webpeel 0.4.0 → 0.5.0

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.
@@ -2,6 +2,56 @@
2
2
  * Structured data extraction using CSS selectors and heuristics
3
3
  */
4
4
  import { load } from 'cheerio';
5
+ /**
6
+ * Extract structured data using an LLM (OpenAI-compatible API)
7
+ */
8
+ export async function extractWithLLM(content, options) {
9
+ const { prompt, schema, llmApiKey, llmModel = 'gpt-4o-mini', llmBaseUrl = 'https://api.openai.com/v1' } = options;
10
+ if (!llmApiKey)
11
+ throw new Error('LLM extraction requires llmApiKey');
12
+ if (!prompt && !schema)
13
+ throw new Error('LLM extraction requires prompt or schema');
14
+ // Truncate content to ~4000 tokens to keep costs low
15
+ const maxChars = 16000;
16
+ const truncatedContent = content.length > maxChars
17
+ ? content.slice(0, maxChars) + '\n\n[Content truncated]'
18
+ : content;
19
+ const systemPrompt = schema
20
+ ? `Extract structured data from the following web page content. Return a JSON object matching this schema:\n${JSON.stringify(schema, null, 2)}\n\nReturn ONLY valid JSON, no explanation.`
21
+ : `Extract structured data from the following web page content based on this instruction: ${prompt}\n\nReturn ONLY valid JSON, no explanation.`;
22
+ const { fetch: undiciFetch } = await import('undici');
23
+ const response = await undiciFetch(`${llmBaseUrl}/chat/completions`, {
24
+ method: 'POST',
25
+ headers: {
26
+ 'Content-Type': 'application/json',
27
+ 'Authorization': `Bearer ${llmApiKey}`,
28
+ },
29
+ body: JSON.stringify({
30
+ model: llmModel,
31
+ messages: [
32
+ { role: 'system', content: systemPrompt },
33
+ { role: 'user', content: truncatedContent },
34
+ ],
35
+ temperature: 0,
36
+ response_format: { type: 'json_object' },
37
+ }),
38
+ });
39
+ if (!response.ok) {
40
+ const errorText = await response.text();
41
+ throw new Error(`LLM API error ${response.status}: ${errorText}`);
42
+ }
43
+ const result = await response.json();
44
+ const responseContent = result.choices?.[0]?.message?.content;
45
+ if (!responseContent) {
46
+ throw new Error('LLM returned empty response');
47
+ }
48
+ try {
49
+ return JSON.parse(responseContent);
50
+ }
51
+ catch {
52
+ throw new Error(`LLM returned invalid JSON: ${responseContent.slice(0, 200)}`);
53
+ }
54
+ }
5
55
  export function extractStructured(html, options) {
6
56
  const $ = load(html);
7
57
  const result = {};
@@ -1 +1 @@
1
- {"version":3,"file":"extract.js","sourceRoot":"","sources":["../../src/core/extract.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,SAAS,CAAC;AAG/B,MAAM,UAAU,iBAAiB,CAAC,IAAY,EAAE,OAAuB;IACrE,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;IACrB,MAAM,MAAM,GAAwB,EAAE,CAAC;IAEvC,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtB,iCAAiC;QACjC,KAAK,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;YAClE,MAAM,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;YAC7B,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC1B,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACvB,CAAC;iBAAM,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACjC,MAAM,CAAC,KAAK,CAAC,GAAG,QAAQ,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;YACjD,CAAC;iBAAM,CAAC;gBACN,MAAM,CAAC,KAAK,CAAC,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC;YACrE,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,2CAA2C;QAC3C,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,CAAC,UAAU,IAAI,OAAO,CAAC,MAAM,CAAC;QAC/D,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAoB,EAAE,CAAC;YAC1E,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,SAAS;gBAAE,SAAS,CAAC,gCAAgC;YAE3E,8CAA8C;YAC9C,MAAM,UAAU,GAAG,KAAK,CAAC,WAAW,EAAE,CAAC;YACvC,MAAM,UAAU,GAAG;gBACjB,cAAc,UAAU,IAAI;gBAC5B,SAAS,UAAU,GAAG;gBACtB,IAAI,UAAU,EAAE;gBAChB,IAAI,UAAU,EAAE;gBAChB,YAAY,UAAU,IAAI;gBAC1B,cAAc,UAAU,IAAI;gBAC5B,qBAAqB,UAAU,IAAI;aACpC,CAAC;YAEF,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;gBAC7B,MAAM,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC;gBAC1B,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClB,IAAI,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;oBACnD,IAAI,KAAK,EAAE,CAAC;wBACV,gCAAgC;wBAChC,IAAI,IAAI,EAAE,IAAI,KAAK,QAAQ,EAAE,CAAC;4BAC5B,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC;4BACvD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;gCAChB,MAAM,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC;gCACpB,MAAM;4BACR,CAAC;wBACH,CAAC;6BAAM,IAAI,IAAI,EAAE,IAAI,KAAK,SAAS,EAAE,CAAC;4BACpC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;4BACnE,MAAM;wBACR,CAAC;6BAAM,IAAI,IAAI,EAAE,IAAI,KAAK,OAAO,EAAE,CAAC;4BAClC,8BAA8B;4BAC9B,MAAM,SAAS,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC;4BACjE,MAAM,CAAC,KAAK,CAAC,GAAG,SAAS,CAAC;4BAC1B,MAAM;wBACR,CAAC;6BAAM,CAAC;4BACN,MAAM,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC;4BACtB,MAAM;wBACR,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;YAED,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,SAAS,EAAE,CAAC;gBAChC,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
1
+ {"version":3,"file":"extract.js","sourceRoot":"","sources":["../../src/core/extract.