webforai 1.1.6 → 1.2.2

package/dist/cjs/extract-hast/readability.js

@@ -0,0 +1,150 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var readability_exports = {};
+__export(readability_exports, {
+  readabilityExtractHast: () => readabilityExtractHast
+});
+module.exports = __toCommonJS(readability_exports);
+var import_hast_util_select = require("hast-util-select");
+var import_hast_util_to_string = require("hast-util-to-string");
+var import_unist_util_filter = require("unist-util-filter");
+var import_unist_util_parents = require("unist-util-parents");
+var import_utils = require("./utils");
+const UNLIKELY_ROLES = ["menu", "menubar", "complementary", "navigation", "alert", "alertdialog", "dialog"];
+const REGEXPS = {
+  hidden: /hidden|invisible|fallback-image/i,
+  byline: /byline|author|dateline|writtenby|p-author/i,
+  unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote|speechify-ignore/i,
+  okMaybeItsACandidate: /and|article|body|column|content|main|shadow|code/i
+};
+const BODY_SELECTORS = ["article", "#article", ".article_body", ".article-body", "#content", ".entry", "main"];
+const PARAGRAPH_TAGS = ["p", "div", "section", "article", "main", "ul", "ol", "li"];
+const BASE_MINIMAL_LENGTH = {
+  ja: 200,
+  en: 500
+};
+const tryGetLang = (node) => {
+  if (node.type !== "element")
+    return;
+  const element = node;
+  if (element.tagName !== "html")
+    return;
+  const langAttr = element.properties.lang || element.properties["xml:lang"];
+  if (langAttr)
+    return langAttr;
+};
+const metadataFilter = (node) => !["comment", "doctype"].includes(node.type);
+const universalElementFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  if (["script", "aside"].includes(element.tagName))
+    return false;
+  if (["hidden", "aria-hidden"].some((key) => element.properties[key]))
+    return false;
+  if ((0, import_utils.classnames)(element).some((classname) => REGEXPS.hidden.test(classname)))
+    return false;
+  if (element.tagName === "dialog")
+    return false;
+  if (element.properties.role === "dialog" && element.properties["aria-modal"])
+    return false;
+  if (element.properties.rel === "author" && (0, import_utils.isStrInclude)(element.properties.itemprop, "author"))
+    return false;
+  if (REGEXPS.byline.test((0, import_utils.matchString)(element)))
+    return false;
+  if (element.properties.role && UNLIKELY_ROLES.includes(element.properties.role))
+    return false;
+  return true;
+};
+const unlikelyElementFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  const match = (0, import_utils.matchString)(element);
+  if (["body", "article", "main", "section", "a"].includes(element.tagName))
+    return true;
+  if ((0, import_utils.hasAncestors)(element, ["table", "code"], 3))
+    return true;
+  if (REGEXPS.unlikelyCandidates.test(match) && !REGEXPS.okMaybeItsACandidate.test(match))
+    return false;
+  return true;
+};
+const removeEmptyFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  if (!PARAGRAPH_TAGS.includes(element.tagName))
+    return true;
+  const text = (0, import_hast_util_to_string.toString)(element);
+  if (text.length < 10)
+    return false;
+  return true;
+};
+const readabilityExtractHast = (hast) => {
+  const lang = String((0, import_hast_util_select.select)("html", hast)?.properties.lang || tryGetLang(hast) || "en");
+  const body = (0, import_hast_util_select.select)("body", hast) ?? hast;
+  const proxiedHast = (0, import_unist_util_parents.parents)(body);
+  let baseFilterd = (0, import_unist_util_filter.filter)(proxiedHast, (node) => {
+    if (!metadataFilter(node))
+      return false;
+    if (!universalElementFilter(node))
+      return false;
+    return true;
+  });
+  if (!baseFilterd)
+    return { type: "root", children: [] };
+  const baseText = (0, import_hast_util_to_string.toString)(baseFilterd);
+  let minimalLength = lang in BASE_MINIMAL_LENGTH ? BASE_MINIMAL_LENGTH[lang] : 500;
+  if (baseText.length > minimalLength) {
+    const filterd = (0, import_unist_util_filter.filter)(baseFilterd, (node) => {
+      if (!unlikelyElementFilter(node))
+        return false;
+      return true;
+    });
+    if (filterd)
+      baseFilterd = filterd;
+  } else
+    minimalLength = Math.max(0, baseText.length - 200);
+  let bodyTree = baseFilterd;
+  for (const selector of BODY_SELECTORS) {
+    const body2 = { type: "root", children: (0, import_hast_util_select.selectAll)(selector, baseFilterd) };
+    const bodyText = (0, import_hast_util_to_string.toString)(body2);
+    if (bodyText.length < 25)
+      continue;
+    const links = (0, import_hast_util_select.selectAll)("a", body2);
+    const linkText = links.map((link) => (0, import_hast_util_to_string.toString)(link)).join("");
+    const linkDensity = linkText.length / bodyText.length;
+    if (linkDensity > 0.4)
+      continue;
+    if (bodyText.length > minimalLength) {
+      bodyTree = body2;
+      break;
+    }
+  }
+  const finalTree = (0, import_unist_util_filter.filter)(bodyTree, (node) => {
+    if (!removeEmptyFilter(node))
+      return false;
+    return true;
+  });
+  return finalTree;
+};
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  readabilityExtractHast
+});

