webforai 1.5.1 → 1.6.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (79) hide show
  1. package/dist/bin.js +258 -278
  2. package/dist/cjs/extract-hast/index.js +4 -3
  3. package/dist/cjs/extract-hast/readability.js +28 -61
  4. package/dist/cjs/extract-hast/strongly.js +2 -1
  5. package/dist/cjs/extract-hast/utils.js +0 -15
  6. package/dist/cjs/html-to-mdast.js +6 -6
  7. package/dist/cjs/link-replacer.js +10 -5
  8. package/dist/cjs/loaders/playwright.js +36 -7
  9. package/dist/cjs/{cli/constants.js → mdast-handlers/custom-br-handler.js} +9 -12
  10. package/dist/cjs/mdast-to-markdown.js +1 -1
  11. package/dist/cjs/utils/hast-utils.js +71 -0
  12. package/dist/extract-hast/index.js +4 -3
  13. package/dist/extract-hast/readability.js +29 -62
  14. package/dist/extract-hast/strongly.js +2 -1
  15. package/dist/extract-hast/utils.js +0 -14
  16. package/dist/html-to-mdast.js +6 -6
  17. package/dist/link-replacer.js +10 -5
  18. package/dist/loaders/playwright.js +36 -7
  19. package/dist/mdast-handlers/custom-br-handler.js +8 -0
  20. package/dist/mdast-to-markdown.js +1 -1
  21. package/dist/types/cli/bin.js.map +1 -1
  22. package/dist/types/cli/commands/webforai/index.d.ts +6 -0
  23. package/dist/types/cli/commands/webforai/index.js.map +1 -0
  24. package/dist/types/cli/commands/webforai/loadHtml.d.ts +3 -0
  25. package/dist/types/cli/commands/webforai/loadHtml.js.map +1 -0
  26. package/dist/types/extract-hast/index.d.ts +9 -4
  27. package/dist/types/extract-hast/index.js.map +1 -1
  28. package/dist/types/extract-hast/readability.d.ts +2 -1
  29. package/dist/types/extract-hast/readability.js.map +1 -1
  30. package/dist/types/extract-hast/strongly.d.ts +2 -1
  31. package/dist/types/extract-hast/strongly.js.map +1 -1
  32. package/dist/types/extract-hast/utils.d.ts +0 -1
  33. package/dist/types/extract-hast/utils.js.map +1 -1
  34. package/dist/types/html-to-mdast.d.ts +4 -0
  35. package/dist/types/html-to-mdast.js.map +1 -1
  36. package/dist/types/link-replacer.d.ts +1 -1
  37. package/dist/types/link-replacer.js.map +1 -1
  38. package/dist/types/loaders/playwright.d.ts +8 -2
  39. package/dist/types/loaders/playwright.js.map +1 -1
  40. package/dist/types/mdast-handlers/custom-br-handler.d.ts +2 -0
  41. package/dist/types/mdast-handlers/custom-br-handler.js.map +1 -0
  42. package/dist/types/mdast-to-markdown.js.map +1 -1
  43. package/dist/types/utils/hast-utils.d.ts +4 -0
  44. package/dist/types/utils/hast-utils.js.map +1 -0
  45. package/dist/utils/hast-utils.js +46 -0
  46. package/package.json +22 -10
  47. package/dist/cjs/cli/bin.js +0 -34
  48. package/dist/cjs/cli/commands/handlers/index.js +0 -84
  49. package/dist/cjs/cli/commands/handlers/loadHtml.js +0 -127
  50. package/dist/cjs/cli/helpers/assertContinue.js +0 -34
  51. package/dist/cjs/cli/helpers/inputOutputPath.js +0 -89
  52. package/dist/cjs/cli/helpers/inputSourcePath.js +0 -64
  53. package/dist/cjs/cli/helpers/selectExtractMode.js +0 -39
  54. package/dist/cjs/cli/helpers/selectLoader.js +0 -44
  55. package/dist/cjs/cli/utils.js +0 -122
  56. package/dist/cjs/html-to-markdown.test.js +0 -118
  57. package/dist/cjs/link-replacer.test.js +0 -47
  58. package/dist/cjs/loaders/fetch.test.js +0 -9
  59. package/dist/cjs/loaders/playwright.test.js +0 -17
  60. package/dist/cjs/loaders/puppeteer.test.js +0 -17
  61. package/dist/cli/bin.js +0 -12
  62. package/dist/cli/commands/handlers/index.js +0 -51
  63. package/dist/cli/commands/handlers/loadHtml.js +0 -94
  64. package/dist/cli/constants.js +0 -9
  65. package/dist/cli/helpers/assertContinue.js +0 -11
  66. package/dist/cli/helpers/inputOutputPath.js +0 -56
  67. package/dist/cli/helpers/inputSourcePath.js +0 -31
  68. package/dist/cli/helpers/selectExtractMode.js +0 -16
  69. package/dist/cli/helpers/selectLoader.js +0 -21
  70. package/dist/cli/utils.js +0 -85
  71. package/dist/html-to-markdown.test.js +0 -118
  72. package/dist/link-replacer.test.js +0 -47
  73. package/dist/loaders/fetch.test.js +0 -9
  74. package/dist/loaders/playwright.test.js +0 -17
  75. package/dist/loaders/puppeteer.test.js +0 -17
  76. package/dist/types/cli/commands/handlers/index.d.ts +0 -13
  77. package/dist/types/cli/commands/handlers/index.js.map +0 -1
  78. package/dist/types/cli/commands/handlers/loadHtml.d.ts +0 -2
  79. package/dist/types/cli/commands/handlers/loadHtml.js.map +0 -1
@@ -12,22 +12,8 @@ var isStrInclude = (value, match) => {
12
12
  }
13
13
  return false;
14
14
  };
15
- var hasAncestors = (element, tagNames, depth = 3) => {
16
- let parent = element.parent;
17
- for (let i = 0; i < depth; i++) {
18
- if (!parent || parent.type !== "element") {
19
- return false;
20
- }
21
- if (tagNames.includes(parent.tagName)) {
22
- return true;
23
- }
24
- parent = parent.parent;
25
- }
26
- return false;
27
- };
28
15
  export {
29
16
  classnames,
30
- hasAncestors,
31
17
  isStrInclude,
32
18
  matchString
33
19
  };
@@ -8,12 +8,13 @@ import { customCodeHandler } from "./mdast-handlers/custom-code-handler.js";
8
8
  import { customDivHandler } from "./mdast-handlers/custom-div-handler.js";
9
9
  import { customImgHandler } from "./mdast-handlers/custom-img-handler.js";
10
10
  import { customTableHandler } from "./mdast-handlers/custom-table-handler.js";
11
- import { emptyHandler } from "./mdast-handlers/empty-handler.js";
12
11
  import { mathHandler } from "./mdast-handlers/math-handler.js";
12
+ import { getLangFromHast, getLangFromStr, getUrlFromHast } from "./utils/hast-utils.js";
13
13
  var htmlToMdast = (htmlOrHast, options) => {
14
- const { extractors } = options || {};
15
- const hast = typeof htmlOrHast === "string" ? fromHtml(htmlOrHast, { fragment: true }) : htmlOrHast;
16
- const extractedHast = extractHast(hast, extractors);
14
+ const { extractors, url: defaultUrl, lang: defaultLang } = options || {};
15
+ const [lang, hast] = typeof htmlOrHast === "string" ? [defaultLang || getLangFromStr(htmlOrHast), fromHtml(htmlOrHast, { fragment: true })] : [defaultLang || getLangFromHast(htmlOrHast), htmlOrHast];
16
+ const url = defaultUrl || getUrlFromHast(hast);
17
+ const extractedHast = extractHast({ hast, lang, url }, extractors);
17
18
  const mdast = toMdast(extractedHast, {
18
19
  handlers: {
19
20
  math: mathHandler,
@@ -21,8 +22,7 @@ var htmlToMdast = (htmlOrHast, options) => {
21
22
  pre: customCodeHandler,
22
23
  a: customAHandler({ asText: options?.linkAsText }),
23
24
  img: customImgHandler({ hideImage: options?.hideImage }),
24
- table: customTableHandler({ asText: options?.tableAsText }),
25
- br: emptyHandler
25
+ table: customTableHandler({ asText: options?.tableAsText })
26
26
  }
27
27
  });
28
28
  const extractedMdast = extractMdast(mdast);
@@ -1,11 +1,16 @@
1
1
  // src/link-replacer.ts
2