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,SAAS,CAAC;AAG/B;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,OAAe,EACf,OAAuB;IAEvB,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,GAAG,aAAa,EAAE,UAAU,GAAG,2BAA2B,EAAE,GAAG,OAAO,CAAC;IAElH,IAAI,CAAC,SAAS;QAAE,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAC;IACrE,IAAI,CAAC,MAAM,IAAI,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,0CAA0C,CAAC,CAAC;IAEpF,qDAAqD;IACrD,MAAM,QAAQ,GAAG,KAAK,CAAC;IACvB,MAAM,gBAAgB,GAAG,OAAO,CAAC,MAAM,GAAG,QAAQ;QAChD,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,GAAG,yBAAyB;QACxD,CAAC,CAAC,OAAO,CAAC;IAEZ,MAAM,YAAY,GAAG,MAAM;QACzB,CAAC,CAAC,4GAA4G,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,6CAA6C;QAC1L,CAAC,CAAC,0FAA0F,MAAM,6CAA6C,CAAC;IAElJ,MAAM,EAAE,KAAK,EAAE,WAAW,EAAE,GAAG,MAAM,MAAM,CAAC,QAAQ,CAAC,CAAC;IAEtD,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,GAAG,UAAU,mBAAmB,EAAE;QACnE,MAAM,EAAE,MAAM;QACd,OAAO,EAAE;YACP,cAAc,EAAE,kBAAkB;YAClC,eAAe,EAAE,UAAU,SAAS,EAAE;SACvC;QACD,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;YACnB,KAAK,EAAE,QAAQ;YACf,QAAQ,EAAE;gBACR,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,YAAY,EAAE;gBACzC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,gBAAgB,EAAE;aAC5C;YACD,WAAW,EAAE,CAAC;YACd,eAAe,EAAE,EAAE,IAAI,EAAE,aAAa,EAAE;SACzC,CAAC;KACH,CAAC,CAAC;IAEH,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;QACjB,MAAM,SAAS,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;QACxC,MAAM,IAAI,KAAK,CAAC,iBAAiB,QAAQ,CAAC,MAAM,KAAK,SAAS,EAAE,CAAC,CAAC;IACpE,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAS,CAAC;IAC5C,MAAM,eAAe,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC;IAE9D,IAAI,CAAC,eAAe,EAAE,CAAC;QACrB,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;IACjD,CAAC;IAED,IAAI,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;IACrC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,KAAK,CAAC,8BAA8B,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;IACjF,CAAC;AACH,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,IAAY,EAAE,OAAuB;IACrE,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;IACrB,MAAM,MAAM,GAAwB,EAAE,CAAC;IAEvC,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtB,iCAAiC;QACjC,KAAK,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;YAClE,MAAM,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;YAC7B,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC1B,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACvB,CAAC;iBAAM,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACjC,MAAM,CAAC,KAAK,CAAC,GAAG,QAAQ,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;YACjD,CAAC;iBAAM,CAAC;gBACN,MAAM,CAAC,KAAK,CAAC,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC;YACrE,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,2CAA2C;QAC3C,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,CAAC,UAAU,IAAI,OAAO,CAAC,MAAM,CAAC;QAC/D,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAoB,EAAE,CAAC;YAC1E,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,SAAS;gBAAE,SAAS,CAAC,gCAAgC;YAE3E,8CAA8C;YAC9C,MAAM,UAAU,GAAG,KAAK,CAAC,WAAW,EAAE,CAAC;YACvC,MAAM,UAAU,GAAG;gBACjB,cAAc,UAAU,IAAI;gBAC5B,SAAS,UAAU,GAAG;gBACtB,IAAI,UAAU,EAAE;gBAChB,IAAI,UAAU,EAAE;gBAChB,YAAY,UAAU,IAAI;gBAC1B,cAAc,UAAU,IAAI;gBAC5B,qBAAqB,UAAU,IAAI;aACpC,CAAC;YAEF,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;gBAC7B,MAAM,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC;gBAC1B,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClB,IAAI,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;oBACnD,IAAI,KAAK,EAAE,CAAC;wBACV,gCAAgC;wBAChC,IAAI,IAAI,EAAE,IAAI,KAAK,QAAQ,EAAE,CAAC;4BAC5B,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC;4BACvD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;gCAChB,MAAM,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC;gCACpB,MAAM;4BACR,CAAC;wBACH,CAAC;6BAAM,IAAI,IAAI,EAAE,IAAI,KAAK,SAAS,EAAE,CAAC;4BACpC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;4BACnE,MAAM;wBACR,CAAC;6BAAM,IAAI,IAAI,EAAE,IAAI,KAAK,OAAO,EAAE,CAAC;4BAClC,8BAA8B;4BAC9B,MAAM,SAAS,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC;4BACjE,MAAM,CAAC,KAAK,CAAC,GAAG,SAAS,CAAC;4BAC1B,MAAM;wBACR,CAAC;6BAAM,CAAC;4BACN,MAAM,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC;4BACtB,MAAM;wBACR,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;YAED,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,SAAS,EAAE,CAAC;gBAChC,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
@@ -15,6 +15,10 @@ export interface MapOptions {
15
15
  includePatterns?: string[];
16
16
  /** Exclude URL patterns matching these regexes */
17
17
  excludePatterns?: string[];
18
+ /** Filter URLs by relevance to this search query */
19
+ search?: string;
20
+ /** Only return URLs matching these content types */
21
+ contentTypeFilter?: string[];
18
22
  }
19
23
  export interface MapResult {
20
24
  /** All discovered URLs (deduplicated) */
@@ -1 +1 @@
1
- {"version":3,"file":"map.d.ts","sourceRoot":"","sources":["../../src/core/map.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAKH,MAAM,WAAW,UAAU;IACzB,2CAA2C;IAC3C,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,8DAA8D;IAC9D,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,+CAA+C;IAC/C,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,iDAAiD;IACjD,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,uDAAuD;IACvD,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;IAC3B,kDAAkD;IAClD,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;CAC5B;AAED,MAAM,WAAW,SAAS;IACxB,yCAAyC;IACzC,IAAI,EAAE,MAAM,EAAE,CAAC;IACf,wBAAwB;IACxB,WAAW,EAAE,MAAM,EAAE,CAAC;IACtB,4BAA4B;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,yBAAyB;IACzB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,wBAAsB,SAAS,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,GAAE,UAAe,GAAG,OAAO,CAAC,SAAS,CAAC,CA4D9F"}
1
+ {"version":3,"file":"map.d.ts","sourceRoot":"","sources":["../../src/core/map.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAKH,MAAM,WAAW,UAAU;IACzB,2CAA2C;IAC3C,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,8DAA8D;IAC9D,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,+CAA+C;IAC/C,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,iDAAiD;IACjD,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,uDAAuD;IACvD,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;IAC3B,kDAAkD;IAClD,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;IAC3B,oDAAoD;IACpD,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,oDAAoD;IACpD,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC;CAC9B;AAED,MAAM,WAAW,SAAS;IACxB,yCAAyC;IACzC,IAAI,EAAE,MAAM,EAAE,CAAC;IACf,wBAAwB;IACxB,WAAW,EAAE,MAAM,EAAE,CAAC;IACtB,4BAA4B;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,yBAAyB;IACzB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,wBAAsB,SAAS,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,GAAE,UAAe,GAAG,OAAO,CAAC,SAAS,CAAC,CAuH9F"}
