC,IAAI,MAAM,CAAC;KACV,WAAW,EAAE;KACb,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC;KAC3B,OAAO,CAAC,kBAAkB,EAAE,GAAG,CAAC;KAChC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC;KACvB,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC;AAE5B,MAAM,CAAC,MAAM,MAAM,GAAG,CAAC,CAA4B,EAAE,IAAa,EAAU,EAAE;IAC5E,MAAM,YAAY,GAAG,IAAI,IAAI,UAAU,CAAC,QAAQ,EAAE,IAAI,CAAC;IACvD,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,YAAY,CAAC,CAAC,IAAI,CAAC;IAC7C,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,CAAC,IAAI,EAAE,CAAC;IACjB,CAAC;AACH,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,CAA4B,EAAU,EAAE,CAClE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAEvE,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAA4B,EAAU,EAAE,CACjE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;AAE9F,MAAM,CAAC,MAAM,wBAAwB,GAAG,CAAC,CAA4B,EAAU,EAAE;IAC/E,IAAI,CAAC,CAAC;QAAE,OAAO,CAAC,IAAI,EAAE,CAAC;IACvB,OAAO,MAAM,CAAC,CAAC,CAAC;SACb,OAAO,CAAC,wCAAwC,EAAE,EAAE,CAAC;SACrD,OAAO,CAAC,wCAAwC,EAAE,EAAE,CAAC;SACrD,OAAO,CAAC,6DAA6D,EAAE,EAAE,CAAC;SAC1E,OAAO,CAAC,6DAA6D,EAAE,EAAE,CAAC;SAC1E,OAAO,CAAC,0BAA0B,EAAE,EAAE,CAAC;SACvC,OAAO,CAAC,0BAA0B,EAAE,EAAE,CAAC;SACvC,OAAO,CAAC,uDAAuD,EAAE,EAAE,CAAC;SACpE,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,IAAI,EAAE,CAAC;AACZ,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,CAA4B,EAAE,UAAU,GAAG,KAAK,EAAU,EAAE;IACnF,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAC;IAClB,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;IACvB,IAAI,UAAU;QAAE,CAAC,GAAG,wBAAwB,CAAC,CAAC,CAAC,CAAC;IAChD,OAAO,CAAC;SACL,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC;SACtB,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC;SACpB,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC;SACrB,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC;SACpB,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC;SACrB,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;AAC3B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,CAA4B,EAAU,EAAE;IACxE,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAC;IAClB,OAAO,MAAM,CAAC,CAAC,CAAC;SACb,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC;SACtB,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC;SACpB,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC;SACrB,OAAO,CAAC,QAAQ,EAAE,MAAM,CAAC;SACzB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,IAAI,EAAE,CAAC;AACZ,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,CAA4B,EAAU,EAAE,CAC3E,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAA4B,EAAE,CAAY,EAAU,EAAE;IAC/E,MAAM,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC;IACpC,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAChE,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC;AACxD,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,CAA4B,EAAU,EAAE;IACtE,MAAM,CAAC,GAAG,UAAU,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAC7B,MAAM,EAAE,GAAG,UAAU,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAC9B,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACxB,IAAI,EAAE,GAAG,CAAC;QAAE,OAAO,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;IACnD,OAAO,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AACxC,CAAC,CAAC;AAEF,mCAAmC;AAEnC,