- var linkReplacer = (markdown, link) => {
3
- const regex = /(!?\[.*?\]\()(.+?)(\))/g;
4
- return markdown.replace(regex, (match, pre, thisUrl, post) => {
5
- if (thisUrl.startsWith("http") || thisUrl.startsWith("#")) {
2
+ var linkReplacer = (markdown, base) => {
3
+ const regex = /(!?\[.*?\]\()([^)\s]+)(\))/g;
4
+ return markdown.replace(regex, (match, pre, url, post) => {
5
+ if (/^(https?:|#)/.test(url)) {
6
+ return match;
7
+ }
8
+ try {
9
+ const absoluteUrl = new URL(url, base).href;
10
+ return `${pre}${absoluteUrl}${post}`;
11
+ } catch {
6
12
  return match;
7
13
  }
8
- return `${pre}${new URL(thisUrl, link).href}${post}`;
9
14
  });
10
15
  };
11
16
  export {
@@ -1,13 +1,42 @@
1
1
  // src/loaders/playwright.ts
2
- import { chromium } from "playwright";
3
- var loadHtml = async (url, context) => {
4
- const _context = context ?? await chromium.launch({ headless: true });
5
- const page = await _context.newPage();
6
- await page.goto(url, { waitUntil: "load" });
2
+ import { chromium, devices } from "playwright-core";
3
+ var SUPER_BYPASS_DEVICE = {
4
+ userAgent: "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36",
5
+ viewport: { width: 1920, height: 1080 },
6
+ deviceScaleFactor: 1,
7
+ hasTouch: false,
8
+ isMobile: false,
9
+ javaScriptEnabled: true,
10
+ locale: "en-US",
11
+ timezoneId: "America/New_York"
12
+ };
13
+ var loadHtml = async (url, options) => {
14
+ const { browser, waitUntil, timeout, superBypassMode } = options ?? {};
15
+ const _browser = browser ?? await chromium.launch({ headless: true });
16
+ const context = await _browser.newContext(superBypassMode ? SUPER_BYPASS_DEVICE : devices["Desktop Chrome"]);
17
+ if (superBypassMode) {
18
+ await context.addInitScript(() => {
19
+ Object.defineProperty(navigator, "webdriver", { get: () => void 0 });
20
+ Object.defineProperty(navigator, "languages", { get: () => ["en-US", "en"] });
21
+ Object.defineProperty(navigator, "plugins", { get: () => [1, 2, 3, 4, 5] });
22
+ });
23
+ }
24
+ const page = await context.newPage();
25
+ if (superBypassMode) {
26
+ await page.route("**/*.js", (route) => {
27
+ if (route.request().url().includes("captcha-delivery")) {
28
+ return route.abort();
29
+ }
30
+ return route.continue();
31
+ });
32
+ }
33
+ await page.goto(url, { waitUntil: waitUntil ?? "load", timeout });
7
34
  const html = await page.content();
8
35
  await page.close();
9
- if (!context) {
10
- await _context.close();
36
+ if (browser) {
37
+ await context.close();
38
+ } else {
39
+ await _browser.close();
11
40
  }
12
41
  return html;
13
42
  };
@@ -0,0 +1,8 @@
1
+ // src/mdast-handlers/custom-br-handler.ts
2
+ import { defaultHandlers } from "hast-util-to-mdast";
3
+ var customBrHandler = (state, node) => {
4
+ return defaultHandlers.br(state, node);
5
+ };
6
+ export {
7
+ customBrHandler
8
+ };
@@ -10,7 +10,7 @@ var DEFAULT_MDAST_TO_MARKDOWN_OPTIONS = {
10
10
  };
11
11
  var mdastToMarkdown = (mdast, options) => {
12
12
  const { baseUrl, ...toMarkdownOptions } = { ...DEFAULT_MDAST_TO_MARKDOWN_OPTIONS, ...options };
13
- let markdown = toMarkdown(warpRoot(mdast), toMarkdownOptions);
13
+ let markdown = toMarkdown(warpRoot(mdast), toMarkdownOptions).replace(/\*\*\*\*/g, "");
14
14
  if (baseUrl) {
15
15
  markdown = linkReplacer(markdown, baseUrl);
16
16
  }
@@ -1 +1 @@
1
- {"version":3,"file":"bin.js","sourceRoot":"","sources":["../../../src/cli/bin.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,WAAW,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAE7C,OAAO;KACL,IAAI,CAAC,UAAU,CAAC;KAChB,WAAW,CAAC,wDAAwD,CAAC;KACrE,OAAO,CAAC,WAAW,CAAC,OAAO,EAAE,eAAe,EAAE,4BAA4B,CAAC;KAC3E,QAAQ,CAAC,UAAU,EAAE,wBAAwB,CAAC;KAC9C,QAAQ,CAAC,cAAc,EAAE,kCAAkC,CAAC;KAC5D,MAAM,CAAC,eAAe,EAAE,oBAAoB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KAChE,MAAM,CAAC,mBAAmB,EAAE,kBAAkB,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KACpE,MAAM,CAAC,qBAAqB,EAAE,6BAA6B,CAAC;KAC5D,MAAM,CAAC,cAAc,EAAE,kBAAkB,CAAC;KAC1C,MAAM,CAAC,aAAa,EAAE,oCAAoC,CAAC;KAC3D,EAAE,CAAC,cAAc,EAAE,GAAG,EAAE;IACxB,OAAO,CAAC,GAAG,CAAC,KAAK,GAAG,MAAM,CAAC;AAC5B,CAAC,CAAC;KACD,MAAM,CAAC,eAAe,CAAC,CAAC;AAE1B,KAAK,CAAC,EAAE,CAAC,OAAO,CAAC,wBAAwB,WAAW,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;AAEjE,OAAO,CAAC,KAAK,EAAE,CAAC"}
1
+ {"version":3,"file":"bin.js","sourceRoot":"","sources":["../../../src/cli/bin.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,WAAW,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAE7C,OAAO;KACL,IAAI,CAAC,UAAU,CAAC;KAChB,WAAW,CAAC,wDAAwD,CAAC;KACrE,OAAO,CAAC,WAAW,CAAC,OAAO,EAAE,eAAe,EAAE,4BAA4B,CAAC,CAAC;AAE9E,OAAO;KACL,QAAQ,CAAC,UAAU,EAAE,wBAAwB,CAAC;KAC9C,MAAM,CAAC,uBAAuB,EAAE,kCAAkC,CAAC;KACnE,MAAM,CAAC,mBAAmB,EAAE,oBAAoB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KACpE,MAAM,CAAC,uBAAuB,EAAE,kBAAkB,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KACxE,MAAM,CAAC,aAAa,EAAE,oCAAoC,CAAC;KAC3D,MAAM,CAAC,eAAe,CAAC,CAAC;AAE1B,OAAO,CAAC,KAAK,EAAE,CAAC"}
@@ -0,0 +1,6 @@
1
+ export declare const webforaiCommand: (initialPath: string, options: {
2
+ output?: string;
3
+ mode?: string;
4
+ loader?: string;
5
+ debug?: boolean;
6
+ }) => Promise<void>;
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AAC5D,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,WAAW,MAAM,0BAA0B,CAAC;AACnD,OAAO,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC3D,OAAO,EAAE,eAAe,EAAE,MAAM,+BAA+B,CAAC;AAChE,OAAO,EAAE,eAAe,EAAE,MAAM,+BAA+B,CAAC;AAChE,OAAO,EAAE,iBAAiB,EAAE,MAAM,iCAAiC,CAAC;AACpE,OAAO,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAC1D,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AACpC,OAAO,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAC;AAEtC,MAAM,aAAa,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC;AAC/E,MAAM,sBAAsB,GAAG,EAAE,UAAU,EAAE,KAAK,EAAE,WAAW,EAAE,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;AAE3F,MAAM,CAAC,MAAM,eAAe,GAAG,KAAK,EACnC,WAAmB,EACnB,OAA6E,EAC5E,EAAE;IACH,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,wBAAwB,WAAW,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC;IAExE,MAAM,UAAU,GAAG,WAAW,IAAI,CAAC,MAAM,eAAe,EAAE,CAAC,CAAC;IAC5D,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,eAAe,UAAU,EAAE,CAAC,CAAC;IAEvD,MAAM,MAAM,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,IAAI,CAAC,MAAM,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;IACtF,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,WAAW,MAAM,EAAE,CAAC,CAAC;IAE/C,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,IAAI,CAAC,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC,CAAC;IACzE,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,eAAe,UAAU,EAAE,CAAC,CAAC;IAEvD,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,CAAC,MAAM,iBAAiB,EAAE,CAAC,CAAC;IACzD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC;IAE3C,IAAI,IAAY,CAAC;IACjB,MAAM,CAAC,GAAG,OAAO