package/dist/core/map.js CHANGED
@@ -6,19 +6,58 @@ import { discoverSitemap } from './sitemap.js';
6
6
  import { peel } from '../index.js';
7
7
  export async function mapDomain(startUrl, options = {}) {
8
8
  const startTime = Date.now();
9
- const { useSitemap = true, crawlHomepage = true, maxUrls = 5000, timeout = 10000, includePatterns = [], excludePatterns = [], } = options;
9
+ const { useSitemap = true, crawlHomepage = true, maxUrls = 5000, timeout = 10000, includePatterns = [], excludePatterns = [], search, contentTypeFilter = [], } = options;
10
10
  const urlObj = new URL(startUrl);
11
11
  const domain = urlObj.hostname;
12
- const allUrls = new Set();
12
+ const allUrls = new Map(); // URL -> relevance score
13
13
  let sitemapUrls = [];
14
14
  // Compile filter patterns
15
15
  const includeRegexes = includePatterns.map(p => new RegExp(p));
16
16
  const excludeRegexes = excludePatterns.map(p => new RegExp(p));
17
+ // Parse search terms
18
+ const searchTerms = search ? search.toLowerCase().split(/\s+/).filter(t => t.length > 0) : [];
19
+ /**
20
+ * Calculate relevance score for a URL based on search terms
21
+ * Scores based on matches in URL path and title/description if available
22
+ */
23
+ function calculateRelevance(url, title, description) {
24
+ if (searchTerms.length === 0)
25
+ return 1; // No search = all equal
26
+ let score = 0;
27
+ const urlLower = url.toLowerCase();
28
+ const titleLower = (title || '').toLowerCase();
29
+ const descLower = (description || '').toLowerCase();
30
+ for (const term of searchTerms) {
31
+ // URL path matches (highest weight)
32
+ if (urlLower.includes(term))
33
+ score += 3;
34
+ // Title matches
35
+ if (titleLower.includes(term))
36
+ score += 2;
37
+ // Description matches
38
+ if (descLower.includes(term))
39
+ score += 1;
40
+ }
41
+ return score;
42
+ }
43
+ /**
44
+ * Check if URL should be included based on patterns and content type
45
+ */
17
46
  function shouldInclude(url) {
18
47
  if (excludeRegexes.some(r => r.test(url)))
19
48
  return false;
20
49
  if (includeRegexes.length > 0 && !includeRegexes.some(r => r.test(url)))
21
50
  return false;
51
+ // Content type filter (check file extension)
52
+ if (contentTypeFilter.length > 0) {
53
+ const ext = url.split('.').pop()?.toLowerCase() || '';
54
+ const hasMatch = contentTypeFilter.some(type => {
55
+ const typeExt = type.replace(/^\./, '').toLowerCase();
56
+ return ext === typeExt || url.toLowerCase().includes(`.${typeExt}`);
57
+ });
58
+ if (!hasMatch)
59
+ return false;
60
+ }
22
61
  return true;
23
62
  }
24
63
  // Step 1: Sitemap discovery
@@ -29,7 +68,9 @@ export async function mapDomain(startUrl, options = {}) {
29
68
  if (allUrls.size >= maxUrls)
30
69
  break;
31
70
  if (shouldInclude(entry.url)) {
32
- allUrls.add(entry.url);
71
+ // Sitemap entries don't have title/description, just score by URL
72
+ const score = calculateRelevance(entry.url);
73
+ allUrls.set(entry.url, score);
33
74
  }
34
75
  }
35
76
  }
@@ -43,7 +84,10 @@ export async function mapDomain(startUrl, options = {}) {
43
84
  try {
44
85
  const linkUrl = new URL(link);
45
86
  if (linkUrl.hostname === domain && shouldInclude(link)) {
46
- allUrls.add(link);
87
+ const score = calculateRelevance(link);
88
+ if (!allUrls.has(link)) {
89
+ allUrls.set(link, score);
90
+ }
47
91
  }
48
92
  }
49
93
  catch { /* skip invalid URLs */ }
@@ -51,8 +95,19 @@ export async function mapDomain(startUrl, options = {}) {
51
95
  }
52
96
  catch { /* skip homepage crawl errors */ }
53
97
  }
98
+ // Sort URLs by relevance score (highest first), then alphabetically
99
+ const sortedUrls = Array.from(allUrls.entries())
100
+ .sort((a, b) => {
101
+ if (search) {
102
+ // Sort by score first if searching
103
+ if (b[1] !== a[1])
104
+ return b[1] - a[1];
105
+ }
106
+ return a[0].localeCompare(b[0]);
107
+ })
108
+ .map(([url]) => url);
54
109
  return {
55
- urls: Array.from(allUrls).sort(),
110
+ urls: sortedUrls,
56
111
  sitemapUrls,
57
112
  total: allUrls.size,
58
113
  elapsed: Date.now() - startTime,
@@ -1 +1 @@
1
- {"version":3,"file":"map.js","sourceRoot":"","sources":["../../src/core/map.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAC/C,OAAO,EAAE,IAAI,EAAE,MAAM,aAAa,CAAC;AA4BnC,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,QAAgB,EAAE,UAAsB,EAAE;IACxE,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC7B,MAAM,EACJ,UAAU,GAAG,IAAI,EACjB,aAAa,GAAG,IAAI,EACpB,OAAO,GAAG,IAAI,EACd,OAAO,GAAG,KAAK,EACf,eAAe,GAAG,EAAE,EACpB,eAAe,GAAG,EAAE,GACrB,GAAG,OAAO,CAAC;IAEZ,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC;IAC/B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,IAAI,WAAW,GAAa,EAAE,CAAC;IAE/B,0BAA0B;IAC1B,MAAM,cAAc,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAC/D,MAAM,cAAc,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAE/D,SAAS,aAAa,CAAC,GAAW;QAChC,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QACxD,IAAI,cAAc,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QACtF,OAAO,IAAI,CAAC;IACd,CAAC;IAED,4BAA4B;IAC5B,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,MAAM,EAAE,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC,CAAC;QACpE,WAAW,GAAG,OAAO,CAAC,WAAW,CAAC;QAClC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC;YACjC,IAAI,OAAO,CAAC,IAAI,IAAI,OAAO;gBAAE,MAAM;YACnC,IAAI,aAAa,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC7B,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;YACzB,CAAC;QACH,CAAC;IACH,CAAC;IAED,8CAA8C;IAC9C,IAAI,aAAa,IAAI,OAAO,CAAC,IAAI,GAAG,OAAO,EAAE,CAAC;QAC5C,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,CAAC,CAAC;YACjD,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;gBAChC,IAAI,OAAO,CAAC,IAAI,IAAI,OAAO;oBAAE,MAAM;gBACnC,IAAI,CAAC;oBACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;oBAC9B,IAAI,OAAO,CAAC,QAAQ,KAAK,MAAM,IAAI,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC;wBACvD,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;oBACpB,CAAC;gBACH,CAAC;gBAAC,MAAM,CAAC,CAAC,uBAAuB,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAAC,MAAM,CAAC,CAAC,gCAAgC,CAAC,CAAC;IAC9C,CAAC;IAED,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE;QAChC,WAAW;QACX,KAAK,EAAE,OAAO,CAAC,IAAI;QACnB,OAAO,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;KAChC,CAAC;AACJ,CAAC"}