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAAS,EAAU,EAAE;IAC9C,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,KAAK,MAAM,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,kBAAkB;QAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,EAAE,EAAE,WAAW,IAAI,GAAG,CAAC,CAAC;IACrF,OAAO,CAAC,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,CAAC,CAAU,EAAE,OAAgB,EAAE,CAAC,GAAG,CAAC,EAAW,EAAE;IACtE,IAAI,CAAC,OAAO;QAAE,OAAO,CAAC,CAAC;IACvB,IAAI,CAAC,GAAG,EAAE;QAAE,OAAO,sBAAsB,CAAC;IAC1C,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS;QAAE,OAAO,CAAC,CAAC;IAC5C,IAAI,OAAO,CAAC,KAAK,QAAQ;QAAE,OAAO,UAAU,CAAC,CAAC,CAAC,CAAC;IAChD,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,CAAC,KAAK,SAAS;QAAE,OAAO,CAAC,CAAC;IAC9D,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;QAAE,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IACtE,IAAI,OAAO,CAAC,KAAK,QAAQ,EAAE,CAAC;QAC1B,MAAM,CAAC,GAA4B,EAAE,CAAC;QACtC,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,CAA4B,CAAC,EAAE,CAAC;YAClE,IAAI,gBAAgB,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBAAE,CAAC,CAAC,CAAC,CAAC,GAAG,iBAAiB,CAAC;;gBAC3E,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACzC,CAAC;QACD,OAAO,CAAC,CAAC;IACX,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC,CAAC;AAEF,+CAA+C;AAE/C,MAAM,CAAC,MAAM,sBAAsB,GAAG,CACpC,IAA+B,EAC/B,KAAgC,EAChC,IAAY,EACZ,MAAsF,EAC9E,EAAE;IACV,IAAI,CAAC,KAAK;QAAE,OAAO,KAAK,IAAI,EAAE,CAAC;IAC/B,MAAM,EAAE,GAAG,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IAC9B,MAAM,EAAE,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;IACzB,IAAI,MAAM,CAAC,yBAAyB,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACvE,OAAO,wBAAwB,EAAE,CAAC,MAAM,GAAG,CAAC;IAC9C,CAAC;IACD,IAAI,IAAI,KAAK,QAAQ,IAAI,EAAE,CAAC,MAAM,GAAG,MAAM,CAAC,wBAAwB,EAAE,CAAC;QACrE,OAAO,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,oBAAoB,EAAE,CAAC,MAAM,SAAS,CAAC;IAClE,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC,CAAC;AAEF,wCAAwC;AAExC,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAClC,IAA+B,EAC/B,MAAwF,EAC/E,EAAE;IACX,IAAI,CAAC,MAAM,CAAC,yBAAyB,IAAI,CAAC,IAAI;QAAE,OAAO,KAAK,CAAC;IAC7D,OAAO,MAAM,CAAC,0BAA0B,CAAC,IAAI,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;AAC7E,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,IAAY,EAAE,QAAgB,EAAU,EAAE;IAC5E,IAAI,CAAC,IAAI;QAAE,OAAO,QAAQ,CAAC;IAC3B,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;IACtC,IAAI,CAAC,QAAQ,IAAI,CAAC,KAAK;QAAE,OAAO,QAAQ,CAAC;IAEzC,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC7D,IAAI,CAAC,QAAQ,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IACvD,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC;QAAE,OAAO,OAAO,CAAC;IAC/C,IAAI,sDAAsD,CAAC,IAAI,CAAC,QAAQ,CAAC;QAAE,OAAO,OAAO,CAAC;IAC1F,OAAO,QAAQ,CAAC;AAClB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAC1B,IAA+B,EAC/B,GAA8B,EAC9B,MAAe,EACC,EAAE;IAClB,MAAM,GAAG,GAAG,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IAC/B,MAAM,EAAE,GAAG,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC;IAC7B,IAAI,GAAG,CAAC,UAAU,CAAC,SAAS,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,EAAE,CAAC;IACnG,IAAI,GAAG,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,EAAE,CAAC;IAC1F,MAAM,EAAE,GAAG,GAAG,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