,EAAE,CAAC;IACpB,IAAI,CAAC;QACJ,CAAC,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;QAC9B,IAAI,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,MAAM,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC;QACpE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,CAAC;IACrC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QAChB,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,CAAC,yBAAyB,CAAC,CAAC,CAAC;QAC1C,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACrB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACjB,CAAC;IACD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC;IAE3C,MAAM,QAAQ,GAAG,cAAc,CAAC,IAAI,EAAE;QACrC,OAAO,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS;QACnD,GAAG,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,sBAAsB,CAAC;KAC3D,CAAC,CAAC;IACH,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,aAAa,QAAQ,EAAE,CAAC,CAAC;IAEnD,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;IAC3C,MAAM,iBAAiB,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;IAEtF,IAAI,CAAC,iBAAiB,EAAE,CAAC;QACxB,MAAM,EAAE,CAAC,KAAK,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAChD,CAAC;IACD,MAAM,EAAE,CAAC,SAAS,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAEzC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,sBAAsB,UAAU,EAAE,CAAC,CAAC,CAAC;AACxE,CAAC,CAAC"}
@@ -0,0 +1,3 @@
1
+ export declare const loadHtml: (sourcePath: string, loader: string, options: {
2
+ debug?: boolean;
3
+ }) => Promise<string>;
@@ -0,0 +1 @@
1
+ {"version":3,"file":"loadHtml.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/loadHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,EAAE,GAAG,EAAE,MAAM,gBAAgB,CAAC;AACrC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,EAAE,QAAQ,EAAE,MAAM,iBAAiB,CAAC;AAC3C,OAAO,EAAE,QAAQ,IAAI,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACrE,OAAO,EAAE,QAAQ,IAAI,oBAAoB,EAAE,MAAM,6BAA6B,CAAC;AAE/E,MAAM,wBAAwB,GAAG,KAAK,IAAI,EAAE;IAC3C,MAAM,IAAI,GAAG,QAAQ,CAAC,cAAc,EAAE,CAAC;IACvC,MAAM,WAAW,GAAG,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAC7D,OAAO,CAAC,CAAC,WAAW,CAAC;AACtB,CAAC,CAAC;AAEF,MAAM,oBAAoB,GAAG,KAAK,IAAI,EAAE;IACvC,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,8BAA8B,CAAC,CAAC;IACvE,MAAM,aAAa,GAAG,MAAM,EAAE;SAC5B,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,EAAE,OAAO,CAAC;SAC9C,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC,CAAC;SACzC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC;IACpB,OAAO,aAAa,EAAE,OAAO,CAAC;AAC/B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,UAAkB,EAAE,MAAc,EAAE,OAA4B,EAAE,EAAE;IAClG,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QACzE,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,0BAA0B,UAAU,EAAE,CAAC,CAAC;QAClE,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC;QAClD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,YAAY,EAAE,CAAC;QAC7B,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QACzE,MAAM,qBAAqB,GAAG,MAAM,wBAAwB,EAAE,CAAC;QAC/D,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,yBAAyB,qBAAqB,EAAE,CAAC,CAAC;QAE5E,MAAM,SAAS,GAAG,MAAM,oBAAoB,EAAE,CAAC;QAE/C,IAAI,CAAC,qBAAqB,EAAE,CAAC;YAC5B,MAAM,OAAO,GAAG;gBACf,EAAE,CAAC,IAAI,CAAC,oCAAoC,CAAC;gBAC7C,EAAE;gBACF,qEAAqE;gBACrE,EAAE;gBACF,oBAAoB,SAAS,mBAAmB;gBAChD,EAAE;gBACF,6CAA6C;gBAC7C,2HAA2H;gBAC3H,qCAAqC;gBACrC,EAAE;gBACF,iDAAiD;gBACjD,0DAA0D;gBAC1D,6DAA6D;gBAC7D,EAAE;gBACF,yBAAyB,SAAS,eAAe;aACjD,CAAC;YAEF,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,WAAW,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC;YAC3E,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;QAChD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,oBAAoB,CAAC,UAAU,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,uBAAuB,MAAM,EAAE,CAAC,CAAC;AAClD,CAAC,CAAC"}
@@ -1,10 +1,15 @@
1
1
  import type { Nodes as Hast } from "hast";
2
2
  export type ExtractPresets = "strongly" | "readability";
3
- export type Extractor = ((hast: Hast) => Hast) | false | ExtractPresets;
3
+ export type ExtractParams = {
4
+ hast: Hast;
5
+ lang?: string;
6
+ url?: string;
7
+ };
8
+ export type Extractor = ((param: ExtractParams) => Hast) | false | ExtractPresets;
4
9
  export type Extracotrs = Extractor | Extractor[];
5
10
  export declare const PRESET_EXTRACT_HAST: {
6
- strongly: (hast: Hast) => Hast;
7
- readability: (hast: Hast) => Hast;
11
+ strongly: (params: ExtractParams) => Hast;
12
+ readability: (params: ExtractParams) => Hast;
8
13
  };
9
14
  export declare const DEFAULT_EXTRACT_HAST: Extracotrs;
10
- export declare const extractHast: (hast: Hast, extractors?: Extracotrs) => Hast;
15
+ export declare const extractHast: (params: ExtractParams, extractors?: Extracotrs) => Hast;
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extract-hast/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAC;AAMjD,MAAM,CAAC,MAAM,mBAAmB,GAAG;IAClC,QAAQ,EAAE,mBAAmB;IAC7B,WAAW,EAAE,sBAAsB;CACnC,CAAC;AAEF,MAAM,CAAC,MAAM,oBAAoB,GAAe,CAAC,aAAa,CAAC,CAAC;AAEhE,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,IAAU,EAAE,aAAyB,oBAAoB,EAAQ,EAAE;IAC9F,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAE1E,MAAM,SAAS,GACd,WAAW,CAAC,MAAM,CAAO,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE;QAC3C,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACzB,OAAO,GAAG,CAAC;QACZ,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,IAAI,mBAAmB,EAAE,CAAC;YACvE,OAAO,mBAAmB,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC;QAC5C,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;YACrC,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC;QACvB,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,sBAAsB,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAElB,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extract-hast/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAC;AAOjD,MAAM,CAAC,MAAM,mBAAmB,GAAG;IAClC,QAAQ,EAAE,mBAAmB;IAC7B,WAAW,EAAE,sBAAsB;CACnC,CAAC;AAEF,MAAM,CAAC,MAAM,oBAAoB,GAAe,CAAC,aAAa,CAAC,CAAC;AAEhE,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,MAAqB,EAAE,aAAyB,oBAAoB,EAAQ,EAAE;IACzG,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IAC9B,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAE1E,MAAM,SAAS,GACd,WAAW,CAAC,MAAM,CAAO,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE;QAC3C,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACzB,OAAO,GAAG,CAAC;QACZ,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,IAAI,mBAAmB,EAAE,CAAC;YACvE,OAAO,mBAAmB,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QAC5D,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;YACrC,OAAO,SAAS,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,sBAAsB,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAElB,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
@@ -1,2 +1,3 @@
1
1
  import type { Nodes as Hast } from "hast";
2