1
+ {"version":3,"file":"map.js","sourceRoot":"","sources":["../../src/core/map.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAC/C,OAAO,EAAE,IAAI,EAAE,MAAM,aAAa,CAAC;AAgCnC,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,QAAgB,EAAE,UAAsB,EAAE;IACxE,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC7B,MAAM,EACJ,UAAU,GAAG,IAAI,EACjB,aAAa,GAAG,IAAI,EACpB,OAAO,GAAG,IAAI,EACd,OAAO,GAAG,KAAK,EACf,eAAe,GAAG,EAAE,EACpB,eAAe,GAAG,EAAE,EACpB,MAAM,EACN,iBAAiB,GAAG,EAAE,GACvB,GAAG,OAAO,CAAC;IAEZ,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC;IAC/B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC,CAAC,yBAAyB;IACpE,IAAI,WAAW,GAAa,EAAE,CAAC;IAE/B,0BAA0B;IAC1B,MAAM,cAAc,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAC/D,MAAM,cAAc,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAE/D,qBAAqB;IACrB,MAAM,WAAW,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAE9F;;;OAGG;IACH,SAAS,kBAAkB,CAAC,GAAW,EAAE,KAAc,EAAE,WAAoB;QAC3E,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC,CAAC,wBAAwB;QAEhE,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,MAAM,QAAQ,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC;QACnC,MAAM,UAAU,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;QAC/C,MAAM,SAAS,GAAG,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;QAEpD,KAAK,MAAM,IAAI,IAAI,WAAW,EAAE,CAAC;YAC/B,oCAAoC;YACpC,IAAI,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YACxC,gBAAgB;YAChB,IAAI,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;YAC1C,sBAAsB;YACtB,IAAI,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC;gBAAE,KAAK,IAAI,CAAC,CAAC;QAC3C,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;OAEG;IACH,SAAS,aAAa,CAAC,GAAW;QAChC,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QACxD,IAAI,cAAc,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QAEtF,6CAA6C;QAC7C,IAAI,iBAAiB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACjC,MAAM,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;YACtD,MAAM,QAAQ,GAAG,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE;gBAC7C,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;gBACtD,OAAO,GAAG,KAAK,OAAO,IAAI,GAAG,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,IAAI,OAAO,EAAE,CAAC,CAAC;YACtE,CAAC,CAAC,CAAC;YACH,IAAI,CAAC,QAAQ;gBAAE,OAAO,KAAK,CAAC;QAC9B,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;IAED,4BAA4B;IAC5B,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,MAAM,EAAE,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC,CAAC;QACpE,WAAW,GAAG,OAAO,CAAC,WAAW,CAAC;QAClC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC;YACjC,IAAI,OAAO,CAAC,IAAI,IAAI,OAAO;gBAAE,MAAM;YACnC,IAAI,aAAa,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC7B,kEAAkE;gBAClE,MAAM,KAAK,GAAG,kBAAkB,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;gBAC5C,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;YAChC,CAAC;QACH,CAAC;IACH,CAAC;IAED,8CAA8C;IAC9C,IAAI,aAAa,IAAI,OAAO,CAAC,IAAI,GAAG,OAAO,EAAE,CAAC;QAC5C,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,CAAC,CAAC;YACjD,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;gBAChC,IAAI,OAAO,CAAC,IAAI,IAAI,OAAO;oBAAE,MAAM;gBACnC,IAAI,CAAC;oBACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;oBAC9B,IAAI,OAAO,CAAC,QAAQ,KAAK,MAAM,IAAI,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC;wBACvD,MAAM,KAAK,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC;wBACvC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;4BACvB,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;wBAC3B,CAAC;oBACH,CAAC;gBACH,CAAC;gBAAC,MAAM,CAAC,CAAC,uBAAuB,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAAC,MAAM,CAAC,CAAC,gCAAgC,CAAC,CAAC;IAC9C,CAAC;IAED,oEAAoE;IACpE,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC;SAC7C,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACb,IAAI,MAAM,EAAE,CAAC;YACX,mCAAmC;YACnC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;gBAAE,OAAO,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACxC,CAAC;QACD,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAClC,CAAC,CAAC;SACD,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;IAEvB,OAAO;QACL,IAAI,EAAE,UAAU;QAChB,WAAW;QACX,KAAK,EAAE,OAAO,CAAC,IAAI;QACnB,OAAO,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;KAChC,CAAC;AACJ,CAAC"}
package/dist/index.d.ts CHANGED
@@ -9,6 +9,9 @@ export * from './types.js';
9
9
  export { crawl, type CrawlOptions, type CrawlResult, type CrawlProgress } from './core/crawler.js';
10
10
  export { discoverSitemap, type SitemapUrl, type SitemapResult } from './core/sitemap.js';
11
11
  export { mapDomain, type MapOptions, type MapResult } from './core/map.js';
12
+ export { extractBranding, type BrandingProfile } from './core/branding.js';
13
+ export { trackChange, getSnapshot, clearSnapshots, type ChangeResult, type Snapshot } from './core/change-tracking.js';
14
+ export { extractWithLLM } from './core/extract.js';
12
15
  /**
13
16
  * Fetch and extract content from a URL
14
17
  *