;IACvC,IAAI,EAAE,EAAE,CAAC;QACP,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;QAC1D,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;IACtC,CAAC;IACD,IAAI,MAAM,IAAI,oDAAoD,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAC7E,IAAI,EAAE,CAAC,MAAM,IAAI,EAAE;YAAE,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;IAClE,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;AACrC,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,CAA4B,EAAW,EAAE,CACxE,2CAA2C,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC"}
|
package/package.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "@page2ai/core",
|
|
3
|
+
"version": "0.1.0",
|
|
4
|
+
"description": "Extract clean Markdown from any web page HTML. Works in Node.js (via linkedom) and browsers. Powers the Page2AI extension and @page2ai/mcp server.",
|
|
5
|
+
"type": "module",
|
|
6
|
+
"exports": {
|
|
7
|
+
".": {
|
|
8
|
+
"types": "./dist/node/index.d.ts",
|
|
9
|
+
"import": "./dist/node/index.js"
|
|
10
|
+
},
|
|
11
|
+
"./node": {
|
|
12
|
+
"types": "./dist/node/index.d.ts",
|
|
13
|
+
"import": "./dist/node/index.js"
|
|
14
|
+
},
|
|
15
|
+
"./browser": {
|
|
16
|
+
"types": "./dist/browser/index.d.ts",
|
|
17
|
+
"import": "./dist/browser/index.js"
|
|
18
|
+
},
|
|
19
|
+
"./shared/types": {
|
|
20
|
+
"types": "./dist/shared/types.d.ts",
|
|
21
|
+
"import": "./dist/shared/types.js"
|
|
22
|
+
},
|
|
23
|
+
"./shared/profiles": {
|
|
24
|
+
"types": "./dist/shared/profiles.d.ts",
|
|
25
|
+
"import": "./dist/shared/profiles.js"
|
|
26
|
+
}
|
|
27
|
+
},
|
|
28
|
+
"main": "./dist/node/index.js",
|
|
29
|
+
"types": "./dist/node/index.d.ts",
|
|
30
|
+
"files": [
|
|
31
|
+
"dist",
|
|
32
|
+
"README.md",
|
|
33
|
+
"LICENSE"
|
|
34
|
+
],
|
|
35
|
+
"engines": {
|
|
36
|
+
"node": ">=18.0.0"
|
|
37
|
+
},
|
|
38
|
+
"scripts": {
|
|
39
|
+
"build": "tsc -p tsconfig.build.json",
|
|
40
|
+
"build:watch": "tsc -p tsconfig.build.json --watch",
|
|
41
|
+
"test": "vitest run",
|
|
42
|
+
"test:watch": "vitest",
|
|
43
|
+
"typecheck": "tsc --noEmit",
|
|
44
|
+
"prepublishOnly": "npm run build && npm run test"
|
|
45
|
+
},
|
|
46
|
+
"publishConfig": {
|
|
47
|
+
"access": "public"
|
|
48
|
+
},
|
|
49
|
+
"dependencies": {
|
|
50
|
+
"css.escape": "^1.5.1",
|
|
51
|
+
"linkedom": "^0.18.11"
|
|
52
|
+
},
|
|
53
|
+
"devDependencies": {
|
|
54
|
+
"@types/css.escape": "^1.5.2",
|
|
55
|
+
"@types/node": "^22.10.0",
|
|
56
|
+
"typescript": "^5.7.2",
|
|
57
|
+
"vitest": "^3.2.7"
|
|
58
|
+
},
|
|
59
|
+
"keywords": [
|
|
60
|
+
"html-to-markdown",
|
|
61
|
+
"markdown",
|
|
62
|
+
"webpage-to-markdown",
|
|
63
|
+
"web-scraping",
|
|
64
|
+
"llm",
|
|
65
|
+
"ai",
|
|
66
|
+
"mcp",
|
|
67
|
+
"linkedom",
|
|
68
|
+
"readability",
|
|
69
|
+
"page2ai"
|
|
70
|
+
],
|
|
71
|
+
"author": "Igor Saevets, AI Expert and Entrepreneur",
|
|
72
|
+
"license": "MIT",
|
|
73
|
+
"homepage": "https://github.com/igorsaevets/page2ai-core",
|
|
74
|
+
"repository": {
|
|
75
|
+
"type": "git",
|
|
76
|
+
"url": "https://github.com/igorsaevets/page2ai-core.git"
|
|
77
|
+
},
|
|
78
|
+
"bugs": {
|
|
79
|
+
"url": "https://github.com/igorsaevets/page2ai-core/issues"
|
|
80
|
+
}
|
|
81
|
+
}
|