- export declare const readabilityExtractHast: (hast: Hast) => Hast;
2
+ import type { ExtractParams } from ".";
3
+ export declare const readabilityExtractHast: (params: ExtractParams) => Hast;
@@ -1 +1 @@
1
- {"version":3,"file":"readability.js","sourceRoot":"","sources":["../../../src/extract-hast/readability.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAC3C,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAE9E,MAAM,cAAc,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;AAE5G,MAAM,OAAO,GAAG;IACf,MAAM,EAAE,kCAAkC;IAC1C,MAAM,EAAE,4CAA4C;IACpD,yBAAyB,EAAE,8BAA8B;IACzD,kBAAkB,EACjB,yQAAyQ;IAC1Q,oBAAoB,EAAE,mDAAmD;CACzE,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;AAE/G,MAAM,cAAc,GAAG,CAAC,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;AAEpF,MAAM,mBAAmB,GAAG;IAC3B,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;CACP,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,IAAU,EAAE,EAAE;IACjC,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO;IACR,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAChC,IAAI,OAAO,CAAC,OAAO,KAAK,MAAM,EAAE,CAAC;QAChC,OAAO;IACR,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,OAAO,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC;IAC3E,IAAI,QAAQ,EAAE,CAAC;QACd,OAAO,QAAkB,CAAC;IAC3B,CAAC;AACF,CAAC,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,IAAU,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAEnF,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,8BAA8B;IAC9B,IAAI,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yCAAyC;IACzC,IAAI,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACtE,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QAC7E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QAC9E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,KAAK,QAAQ,IAAI,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAChG,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,KAAK,CAAC;IACd,CAAC;IAED,wBAAwB;IACxB,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,CAAC,IAAc,CAAC,EAAE,CAAC;QAC3F,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,qBAAqB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC5C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,6BAA6B;IAC7B,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC3E,OAAO,IAAI,CAAC;IACb,CAAC;IACD,IAAI,YAAY,CAAC,OAAO,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACjD,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,KAAK,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;IAEnC,IAAI,OAAO,CAAC,yBAAyB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,6BAA6B;IAC7B,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzF,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE;IACxC,MAAM,CAAC,GAAG,MAAM,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IAC/B,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IAE7B,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;QACjB,OAAO,KAAK,CAAC;IACd,CAAC;IAED,MAAM,WAAW,GAAG,GAAG,CAAC,UAAU,CAAC,GAAG,EAAE,QAAQ,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;IACpE,MAAM,YAAY,GAAG,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;IAEpE,OAAO,WAAW,KAAK,YAAY,IAAI,WAAW,CAAC;AACpD,CAAC,CAAC;AAEF,MAAM,iBAAiB,GAAG,CAAC,IAAU,EAAE,EAAE;IACxC,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,CAAC,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/C,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1B,OAAO,IAAI,CAAC;IACb,CAAC;IAED,MAAM,IAAI,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;IACnC,IAAI,IAAI,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QACtB,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAQ,EAAE;IAC1D,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,UAAU,CAAC,IAAI,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,CAAC;IACvF,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAC1C,MAAM,QAAQ,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAEpC,MAAM,WAAW,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE;QACzC,IAAI,CAAC,cAAc,CAAC,IAAY,CAAC,EAAE,CAAC;YACnC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,sBAAsB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC3C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,WAAW,EAAE,CAAC;QAClB,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IACvC,CAAC;IAED,MAAM,eAAe,GAAG,YAAY,CAAC,WAAW,CAAC,CAAC;IAClD,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC,GACvB,eAAe,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,GAAG,IAAI;QAC5E,CAAC,CAAE,CAAC,WAAmB,EAAE,eAAe,CAAW;QACnD,CAAC,CAAE,CAAC,IAAY,EAAE,QAAQ,CAAW,CAAC;IAExC,IAAI,aAAa,GAAG,IAAI,IAAI,mBAAmB,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAwC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;IACtH,IAAI,QAAQ,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QACrC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IAED,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,EAAE,IAAI,EAAE,MAAe,EAAE,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,WAAW,CAAC,EAAE,CAAC;QACtF,MAAM,WAAW,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;QAE1C,IAAI,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC7B,SAAS;QACV,CAAC;QAED,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC;QACtD,IAAI,WAAW,GAAG,GAAG,EAAE,CAAC;YACvB,SAAS;QACV,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;YACxC,QAAQ,GAAG,OAAO,CAAC;YACnB,QAAQ,GAAG,WAAW,CAAC;YACvB,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,EAAE;QACnD,IAAI,CAAC,iBAAiB,CAAC,IAAY,CAAC,EAAE,CAAC;YACtC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,qBAAqB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QAED,OAAO,IAAI,CAAC;IACb,CAAC,CAAS,CAAC;IAEX,MAAM,SAAS,GAAG,YAAY,CAAC,iBAAiB,CAAC,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,QAAQ,CAAC;IAC9G,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
1
+ {"version":3,"file":"readability.js","sourceRoot":"","sources":["../../../src/extract-hast/readability.