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAMH,OAAO,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;AAE5C,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAE1D,cAAc,YAAY,CAAC;AAC3B,OAAO,EAAE,KAAK,EAAE,KAAK,YAAY,EAAE,KAAK,WAAW,EAAE,KAAK,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACnG,OAAO,EAAE,eAAe,EAAE,KAAK,UAAU,EAAE,KAAK,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACzF,OAAO,EAAE,SAAS,EAAE,KAAK,UAAU,EAAE,KAAK,SAAS,EAAE,MAAM,eAAe,CAAC;AAE3E;;;;;;;;;;;;;;;GAeG;AACH,wBAAsB,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,GAAE,WAAgB,GAAG,OAAO,CAAC,UAAU,CAAC,CA2MtF;AAED;;;;;;;;;;;;;;GAcG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,MAAM,EAAE,EACd,OAAO,GAAE,WAAW,GAAG;IAAE,WAAW,CAAC,EAAE,MAAM,CAAA;CAAO,GACnD,OAAO,CAAC,CAAC,UAAU,GAAG;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,EAAE,CAAC,CAwB1D;AAED;;;GAGG;AACH,OAAO,EAAE,OAAO,EAAE,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAMH,OAAO,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;AAE5C,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAE1D,cAAc,YAAY,CAAC;AAC3B,OAAO,EAAE,KAAK,EAAE,KAAK,YAAY,EAAE,KAAK,WAAW,EAAE,KAAK,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACnG,OAAO,EAAE,eAAe,EAAE,KAAK,UAAU,EAAE,KAAK,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACzF,OAAO,EAAE,SAAS,EAAE,KAAK,UAAU,EAAE,KAAK,SAAS,EAAE,MAAM,eAAe,CAAC;AAC3E,OAAO,EAAE,eAAe,EAAE,KAAK,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,cAAc,EAAE,KAAK,YAAY,EAAE,KAAK,QAAQ,EAAE,MAAM,2BAA2B,CAAC;AACvH,OAAO,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAC;AAEnD;;;;;;;;;;;;;;;GAeG;AACH,wBAAsB,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,GAAE,WAAgB,GAAG,OAAO,CAAC,UAAU,CAAC,CA8RtF;AAED;;;;;;;;;;;;;;GAcG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,MAAM,EAAE,EACd,OAAO,GAAE,WAAW,GAAG;IAAE,WAAW,CAAC,EAAE,MAAM,CAAA;CAAO,GACnD,OAAO,CAAC,CAAC,UAAU,GAAG;IAAE,GAAG,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,EAAE,CAAC,CAwB1D;AAED;;;GAGG;AACH,OAAO,EAAE,OAAO,EAAE,CAAC"}
package/dist/index.js CHANGED
@@ -13,6 +13,9 @@ export * from './types.js';
13
13
  export { crawl } from './core/crawler.js';
14
14
  export { discoverSitemap } from './core/sitemap.js';
15
15
  export { mapDomain } from './core/map.js';
16
+ export { extractBranding } from './core/branding.js';
17
+ export { trackChange, getSnapshot, clearSnapshots } from './core/change-tracking.js';
18
+ export { extractWithLLM } from './core/extract.js';
16
19
  /**
17
20
  * Fetch and extract content from a URL
18
21
  *
@@ -49,6 +52,10 @@ export async function peel(url, options = {}) {
49
52
  if (actions && actions.length > 0) {
50
53
  render = true;
51
54
  }
55
+ // If branding is requested, force render mode
56
+ if (options.branding) {
57
+ render = true;
58
+ }
52
59
  try {
53
60
  // Fetch the page
54
61
  const fetchResult = await smartFetch(url, {
@@ -168,7 +175,15 @@ export async function peel(url, options = {}) {
168
175
  // Extract structured data if requested
169
176
  let extracted;
170
177
  if (extract && isHTML) {
171
- extracted = extractStructured(fetchResult.html, extract);
178
+ if (extract.llmApiKey && (extract.prompt || extract.schema)) {
179
+ // LLM-powered extraction
180
+ const { extractWithLLM } = await import('./core/extract.js');
181
+ extracted = await extractWithLLM(content, extract);
182
+ }
183
+ else if (extract.selectors || extract.schema) {
184
+ // CSS-based extraction (existing)
185
+ extracted = extractStructured(fetchResult.html, extract);
186
+ }
172
187
  }
173
188
  // Truncate to token budget if requested
174
189
  if (maxTokens && maxTokens > 0) {
@@ -180,6 +195,69 @@ export async function peel(url, options = {}) {
180
195
  const fingerprint = createHash('sha256').update(content).digest('hex').slice(0, 16);
181
196
  // Convert screenshot buffer to base64 if present
182
197
  const screenshotBase64 = fetchResult.screenshot?.toString('base64');
198
+ // Extract branding if requested (requires browser)
199
+ let brandingProfile;
200
+ if (options.branding && render) {
201
+ try {
202
+ // Import playwright and create a page for branding extraction
203
+ const { chromium } = await import('playwright-core');
204
+ const browser = await chromium.launch({ headless: true });
205
+ const page = await browser.newPage({
206
+ userAgent: userAgent || undefined,
207
+ });
208
+ // Navigate to the URL
209
+ await page.goto(fetchResult.url, {
210
+ waitUntil: 'domcontentloaded',
211
+ timeout: timeout || 30000,
212
+ });
213
+ // Wait if specified
214
+ if (wait > 0) {
215
+ await page.waitForTimeout(wait);
216
+ }
217
+ // Extract branding
218
+ const { extractBranding } = await import('./core/branding.js');
219
+ brandingProfile = await extractBranding(page);
220
+ // Clean up
221
+ await browser.close();
222
+ }
223
+ catch (error) {
224
+ console.error('Branding extraction failed:', error);
225
+ }
226
+ }
227
+ // Track content changes if requested
228
+ let changeResult;
229
+ if (options.changeTracking) {
230
+ try {
231
+ const { trackChange } = await import('./core/change-tracking.js');
232
+ changeResult = await trackChange(fetchResult.url, content, fingerprint);
233
+ }
234
+ catch (error) {
235
+ console.error('Change tracking failed:', error);
236
+ }
237
+ }
238
+ // Generate AI summary if requested
239
+ let summaryText;
240
+ if (options.summary && options.llm) {
241
+ try {
242
+ const { extractWithLLM } = await import('./core/extract.js');
243
+ const summaryPrompt = typeof options.summary === 'object' && options.summary.prompt
244
+ ? options.summary.prompt
245
+ : 'Summarize the main points of this content in a concise paragraph.';
246
+ const maxLength = typeof options.summary === 'object' && options.summary.maxLength
247
+ ? options.summary.maxLength
248
+ : 200;
249
+ const result = await extractWithLLM(content, {
250
+ prompt: `${summaryPrompt} Keep it under ${maxLength} words.`,
251
+ llmApiKey: options.llm.apiKey,
252
+ llmModel: options.llm.model,
253
+ llmBaseUrl: options.llm.baseUrl,
254
+ });
255
+ summaryText = result.summary || Object.values(result)[0];
256
+ }
257
+ catch (error) {
258
+ console.error('Summary generation failed:', error);
259
+ }
260
+ }
183
261
  return {
184
262
  url: fetchResult.url,
185
263
  title,
@@ -194,6 +272,9 @@ export async function peel(url, options = {}) {
194
272
  quality,
195
273
  fingerprint,
196
274
  extracted,
275
+ branding: brandingProfile,
276
+ changeTracking: changeResult,
277
+ summary: summaryText,
197
278
  };
198
279
  }
199
280
  catch (error) {