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAEhE,MAAM,cAAc,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;AAE5G,MAAM,OAAO,GAAG;IACf,MAAM,EAAE,kCAAkC;IAC1C,MAAM,EAAE,4CAA4C;IACpD,yBAAyB,EAAE,8BAA8B;IACzD,kBAAkB,EACjB,wRAAwR;IACzR,oBAAoB,EAAE,mDAAmD;CACzE,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;AAEvG,MAAM,cAAc,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;AAEzF,MAAM,mBAAmB,GAAG;IAC3B,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;CACP,CAAC;AAEF,yBAAyB;AACzB,UAAU;AACV,WAAW;AACX,KAAK;AAEL,MAAM,cAAc,GAAG,CAAC,IAAU,EAAE,EAAE;IACrC,OAAO,CAAC,CACP,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QAC1C,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CACnH,CAAC;AACH,CAAC,CAAC;AAEF,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAChD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yCAAyC;IACzC,IAAI,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACtE,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QAC7E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QAC9E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,KAAK,QAAQ,IAAI,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAChG,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,KAAK,CAAC;IACd,CAAC;IAED,wBAAwB;IACxB,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,CAAC,IAAc,CAAC,EAAE,CAAC;QAC3F,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,qBAAqB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC5C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,6BAA6B;IAC7B,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/G,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,KAAK,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;IAEnC,IAAI,OAAO,CAAC,yBAAyB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,6BAA6B;IAC7B,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzF,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,iBAAiB,GAAG,CAAC,IAAU,EAAE,KAAa,EAAE,EAAE;IACvD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IACrE,MAAM,EAAE,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,MAAM,CAAC;IACrC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAE1C,MAAM,oBAAoB,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAY,CAAC,CAAC,CAAC;IAClF,MAAM,wBAAwB,GAAG,oBAAoB,IAAI,YAAY,CAAC,oBAAoB,CAAC,CAAC;IAC5F,IAAI,CAAC,CAAC,oBAAoB,IAAI,wBAAwB,CAAC,EAAE,CAAC;QACzD,OAAO,IAAI,CAAC;IACb,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,CAAC,oBAAoB,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,sBAAsB,CAAC,IAAY,CAAC,CAAC,CAAC;IACjG,MAAM,eAAe,GAAG,WAAW,CAAC,CAAC,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACrE,MAAM,CAAC,QAAQ,EAAE,QAAQ,CAAC,GACzB,eAAe,CAAC,MAAM,GAAG,wBAAwB,CAAC,MAAM,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,GAAG,IAAI;QAC5F,CAAC,CAAE,CAAC,WAAmB,EAAE,eAAe,CAAW;QACnD,CAAC,CAAE,CAAC,oBAA4B,EAAE,wBAAwB,CAAW,CAAC;IAExE,IAAI,aAAa,GAAG,IAAI,IAAI,mBAAmB,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAwC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;IACtH,IAAI,QAAQ,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QACrC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IAED,IAAI,aAAa,GAAS,QAAQ,CAAC;IACnC,IAAI,aAAa,GAAG,QAAQ,CAAC;IAE7B,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,EAAE,IAAI,EAAE,MAAe,EAAE,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,WAAW,CAAC,EAAE,CAAC;QACtF,MAAM,WAAW,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;QAE1C,IAAI,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC7B,SAAS;QACV,CAAC;QAED,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC;QACzD,IAAI,WAAW,GAAG,GAAG,EAAE,CAAC;YACvB,SAAS;QACV,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;YACxC,aAAa,GAAG,OAAO,CAAC;YACxB,aAAa,GAAG,WAAW,CAAC;YAC5B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,aAAa,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,CAAC,iBAAiB,CAAC,IAAY,EAAE,IAAI,CAAC,EAAE,CAAC;YAC5C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,qBAAqB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QAED,OAAO,IAAI,CAAC;IACb,CAAC,CAAS,CAAC;IAEX,MAAM,SAAS,GACd,YAAY,CAAC,iBAAiB,CAAC,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,aAAa,CAAC;IACvG,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
@@ -1,2 +1,3 @@
1
1
  import type { Nodes as Hast } from "hast";
2
- export declare const stronglyExtractHast: (hast: Hast) => Hast;
2
+ import type { ExtractParams } from ".";
3
+ export declare const stronglyExtractHast: (params: ExtractParams) => Hast;
@@ -1 +1 @@
1
- {"version":3,"file":"strongly.js","sourceRoot":"","sources":["../../../src/extract-hast/strongly.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,MAAM,YAAY,GACjB,iPAAiP,CAAC;AAEnP,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,IAAU,EAAQ,EAAE;IACvD,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,eAAe,CAAC,CAAC;IAC/E,IAAI,eAAe,GAAG,IAAI,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;QAClC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACvC,IAAI,OAAO,EAAE,CAAC;YACb,eAAe,GAAG,OAAO,CAAC;YAC1B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,eAAe,GAAG,MAAM,CAAC,eAAe,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC;QACb,CAAC;QACD,MAAM,OAAO,GAAG,IAAe,CAAC;QAChC,MAAM,WAAW,GAAG,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,SAAS,EAAE,CAAC;QAClG,IAAI,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC;YACpC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC,CAAC,CAAC;IAEH,OAAO,eAAuB,CAAC;AAChC,CAAC,CAAC"}
1
+ {"version":3,"file":"strongly.js","sourceRoot":"","sources":["../../../src/extract-hast/strongly.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAG3C,MAAM,YAAY,GACjB,iPAAiP,CAAC;AAEnP,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IAClE,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IACxB,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,eAAe,CAAC,CAAC;IAC/E,IAAI,eAAe,GAAG,IAAI,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;QAClC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACvC,IAAI,OAAO,EAAE,CAAC;YACb,eAAe,GAAG,OAAO,CAAC;YAC1B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,eAAe,GAAG,MAAM,CAAC,eAAe,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC;QACb,CAAC;QACD,MAAM,OAAO,GAAG,IAAe,CAAC;QAChC,MAAM,WAAW,GAAG,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,SAAS,EAAE,CAAC;QAClG,IAAI,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC;YACpC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC,CAAC,CAAC;IAEH,OAAO,eAAuB,CAAC;AAChC,CAAC,CAAC"}
@@ -2,4 +2,3 @@ import type { Element } from "hast";
2
2
  export declare const matchString: (element: Element) => string;
3
3
  export declare const classnames: (element: Element) => string[];
4
4
  export declare const isStrInclude: (value: unknown, match: string) => boolean;
5
- export declare const hasAncestors: (element: Element, tagNames: string[], depth?: number) => boolean;
@@ -1 +1 @@
1
- {"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/extract-hast/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,OAAgB,EAAE,QAAkB,EAAE,KAAK,GAAG,CAAC,EAAE,EAAE;IAC/E,IAAI,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAChC,IAAI,CAAC,MAAM,IAAI,MAAM,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC;YACvC,OAAO,IAAI,CAAC;QACb,CAAC;QACD,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;IACxB,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC"}
1
+ {"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/extract-hast/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC"}
@@ -13,6 +13,10 @@ export type HtmlToMdastOptions = {
13
13
  tableAsText?: boolean;
14
14
  /** Whether to hide images. */
15
15
  hideImage?: boolean;
16
+ /** The language of the HTML. */
17
+ lang?: string;
18
+ /** The URL of the HTML. */
19
+ url?: string;
16
20
  };
17
21
  /**
18
22
  * Converts an HTML string or HAST tree to an MDAST tree.
@@ -1 +1 @@
1
- {"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,EAAmB,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACnE,OAAO,EAAE,iBAAiB,EAAE,MAAM,sCAAsC,CAAC;AACzE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uCAAuC,CAAC;AAC3E,OAAO,EAAE,YAAY,EAAE,MAAM,gCAAgC,CAAC;AAC9D,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAgB5D;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,UAAyB,EAAE,OAA4B,EAAS,EAAE;IAC7F,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IAErC,MAAM,IAAI,GAAG,OAAO,UAAU,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;IAEpG,MAAM,aAAa,GAAG,WAAW,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;IAEpD,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,EAAE;QACpC,QAAQ,EAAE;YACT,IAAI,EAAE,WAAW;YACjB,GAAG,EAAE,gBAAgB;YACrB,GAAG,EAAE,iBAAiB;YACtB,CAAC,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC;YAClD,GAAG,EAAE,gBAAgB,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;YACxD,KAAK,EAAE,kBAAkB,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;YAC3D,EAAE,EAAE,YAAY;SAChB;KACD,CAAC,CAAC;IAEH,MAAM,cAAc,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAE3C,OAAO,cAAc,CAAC;AACvB,CAAC,CAAC"}
1
+ {"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,EAAmB,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACnE,OAAO,EAAE,iBAAiB,EAAE,MAAM,sCAAsC,CAAC;AACzE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uCAAuC,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AAoBrF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,UAAyB,EAAE,OAA4B,EAAS,EAAE;IAC7F,MAAM,EAAE,UAAU,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,WAAW,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IAEzE,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GACjB,OAAO,UAAU,KAAK,QAAQ;QAC7B,CAAC,CAAC,CAAC,WAAW,IAAI,cAAc,CAAC,UAAU,CAAC,EAAE,QAAQ,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC;QACvF,CAAC,CAAC,CAAC,WAAW,IAAI,eAAe,CAAC,UAAU,CAAC,EAAE,UAAU,CAAC,CAAC;IAE7D,MAAM,GAAG,GAAG,UAAU,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC;IAE/C,MAAM,aAAa,GAAG,WAAW,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,UAAU,CAAC,CAAC;IAEnE,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,EAAE;QACpC,QAAQ,EAAE;YACT,IAAI,EAAE,WAAW;YACjB,GAAG,EAAE,gBAAgB;YACrB,GAAG,EAAE,iBAAiB;YACtB,CAAC,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC;YAClD,GAAG,EAAE,gBAAgB,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;YACxD,KAAK,EAAE,kBAAkB,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;SAC3D;KACD,CAAC,CAAC;IAEH,MAAM,cAAc,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAE3C,OAAO,cAAc,CAAC;AACvB,CAAC,CAAC"}
@@ -1 +1 @@
1
- export declare const linkReplacer: (markdown: string, link: string) => string;
1
+ export declare const linkReplacer: (markdown: string, base: string) => string;
@@ -1 +1 @@
1
- {"version":3,"file":"link-replacer.js","sourceRoot":"","sources":["../../src/link-replacer.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,QAAgB,EAAE,IAAY,EAAE,EAAE;IAC9D,MAAM,KAAK,GAAG,yBAAyB,CAAC;IACxC,OAAO,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE;QAC5D,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YAC3D,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,GAAG,GAAG,GAAG,IAAI,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,IAAI,GAAG,IAAI,EAAE,CAAC;IACtD,CAAC,CAAC,CAAC;AACJ,CAAC,CAAC"}
1
+ {"version":3,"file":"link-replacer.js","sourceRoot":"","sources":["../../src/link-replacer.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,QAAgB,EAAE,IAAY,EAAE,EAAE;IAC9D,MAAM,KAAK,GAAG,6BAA6B,CAAC;IAC5C,OAAO,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,EAAE;QACxD,IAAI,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAC9B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC;YACJ,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC;YAC5C,OAAO,GAAG,GAAG,GAAG,WAAW,GAAG,IAAI,EAAE,CAAC;QACtC,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;IACF,CAAC,CAAC,CAAC;AACJ,CAAC,CAAC"}
@@ -1,4 +1,10 @@
1
- import { type Browser, type BrowserContext } from "playwright";
1
+ import { type Browser } from "playwright-core";
2
+ export type LoadHtmlOptions = {
3
+ browser?: Browser;
4
+ timeout?: number;
5
+ waitUntil?: "load" | "domcontentloaded" | "networkidle";
6
+ superBypassMode?: boolean;
7
+ };
2
8
  /**
3
9
  * Useful function for load the HTML of a URL using Playwright.
4
10
  * **Not recommended** for use in production environments.
@@ -13,4 +19,4 @@ import { type Browser, type BrowserContext } from "playwright";
13
19
  * console.log(html);
14
20
  * ```
15
21
  */
16
- export declare const loadHtml: (url: string, context?: BrowserContext | Browser) => Promise<string>;
22
+ export declare const loadHtml: (url: string, options?: LoadHtmlOptions) => Promise<string>;
@@ -1 +1 @@
1
- {"version":3,"file":"playwright.js","sourceRoot":"","sources":["../../../src/loaders/playwright.ts"],"names":[],"mappings":"AAAA,OAAO,EAAqC,QAAQ,EAAE,MAAM,YAAY,CAAC;AAEzE;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,GAAW,EAAE,OAAkC,EAAE,EAAE;IACjF,MAAM,QAAQ,GAAG,OAAO,IAAI,CAAC,MAAM,QAAQ,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IAExE,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,OAAO,EAAE,CAAC;IACtC,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;IAC5C,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;IAClC,MAAM,IAAI,CAAC,KAAK,EAAE,CAAC;IAEnB,IAAI,CAAC,OAAO,EAAE,CAAC;QACd,MAAM,QAAQ,CAAC,KAAK,EAAE,CAAC;IACxB,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC"}
1
+ {"version":3,"file":"playwright.js","sourceRoot":"","sources":["../../../src/loaders/playwright.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,QAAQ,EAAE,OAAO,EAAE,MAAM,iBAAiB,CAAC;AASlE,MAAM,mBAAmB,GAAG;IAC3B,SAAS,EACR,qHAAqH;IACtH,QAAQ,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE;IACvC,iBAAiB,EAAE,CAAC;IACpB,QAAQ,EAAE,KAAK;IACf,QAAQ,EAAE,KAAK;IACf,iBAAiB,EAAE,IAAI;IACvB,MAAM,EAAE,OAAO;IACf,UAAU,EAAE,kBAAkB;CAC9B,CAAC;AAEF;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,GAAW,EAAE,OAAyB,EAAE,EAAE;IACxE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,eAAe,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IACvE,MAAM,QAAQ,GAAG,OAAO,IAAI,CAAC,MAAM,QAAQ,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACxE,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,UAAU,CAAC,eAAe,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC,CAAC;IAE7G,IAAI,eAAe,EAAE,CAAC;QACrB,MAAM,OAAO,CAAC,aAAa,CAAC,GAAG,EAAE;YAChC,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,SAAS,EAAE,CAAC,CAAC;YACxE,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;YAC9E,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,SAAS,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC;QAC7E,CAAC,CAAC,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;IACrC,IAAI,eAAe,EAAE,CAAC;QACrB,MAAM,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE;YACrC,IAAI,KAAK,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,CAAC,kBAAkB,CAAC,EAAE,CAAC;gBACxD,OAAO,KAAK,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YACD,OAAO,KAAK,CAAC,QAAQ,EAAE,CAAC;QACzB,CAAC,CAAC,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,SAAS,IAAI,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC;IAClE,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;IAClC,MAAM,IAAI,CAAC,KAAK,EAAE,CAAC;IAEnB,IAAI,OAAO,EAAE,CAAC;QACb,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;IACvB,CAAC;SAAM,CAAC;QACP,MAAM,QAAQ,CAAC,KAAK,EAAE,CAAC;IACxB,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC"}
@@ -0,0 +1,2 @@
1
+ import { type Handle } from "hast-util-to-mdast";
2
+ export declare const customBrHandler: Handle;
@@ -0,0 +1 @@
1
+ {"version":3,"file":"custom-br-handler.js","sourceRoot":"","sources":["../../../src/mdast-handlers/custom-br-handler.ts"],"names":[],"mappings":"AAAA,OAAO,EAAe,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAElE,MAAM,CAAC,MAAM,eAAe,GAAW,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;IACtD,OAAO,eAAe,CAAC,EAAE,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;AACxC,CAAC,CAAC"}