package/dist/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AACpC,OAAO,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAClD,OAAO,EAAE,cAAc,EAAE,UAAU,EAAE,cAAc,EAAE,aAAa,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAC3J,OAAO,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;AAC5C,OAAO,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AAGtD,cAAc,YAAY,CAAC;AAC3B,OAAO,EAAE,KAAK,EAA2D,MAAM,mBAAmB,CAAC;AACnG,OAAO,EAAE,eAAe,EAAuC,MAAM,mBAAmB,CAAC;AACzF,OAAO,EAAE,SAAS,EAAmC,MAAM,eAAe,CAAC;AAE3E;;;;;;;;;;;;;;;GAeG;AACH,MAAM,CAAC,KAAK,UAAU,IAAI,CAAC,GAAW,EAAE,UAAuB,EAAE;IAC/D,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAE7B,IAAI,EACF,MAAM,GAAG,KAAK,EACd,OAAO,GAAG,KAAK,EACf,IAAI,GAAG,CAAC,EACR,MAAM,GAAG,UAAU,EACnB,OAAO,GAAG,KAAK,EACf,SAAS,EACT,UAAU,GAAG,KAAK,EAClB,kBAAkB,GAAG,KAAK,EAC1B,QAAQ,EACR,OAAO,EACP,OAAO,EACP,OAAO,EACP,GAAG,GAAG,KAAK,EACX,OAAO,EACP,OAAO,EACP,SAAS,GACV,GAAG,OAAO,CAAC;IAEZ,8CAA8C;IAC9C,MAAM,KAAK,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACjD,IAAI,KAAK,EAAE,CAAC;QACV,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,gDAAgD;IAChD,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,6CAA6C;IAC7C,IAAI,OAAO,EAAE,CAAC;QACZ,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,6CAA6C;IAC7C,IAAI,OAAO,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,IAAI,CAAC;QACH,iBAAiB;QACjB,MAAM,WAAW,GAAG,MAAM,UAAU,CAAC,GAAG,EAAE;YACxC,YAAY,EAAE,MAAM;YACpB,OAAO;YACP,MAAM,EAAE,IAAI;YACZ,SAAS;YACT,SAAS,EAAE,OAAO;YAClB,UAAU;YACV,kBAAkB;YAClB,OAAO;YACP,OAAO;YACP,OAAO;SACR,CAAC,CAAC;QAEH,wCAAwC;QACxC,MAAM,EAAE,GAAG,CAAC,WAAW,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;QACzD,MAAM,MAAM,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,EAAE,IAAI,WAAW,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC;QACpH,MAAM,MAAM,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACnC,MAAM,KAAK,GAAG,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QAC9E,MAAM,WAAW,GAAG,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,eAAe,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC;QAEjK,MAAM,YAAY,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;QAEvG,IAAI,OAAe,CAAC;QACpB,IAAI,KAAK,GAAG,EAAE,CAAC;QACf,IAAI,QAAQ,GAAQ,EAAE,CAAC;QACvB,IAAI,KAAK,GAAa,EAAE,CAAC;QACzB,IAAI,OAAO,GAAG,CAAC,CAAC;QAEhB,IAAI,MAAM,EAAE,CAAC;YACX,yBAAyB;YACzB,IAAI,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC;YAC5B,IAAI,QAAQ,EAAE,CAAC;gBACb,IAAI,GAAG,aAAa,CAAC,IAAI,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;YAChD,CAAC;YAED,4CAA4C;YAC5C,MAAM,IAAI,GAAG,eAAe,CAAC,IAAI,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC;YACpD,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;YACnB,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;YACzB,KAAK,GAAG,YAAY,CAAC,IAAI,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC;YAE5C,kEAAkE;YAClE,IAAI,WAAW,GAAG,IAAI,CAAC;YACvB,IAAI,CAAC,GAAG,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACtB,MAAM,QAAQ,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;gBACzC,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;oBACtB,WAAW,GAAG,QAAQ,CAAC,IAAI,CAAC;gBAC9B,CAAC;YACH,CAAC;YAED,QAAQ,MAAM,EAAE,CAAC;gBACf,KAAK,MAAM;oBACT,OAAO,GAAG,WAAW,CAAC;oBACtB,MAAM;gBACR,KAAK,MAAM;oBACT,OAAO,GAAG,UAAU,CAAC,WAAW,CAAC,CAAC;oBAClC,MAAM;gBACR,KAAK,UAAU,CAAC;gBAChB;oBACE,OAAO,GAAG,cAAc,CAAC,WAAW,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;oBAC/C,MAAM;YACV,CAAC;YAED,OAAO,GAAG,gBAAgB,CAAC,OAAO,EAAE,WAAW,CAAC,IAAI,CAAC,CAAC;QACxD,CAAC;aAAM,IAAI,MAAM,EAAE,CAAC;YAClB,+BAA+B;YAC/B,IAAI,CAAC;gBACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;gBAC5C,OAAO,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;gBAC1C,KAAK,GAAG,eAAe,CAAC;gBAExB,uCAAuC;gBACvC,MAAM,QAAQ,GAAG,4BAA4B,CAAC;gBAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;gBAC5C,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;YAC9B,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;gBAC3B,KAAK,GAAG,2BAA2B,CAAC;YACtC,CAAC;YACD,OAAO,GAAG,GAAG,CAAC,CAAC,qCAAqC;QACtD,CAAC;aAAM,IAAI,KAAK,EAAE,CAAC;YACjB,4CAA4C;YAC5C,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;gBAE1E,yBAAyB;gBACzB,MAAM,KAAK,GAAG,CAAC,CAAC,aAAa,CAAC,CAAC;gBAC/B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACrB,KAAK,GAAG,CAAC,CAAC,+BAA+B,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,eAAe,CAAC;oBAC7E,MAAM,SAAS,GAAa,EAAE,CAAC;oBAC/B,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;wBACnB,MAAM,SAAS,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC;wBACrD,MAAM,QAAQ,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;wBACpG,MAAM,QAAQ,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,+BAA+B,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;wBAC1F,SAAS,CAAC,IAAI,CAAC,MAAM,SAAS,KAAK,QAAQ,KAAK,QAAQ,EAAE,CAAC,CAAC;wBAC5D,IAAI,QAAQ;4BAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;oBACrC,CAAC,CAAC,CAAC;oBACH,OAAO,GAAG,KAAK,KAAK,OAAO,SAAS,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC;gBAC7D,CAAC;qBAAM,CAAC;oBACN,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;oBAC3B,KAAK,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,cAAc,CAAC;gBACtD,CAAC;YACH,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;gBAC3B,KAAK,GAAG,cAAc,CAAC;YACzB,CAAC;YACD,OAAO,GAAG,GAAG,CAAC;QAChB,CAAC;aAAM,CAAC;YACN,0CAA0C;YAC1C,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;YAC3B,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,eAAe,CAAC;YAE5D,+BAA+B;YAC/B,MAAM,QAAQ,GAAG,4BAA4B,CAAC;YAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;YAC5C,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;YAC5B,OAAO,GAAG,GAAG,CAAC;QAChB,CAAC;QAED,uCAAuC;QACvC,IAAI,SAA0C,CAAC;QAC/C,IAAI,OAAO,IAAI,MAAM,EAAE,CAAC;YACtB,SAAS,GAAG,iBAAiB,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;QAC3D,CAAC;QAED,wCAAwC;QACxC,IAAI,SAAS,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YAC/B,OAAO,GAAG,qBAAqB,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;QACtD,CAAC;QAED,kDAAkD;QAClD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC;QACvC,MAAM,MAAM,GAAG,cAAc,CAAC,OAAO,CAAC,CAAC;QACvC,MAAM,WAAW,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAEpF,iDAAiD;QACjD,MAAM,gBAAgB,GAAG,WAAW,CAAC,UAAU,EAAE,QAAQ,CAAC,QAAQ,CAAC,CAAC;QAEpE,OAAO;YACL,GAAG,EAAE,WAAW,CAAC,GAAG;YACpB,KAAK;YACL,OAAO;YACP,QAAQ;YACR,KAAK;YACL,MAAM;YACN,MAAM,EAAE,WAAW,CAAC,MAAM;YAC1B,OAAO;YACP,UAAU,EAAE,gBAAgB;YAC5B,WAAW,EAAE,YAAY;YACzB,OAAO;YACP,WAAW;YACX,SAAS;SACV,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,sCAAsC;QACtC,MAAM,OAAO,EAAE,CAAC;QAChB,MAAM,KAAK,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;;GAcG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,IAAc,EACd,UAAkD,EAAE;IAEpD,MAAM,EAAE,WAAW,GAAG,CAAC,EAAE,GAAG,QAAQ,EAAE,GAAG,OAAO,CAAC;IACjD,MAAM,OAAO,GAAoD,EAAE,CAAC;IAEpE,qBAAqB;IACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,EAAE,CAAC;QAClD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,CAAC;QAC7C,MAAM,YAAY,GAAG,MAAM,OAAO,CAAC,UAAU,CAC3C,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,CACtC,CAAC;QAEF,YAAY,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACjC,IAAI,MAAM,CAAC,MAAM,KAAK,WAAW,EAAE,CAAC;gBAClC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,IAAI,CAAC;oBACX,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;oBACb,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,IAAI,eAAe;iBACjD,CAAC,CAAC;YACL,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;GAGG;AACH,OAAO,EAAE,OAAO,EAAE,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AACpC,OAAO,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAClD,OAAO,EAAE,cAAc,EAAE,UAAU,EAAE,cAAc,EAAE,aAAa,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAC3J,OAAO,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;AAC5C,OAAO,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AAGtD,cAAc,YAAY,CAAC;AAC3B,OAAO,EAAE,KAAK,EAA2D,MAAM,mBAAmB,CAAC;AACnG,OAAO,EAAE,eAAe,EAAuC,MAAM,mBAAmB,CAAC;AACzF,OAAO,EAAE,SAAS,EAAmC,MAAM,eAAe,CAAC;AAC3E,OAAO,EAAE,eAAe,EAAwB,MAAM,oBAAoB,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,cAAc,EAAoC,MAAM,2BAA2B,CAAC;AACvH,OAAO,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAC;AAEnD;;;;;;;;;;;;;;;GAeG;AACH,MAAM,CAAC,KAAK,UAAU,IAAI,CAAC,GAAW,EAAE,UAAuB,EAAE;IAC/D,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAE7B,IAAI,EACF,MAAM,GAAG,KAAK,EACd,OAAO,GAAG,KAAK,EACf,IAAI,GAAG,CAAC,EACR,MAAM,GAAG,UAAU,EACnB,OAAO,GAAG,KAAK,EACf,SAAS,EACT,UAAU,GAAG,KAAK,EAClB,kBAAkB,GAAG,KAAK,EAC1B,QAAQ,EACR,OAAO,EACP,OAAO,EACP,OAAO,EACP,GAAG,GAAG,KAAK,EACX,OAAO,EACP,OAAO,EACP,SAAS,GACV,GAAG,OAAO,CAAC;IAEZ,8CAA8C;IAC9C,MAAM,KAAK,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACjD,IAAI,KAAK,EAAE,CAAC;QACV,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,gDAAgD;IAChD,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,6CAA6C;IAC7C,IAAI,OAAO,EAAE,CAAC;QACZ,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,6CAA6C;IAC7C,IAAI,OAAO,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,8CAA8C;IAC9C,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC;QACrB,MAAM,GAAG,IAAI,CAAC;IAChB,CAAC;IAED,IAAI,CAAC;QACH,iBAAiB;QACjB,MAAM,WAAW,GAAG,MAAM,UAAU,CAAC,GAAG,EAAE;YACxC,YAAY,EAAE,MAAM;YACpB,OAAO;YACP,MAAM,EAAE,IAAI;YACZ,SAAS;YACT,SAAS,EAAE,OAAO;YAClB,UAAU;YACV,kBAAkB;YAClB,OAAO;YACP,OAAO;YACP,OAAO;SACR,CAAC,CAAC;QAEH,wCAAwC;QACxC,MAAM,EAAE,GAAG,CAAC,WAAW,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;QACzD,MAAM,MAAM,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,EAAE,IAAI,WAAW,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC;QACpH,MAAM,MAAM,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACnC,MAAM,KAAK,GAAG,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QAC9E,MAAM,WAAW,GAAG,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,eAAe,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC;QAEjK,MAAM,YAAY,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;QAEvG,IAAI,OAAe,CAAC;QACpB,IAAI,KAAK,GAAG,EAAE,CAAC;QACf,IAAI,QAAQ,GAAQ,EAAE,CAAC;QACvB,IAAI,KAAK,GAAa,EAAE,CAAC;QACzB,IAAI,OAAO,GAAG,CAAC,CAAC;QAEhB,IAAI,MAAM,EAAE,CAAC;YACX,yBAAyB;YACzB,IAAI,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC;YAC5B,IAAI,QAAQ,EAAE,CAAC;gBACb,IAAI,GAAG,aAAa,CAAC,IAAI,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;YAChD,CAAC;YAED,4CAA4C;YAC5C,MAAM,IAAI,GAAG,eAAe,CAAC,IAAI,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC;YACpD,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;YACnB,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;YACzB,KAAK,GAAG,YAAY,CAAC,IAAI,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC;YAE5C,kEAAkE;YAClE,IAAI,WAAW,GAAG,IAAI,CAAC;YACvB,IAAI,CAAC,GAAG,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACtB,MAAM,QAAQ,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;gBACzC,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;oBACtB,WAAW,GAAG,QAAQ,CAAC,IAAI,CAAC;gBAC9B,CAAC;YACH,CAAC;YAED,QAAQ,MAAM,EAAE,CAAC;gBACf,KAAK,MAAM;oBACT,OAAO,GAAG,WAAW,CAAC;oBACtB,MAAM;gBACR,KAAK,MAAM;oBACT,OAAO,GAAG,UAAU,CAAC,WAAW,CAAC,CAAC;oBAClC,MAAM;gBACR,KAAK,UAAU,CAAC;gBAChB;oBACE,OAAO,GAAG,cAAc,CAAC,WAAW,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;oBAC/C,MAAM;YACV,CAAC;YAED,OAAO,GAAG,gBAAgB,CAAC,OAAO,EAAE,WAAW,CAAC,IAAI,CAAC,CAAC;QACxD,CAAC;aAAM,IAAI,MAAM,EAAE,CAAC;YAClB,+BAA+B;YAC/B,IAAI,CAAC;gBACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;gBAC5C,OAAO,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;gBAC1C,KAAK,GAAG,eAAe,CAAC;gBAExB,uCAAuC;gBACvC,MAAM,QAAQ,GAAG,4BAA4B,CAAC;gBAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;gBAC5C,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;YAC9B,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;gBAC3B,KAAK,GAAG,2BAA2B,CAAC;YACtC,CAAC;YACD,OAAO,GAAG,GAAG,CAAC,CAAC,qCAAqC;QACtD,CAAC;aAAM,IAAI,KAAK,EAAE,CAAC;YACjB,4CAA4C;YAC5C,IAAI,CAAC;gBACH,MAAM,CAAC,GAAG,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;gBAE1E,yBAAyB;gBACzB,MAAM,KAAK,GAAG,CAAC,CAAC,aAAa,CAAC,CAAC;gBAC/B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACrB,KAAK,GAAG,CAAC,CAAC,+BAA+B,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,eAAe,CAAC;oBAC7E,MAAM,SAAS,GAAa,EAAE,CAAC;oBAC/B,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;wBACnB,MAAM,SAAS,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC;wBACrD,MAAM,QAAQ,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;wBACpG,MAAM,QAAQ,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,+BAA+B,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;wBAC1F,SAAS,CAAC,IAAI,CAAC,MAAM,SAAS,KAAK,QAAQ,KAAK,QAAQ,EAAE,CAAC,CAAC;wBAC5D,IAAI,QAAQ;4BAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;oBACrC,CAAC,CAAC,CAAC;oBACH,OAAO,GAAG,KAAK,KAAK,OAAO,SAAS,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC;gBAC7D,CAAC;qBAAM,CAAC;oBACN,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;oBAC3B,KAAK,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,cAAc,CAAC;gBACtD,CAAC;YACH,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;gBAC3B,KAAK,GAAG,cAAc,CAAC;YACzB,CAAC;YACD,OAAO,GAAG,GAAG,CAAC;QAChB,CAAC;aAAM,CAAC;YACN,0CAA0C;YAC1C,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;YAC3B,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,eAAe,CAAC;YAE5D,+BAA+B;YAC/B,MAAM,QAAQ,GAAG,4BAA4B,CAAC;YAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;YAC5C,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;YAC5B,OAAO,GAAG,GAAG,CAAC;QAChB,CAAC;QAED,uCAAuC;QACvC,IAAI,SAA0C,CAAC;QAC/C,IAAI,OAAO,IAAI,MAAM,EAAE,CAAC;YACtB,IAAI,OAAO,CAAC,SAAS,IAAI,CAAC,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5D,yBAAyB;gBACzB,MAAM,EAAE,cAAc,EAAE,GAAG,MAAM,MAAM,CAAC,mBAAmB,CAAC,CAAC;gBAC7D,SAAS,GAAG,MAAM,cAAc,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YACrD,CAAC;iBAAM,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;gBAC/C,kCAAkC;gBAClC,SAAS,GAAG,iBAAiB,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;YAC3D,CAAC;QACH,CAAC;QAED,wCAAwC;QACxC,IAAI,SAAS,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YAC/B,OAAO,GAAG,qBAAqB,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;QACtD,CAAC;QAED,kDAAkD;QAClD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC;QACvC,MAAM,MAAM,GAAG,cAAc,CAAC,OAAO,CAAC,CAAC;QACvC,MAAM,WAAW,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAEpF,iDAAiD;QACjD,MAAM,gBAAgB,GAAG,WAAW,CAAC,UAAU,EAAE,QAAQ,CAAC,QAAQ,CAAC,CAAC;QAEpE,mDAAmD;QACnD,IAAI,eAAyE,CAAC;QAC9E,IAAI,OAAO,CAAC,QAAQ,IAAI,MAAM,EAAE,CAAC;YAC/B,IAAI,CAAC;gBACH,8DAA8D;gBAC9D,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,MAAM,CAAC,iBAAiB,CAAC,CAAC;gBACrD,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC;gBAC1D,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,CAAC;oBACjC,SAAS,EAAE,SAAS,IAAI,SAAS;iBAClC,CAAC,CAAC;gBAEH,sBAAsB;gBACtB,MAAM,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,EAAE;oBAC/B,SAAS,EAAE,kBAAkB;oBAC7B,OAAO,EAAE,OAAO,IAAI,KAAK;iBAC1B,CAAC,CAAC;gBAEH,oBAAoB;gBACpB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;oBACb,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;gBAClC,CAAC;gBAED,mBAAmB;gBACnB,MAAM,EAAE,eAAe,EAAE,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAAC,CAAC;gBAC/D,eAAe,GAAG,MAAM,eAAe,CAAC,IAAI,CAAC,CAAC;gBAE9C,WAAW;gBACX,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;YACxB,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CAAC,6BAA6B,EAAE,KAAK,CAAC,CAAC;YACtD,CAAC;QACH,CAAC;QAED,qCAAqC;QACrC,IAAI,YAA0E,CAAC;QAC/E,IAAI,OAAO,CAAC,cAAc,EAAE,CAAC;YAC3B,IAAI,CAAC;gBACH,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAC;gBAClE,YAAY,GAAG,MAAM,WAAW,CAAC,WAAW,CAAC,GAAG,EAAE,OAAO,EAAE,WAAW,CAAC,CAAC;YAC1E,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CAAC,yBAAyB,EAAE,KAAK,CAAC,CAAC;YAClD,CAAC;QACH,CAAC;QAED,mCAAmC;QACnC,IAAI,WAA+B,CAAC;QACpC,IAAI,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,GAAG,EAAE,CAAC;YACnC,IAAI,CAAC;gBACH,MAAM,EAAE,cAAc,EAAE,GAAG,MAAM,MAAM,CAAC,mBAAmB,CAAC,CAAC;gBAC7D,MAAM,aAAa,GAAG,OAAO,OAAO,CAAC,OAAO,KAAK,QAAQ,IAAI,OAAO,CAAC,OAAO,CAAC,MAAM;oBACjF,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM;oBACxB,CAAC,CAAC,mEAAmE,CAAC;gBACxE,MAAM,SAAS,GAAG,OAAO,OAAO,CAAC,OAAO,KAAK,QAAQ,IAAI,OAAO,CAAC,OAAO,CAAC,SAAS;oBAChF,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,SAAS;oBAC3B,CAAC,CAAC,GAAG,CAAC;gBAER,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,OAAO,EAAE;oBAC3C,MAAM,EAAE,GAAG,aAAa,kBAAkB,SAAS,SAAS;oBAC5D,SAAS,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM;oBAC7B,QAAQ,EAAE,OAAO,CAAC,GAAG,CAAC,KAAK;oBAC3B,UAAU,EAAE,OAAO,CAAC,GAAG,CAAC,OAAO;iBAChC,CAAC,CAAC;gBACH,WAAW,GAAG,MAAM,CAAC,OAAO,IAAI,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAW,CAAC;YACrE,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CAAC,4BAA4B,EAAE,KAAK,CAAC,CAAC;YACrD,CAAC;QACH,CAAC;QAED,OAAO;YACL,GAAG,EAAE,WAAW,CAAC,GAAG;YACpB,KAAK;YACL,OAAO;YACP,QAAQ;YACR,KAAK;YACL,MAAM;YACN,MAAM,EAAE,WAAW,CAAC,MAAM;YAC1B,OAAO;YACP,UAAU,EAAE,gBAAgB;YAC5B,WAAW,EAAE,YAAY;YACzB,OAAO;YACP,WAAW;YACX,SAAS;YACT,QAAQ,EAAE,eAAe;YACzB,cAAc,EAAE,YAAY;YAC5B,OAAO,EAAE,WAAW;SACrB,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,sCAAsC;QACtC,MAAM,OAAO,EAAE,CAAC;QAChB,MAAM,KAAK,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;;GAcG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,IAAc,EACd,UAAkD,EAAE;IAEpD,MAAM,EAAE,WAAW,GAAG,CAAC,EAAE,GAAG,QAAQ,EAAE,GAAG,OAAO,CAAC;IACjD,MAAM,OAAO,GAAoD,EAAE,CAAC;IAEpE,qBAAqB;IACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,EAAE,CAAC;QAClD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,CAAC;QAC7C,MAAM,YAAY,GAAG,MAAM,OAAO,CAAC,UAAU,CAC3C,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,CACtC,CAAC;QAEF,YAAY,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;YACjC,IAAI,MAAM,CAAC,MAAM,KAAK,WAAW,EAAE,CAAC;gBAClC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,IAAI,CAAC;oBACX,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;oBACb,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,IAAI,eAAe;iBACjD,CAAC,CAAC;YACL,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;GAGG;AACH,OAAO,EAAE,OAAO,EAAE,CAAC"}