@@ -1 +1 @@
1
- {"version":3,"file":"mdast-to-markdown.js","sourceRoot":"","sources":["../../src/mdast-to-markdown.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,EAAqC,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEvF,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAY/C;;GAEG;AACH,MAAM,CAAC,MAAM,iCAAiC,GAA2B;IACxE,UAAU,EAAE,CAAC,aAAa,EAAE,EAAE,cAAc,EAAE,CAAC;IAC/C,MAAM,EAAE,GAAG;CACX,CAAC;AAEF;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,KAA4B,EAAE,OAAgC,EAAU,EAAE;IACzG,MAAM,EAAE,OAAO,EAAE,GAAG,iBAAiB,EAAE,GAAG,EAAE,GAAG,iCAAiC,EAAE,GAAG,OAAO,EAAE,CAAC;IAE/F,IAAI,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,iBAAiB,CAAC,CAAC;IAE9D,IAAI,OAAO,EAAE,CAAC;QACb,QAAQ,GAAG,YAAY,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC5C,CAAC;IAED,OAAO,QAAQ,CAAC;AACjB,CAAC,CAAC"}
1
+ {"version":3,"file":"mdast-to-markdown.js","sourceRoot":"","sources":["../../src/mdast-to-markdown.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,EAAqC,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEvF,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAY/C;;GAEG;AACH,MAAM,CAAC,MAAM,iCAAiC,GAA2B;IACxE,UAAU,EAAE,CAAC,aAAa,EAAE,EAAE,cAAc,EAAE,CAAC;IAC/C,MAAM,EAAE,GAAG;CACX,CAAC;AAEF;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,KAA4B,EAAE,OAAgC,EAAU,EAAE;IACzG,MAAM,EAAE,OAAO,EAAE,GAAG,iBAAiB,EAAE,GAAG,EAAE,GAAG,iCAAiC,EAAE,GAAG,OAAO,EAAE,CAAC;IAE/F,IAAI,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,iBAAiB,CAAC,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;IAEvF,IAAI,OAAO,EAAE,CAAC;QACb,QAAQ,GAAG,YAAY,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC5C,CAAC;IAED,OAAO,QAAQ,CAAC;AACjB,CAAC,CAAC"}
@@ -0,0 +1,4 @@
1
+ import type { Nodes as Hast } from "hast";
2
+ export declare const getLangFromHast: (node: Hast) => string | undefined;
3
+ export declare const getLangFromStr: (str: string) => string | undefined;
4
+ export declare const getUrlFromHast: (node: Hast) => string | undefined;
@@ -0,0 +1 @@
1
+ {"version":3,"file":"hast-utils.js","sourceRoot":"","sources":["../../../src/utils/hast-utils.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAErD,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAClC,IAAI,IAAI,IAAI,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtD,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;IAC7B,CAAC;IACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO;IACR,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAChC,IAAI,OAAO,CAAC,OAAO,KAAK,MAAM,EAAE,CAAC;QAChC,OAAO;IACR,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,OAAO,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC;IAC3E,IAAI,QAAQ,EAAE,CAAC;QACd,OAAO,QAAkB,CAAC;IAC3B,CAAC;IAED,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,GAAW,EAAE,EAAE;IAC7C,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,uBAAuB,CAAC,CAAC;IACjD,IAAI,KAAK,EAAE,CAAC;QACX,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC;IACjB,CAAC;IACD,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,IAAU,EAAsB,EAAE;IAChE,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,SAAS,CAAC;IAClB,CAAC;IAED,MAAM,iBAAiB,GAAG,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACpD,MAAM,QAAQ,GAAG,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAEzC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC7B,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,CAAC,QAAQ,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;QAClE,IAAI,OAAO,QAAQ,KAAK,QAAQ,IAAI,iBAAiB,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1E,OAAO,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;QAC1F,CAAC;IACF,CAAC;IAED,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
@@ -0,0 +1,46 @@
1
+ // src/utils/hast-utils.ts
2
+ import { select, selectAll } from "hast-util-select";
3
+ var getLangFromHast = (node) => {
4
+ const html = select("html", node);
5
+ if (html && typeof html.properties.lang === "string") {
6
+ return html.properties.lang;
7
+ }
8
+ if (node.type !== "element") {
9
+ return;
10
+ }
11
+ const element = node;
12
+ if (element.tagName !== "html") {
13
+ return;
14
+ }
15
+ const langAttr = element.properties.lang || element.properties["xml:lang"];
16
+ if (langAttr) {
17
+ return langAttr;
18
+ }
19
+ return void 0;
20
+ };
21
+ var getLangFromStr = (str) => {
22
+ const match = str.match(/lang=["']([^"']+)["']/);
23
+ if (match) {
24
+ return match[1];
25
+ }
26
+ return void 0;
27
+ };
28
+ var getUrlFromHast = (node) => {
29
+ if (node.type !== "element") {
30
+ return void 0;
31
+ }
32
+ const metaTagAttributes = ["og:url", "twitter:url"];
33
+ const metaTags = selectAll("meta", node);
34
+ for (const meta of metaTags) {
35
+ const property = meta.properties.property || meta.properties.name;
36
+ if (typeof property === "string" && metaTagAttributes.includes(property)) {
37
+ return typeof meta.properties.content === "string" ? meta.properties.content : void 0;
38
+ }
39
+ }
40
+ return void 0;
41
+ };
42
+ export {
43
+ getLangFromHast,
44
+ getLangFromStr,
45
+ getUrlFromHast
46
+ };
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "webforai",
3
- "version": "1.5.1",
3
+ "version": "1.6.1",
4
4
  "description": "A library that provides a web interface for AI",
5
5
  "author": "inaridiy",
6
6
  "license": "MIT",
@@ -80,20 +80,31 @@
80
80
  },
81
81
  "peerDependencies": {
82
82
  "@cloudflare/puppeteer": ">=0.0.6",
83
- "playwright": ">=1.4",
83
+ "playwright-core": ">=1.4",
84
84
  "puppeteer": ">=22"
85
85
  },
86
+ "peerDependenciesMeta": {
87
+ "@cloudflare/puppeteer": {
88
+ "optional": true
89
+ },
90
+ "playwright-core": {
91
+ "optional": false
92
+ },
93
+ "puppeteer": {
94
+ "optional": true
95
+ }
96
+ },
86
97
  "dependencies": {
87
98
  "@clack/prompts": "^0.7.0",
99
+ "boxen": "^8.0.1",
88
100
  "commander": "^12.1.0",
89
- "fastest-levenshtein": "^1.0.16",
90
- "hast-util-from-html": "^2.0.1",
91
- "hast-util-select": "^6.0.1",
92
- "hast-util-to-html": "^9.0.0",
101
+ "hast-util-from-html": "^2.0.3",
102
+ "hast-util-select": "^6.0.2",
103
+ "hast-util-to-html": "^9.0.3",
93
104
  "hast-util-to-mdast": "^10.1.0",
94
105
  "hast-util-to-string": "^3.0.0",
95
106
  "hast-util-to-text": "^4.0.0",
96
- "mathml-to-latex": "^1.3.0",
107
+ "mathml-to-latex": "^1.4.1",
97
108
  "mdast-util-gfm": "^3.0.0",
98
109
  "mdast-util-math": "^3.0.0",
99
110
  "mdast-util-to-markdown": "^2.1.0",
@@ -107,16 +118,17 @@
107
118
  "@tsconfig/recommended": "^1.0.3",
108
119
  "@types/hast": "^3.0.2",
109
120
  "@types/mdast": "^4.0.2",
110
- "@types/node": "^20.11.0",
121
+ "@types/node": "^20.14.10",
111
122
  "arg": "^5.0.2",
112
123
  "esbuild": "^0.19.11",
124
+ "fastest-levenshtein": "^1.0.16",
113
125
  "glob": "^10.3.10",
114
126
  "ncp": "^2.0.0",
115
127
  "np": "^9.2.0",
116
- "playwright": "^1.40.1",
128
+ "playwright-core": "^1.40.1",
117
129
  "puppeteer": "^23.2.2",
118
130
  "rimraf": "^5.0.5",
119
- "tsx": "^4.7.0",
131
+ "tsx": "^4.19.1",
120
132
  "typescript": "^5.4.5"
121
133
  },
122
134
  "scripts": {
@@ -1,34 +0,0 @@
1
- "use strict";
2
- var __create = Object.create;
3
- var __defProp = Object.defineProperty;
4
- var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
5
- var __getOwnPropNames = Object.getOwnPropertyNames;
6
- var __getProtoOf = Object.getPrototypeOf;
7
- var __hasOwnProp = Object.prototype.hasOwnProperty;
8
- var __copyProps = (to, from, except, desc) => {
9
- if (from && typeof from === "object" || typeof from === "function") {
10
- for (let key of __getOwnPropNames(from))
11
- if (!__hasOwnProp.call(to, key) && key !== except)
12
- __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
13
- }
14
- return to;
15
- };
16
- var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
17
- // If the importer is in node compatibility mode or this is not an ESM
18
- // file that has been converted to a CommonJS file using a Babel-
19
- // compatible transform (i.e. "__esModule" has not been set), then set
20
- // "default" to the CommonJS "module.exports" for node compatibility.
21
- isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
22
- mod
23
- ));
24
- var import_prompts = require("@clack/prompts");
25
- var import_commander = require("commander");
26
- var import_picocolors = __toESM(require("picocolors"), 1);
27
- var import_package = __toESM(require("../../package.json"), 1);
28
- var import_handlers = require("./commands/handlers");
29
- var import_constants = require("./constants");
30
- import_commander.program.name("webforai").description("CLI tool for ultra-precise HTML to Markdown conversion").version(import_package.default.version, "-v, --version", "output the current version").argument("[source]", "URL or path to process").argument("[outputPath]", "Path to output file or directory").option("--mode <mode>", `Processing mode (${import_constants.MODES.join(", ")})`).option("--loader <loader>", `Loader to use (${import_constants.LOADERS.join(", ")})`).option("--baseUrl <baseUrl>", "Base URL for relative links").option("-o, --stdout", "output to stdout").option("-d, --debug", "output extra debugging information").on("option:debug", () => {
31
- process.env.DEBUG = "true";
32
- }).action(import_handlers.webforaiCommand);
33
- (0, import_prompts.intro)(import_picocolors.default.inverse(`webforai CLI version ${import_package.default.version}`));
34
- import_commander.program.parse();
@@ -1,84 +0,0 @@
1
- "use strict";
2
- var __create = Object.create;
3
- var __defProp = Object.defineProperty;
4
- var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
5
- var __getOwnPropNames = Object.getOwnPropertyNames;
6
- var __getProtoOf = Object.getPrototypeOf;
7
- var __hasOwnProp = Object.prototype.hasOwnProperty;
8
- var __export = (target, all) => {
9
- for (var name in all)
10
- __defProp(target, name, { get: all[name], enumerable: true });
11
- };
12
- var __copyProps = (to, from, except, desc) => {
13
- if (from && typeof from === "object" || typeof from === "function") {
14
- for (let key of __getOwnPropNames(from))
15
- if (!__hasOwnProp.call(to, key) && key !== except)
16
- __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
17
- }
18
- return to;
19
- };
20
- var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
21
- // If the importer is in node compatibility mode or this is not an ESM
22
- // file that has been converted to a CommonJS file using a Babel-
23
- // compatible transform (i.e. "__esModule" has not been set), then set
24
- // "default" to the CommonJS "module.exports" for node compatibility.
25
- isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
26
- mod
27
- ));
28
- var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
29
- var handlers_exports = {};
30
- __export(handlers_exports, {
31
- webforaiCommand: () => webforaiCommand
32
- });
33
- module.exports = __toCommonJS(handlers_exports);
34
- var import_node_fs = __toESM(require("node:fs"), 1);
35
- var import_node_path = __toESM(require("node:path"), 1);
36
- var import_prompts = require("@clack/prompts");
37
- var import_html_to_markdown = require("../../../html-to-markdown");
38
- var import_inputOutputPath = require("../../helpers/inputOutputPath");
39
- var import_inputSourcePath = require("../../helpers/inputSourcePath");
40
- var import_selectExtractMode = require("../../helpers/selectExtractMode");
41
- var import_selectLoader = require("../../helpers/selectLoader");
42
- var import_utils = require("../../utils");
43
- var import_loadHtml = require("./loadHtml");
44
- const aiModeOptions = { linkAsText: true, tableAsText: true, hideImage: true };
45
- const readabilityModeOptions = { linkAsText: false, tableAsText: false, hideImage: false };
46
- const webforaiCommand = async (initialPath, initialOutputPath, options) => {
47
- const sourcePath = initialPath ?? await (0, import_inputSourcePath.inputSourcePath)();
48
- if (options.debug) {
49
- console.debug(`sourcePath: ${sourcePath}`);
50
- }
51
- const loader = (0, import_utils.isUrl)(sourcePath) ? { type: "web", loader: options.loader ?? await (0, import_selectLoader.selectLoader)() } : { type: "local", loader: null };
52
- if (options.debug) {
53
- console.debug(`loader: ${loader}`);
54
- }
55
- const outputPath = initialOutputPath ?? await (0, import_inputOutputPath.inputOutputPath)(sourcePath);
56
- if (options.debug) {
57
- console.debug(`outputPath: ${outputPath}`);
58
- }
59
- const mode = options.mode ?? await (0, import_selectExtractMode.selectExtractMode)();
60
- if (options.debug) {
61
- console.debug(`mode: ${mode}`);
62
- }
63
- const content = await (0, import_loadHtml.loadHtml)(sourcePath, loader);
64
- if (options.debug) {
65
- console.debug(`content: ${content}`);
66
- }
67
- const markdown = (0, import_html_to_markdown.htmlToMarkdown)(content, {
68
- baseUrl: (0, import_utils.isUrl)(sourcePath) ? sourcePath : options.baseUrl,
69
- ...mode === "ai" ? aiModeOptions : readabilityModeOptions
70
- });
71
- if (options.debug) {
72
- console.debug(`markdown: ${markdown}`);
73
- }
74
- const directory = import_node_path.default.dirname(outputPath);
75
- if (!import_node_fs.default.existsSync(directory)) {
76
- import_node_fs.default.mkdirSync(directory, { recursive: true });
77
- }
78
- import_node_fs.default.writeFileSync(outputPath, markdown);
79
- (0, import_prompts.outro)("Markdown conversion is complete!");
80
- };
81
- // Annotate the CommonJS export names for ESM import in node:
82
- 0 && (module.exports = {
83
- webforaiCommand
84
- });