package/dist/cjs/extract-hast/strongly.js

@@ -22,26 +22,8 @@ __export(strongly_exports, {
 });
 module.exports = __toCommonJS(strongly_exports);
 var import_hast_util_select = require("hast-util-select");
-var import_hast_util_to_text = require("hast-util-to-text");
 var import_unist_util_filter = require("unist-util-filter");
-const evaluateTextComplexity = (text) => {
-  const words = text.trim().split(/\s+/);
-  const wordCount = words.length;
-  const charCount = text.length;
-  const avgWordLength = charCount / wordCount;
-  const uniqueWords = new Set(words);
-  const uniqueWordCount = uniqueWords.size;
-  const complexityScore = avgWordLength * 0.4 + uniqueWordCount / wordCount * 100 * 0.6;
-  return {
-    wordCount,
-    charCount,
-    avgWordLength,
-    uniqueWordCount,
-    complexityScore
-  };
-};
 const REMOVE_REGEX = /fallback|icon|header|nav|speechify-ignore|assist|-ad-|banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|assist|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i;
-const HIDDEN_REGEX = /hidden|^hid$| hid$| hid |^hid /i;
 const stronglyExtractHast = (hast) => {
   const selectors = ["article", "#article", "#content", "main", "div.container"];
   let selectedElement = hast;
@@ -61,10 +43,6 @@ const stronglyExtractHast = (hast) => {
     const elementName = `${element.tagName} ${element.properties.id} ${element.properties.className}`;
     if (REMOVE_REGEX.test(elementName))
       return false;
-    const isHidden = HIDDEN_REGEX.test(elementName);
-    const complexity = evaluateTextComplexity((0, import_hast_util_to_text.toText)(element));
-    if (isHidden && complexity.complexityScore < 40)
-      return false;
     return true;
   });
   return removedElements;

package/dist/cjs/extract-hast/utils.js

@@ -0,0 +1,55 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var utils_exports = {};
+__export(utils_exports, {
+  classnames: () => classnames,
+  hasAncestors: () => hasAncestors,
+  isStrInclude: () => isStrInclude,
+  matchString: () => matchString
+});
+module.exports = __toCommonJS(utils_exports);
+const matchString = (element) => `${element.tagName} ${element.properties.id} ${classnames(element).join(" ")} }`;
+const classnames = (element) => {
+  if (Array.isArray(element.properties.className))
+    return element.properties.className;
+  return [];
+};
+const isStrInclude = (value, match) => {
+  if (typeof value === "string")
+    return value.includes(match);
+  return false;
+};
+const hasAncestors = (element, tagNames, depth = 3) => {
+  let parent = element.parent;
+  for (let i = 0; i < depth; i++) {
+    if (!parent || parent.type !== "element")
+      return false;
+    if (tagNames.includes(parent.tagName))
+      return true;
+    parent = parent.parent;
+  }
+  return false;
+};
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  classnames,
+  hasAncestors,
+  isStrInclude,
+  matchString
+});

package/dist/cjs/html-to-mdast.js

@@ -24,13 +24,17 @@ module.exports = __toCommonJS(html_to_mdast_exports);
 var import_hast_util_from_html = require("hast-util-from-html");
 var import_hast_util_to_mdast = require("hast-util-to-mdast");
 var import_extract_hast = require("./extract-hast");
+var import_readability = require("./extract-hast/readability");
 var import_custom_a_handler = require("./mdast-handlers/custom-a-handler");
 var import_custom_div_handler = require("./mdast-handlers/custom-div-handler");
+var import_custom_img_handler = require("./mdast-handlers/custom-img-handler");
+var import_custom_table_handler = require("./mdast-handlers/custom-table-handler");
 var import_math_handler = require("./mdast-handlers/math-handler");
 const PRESET_EXTRACT_HAST = {
-  strongly: import_extract_hast.stronglyExtractHast
+  strongly: import_extract_hast.stronglyExtractHast,
+  readability: import_readability.readabilityExtractHast
 };
-const DEFAULT_EXTRACT_HAST = ["strongly"];
+const DEFAULT_EXTRACT_HAST = ["readability"];
 const htmlToMdast = (html, options) => {
   const { extractHast = DEFAULT_EXTRACT_HAST } = options || {};
   const _extractHast = Array.isArray(extractHast) ? extractHast : [extractHast];
@@ -46,13 +50,13 @@ const htmlToMdast = (html, options) => {
   }, hast) || hast;
   const mdast = (0, import_hast_util_to_mdast.toMdast)(extractedHast, {
     handlers: {
-      div: import_custom_div_handler.customDivHandler,
       math: import_math_handler.mathHandler,
+      div: import_custom_div_handler.customDivHandler,
       a: (0, import_custom_a_handler.customAHandler)({ asText: options?.linkAsText }),
+      img: (0, import_custom_img_handler.customImgHandler)({ hideImage: options?.hideImage }),
+      table: (0, import_custom_table_handler.customTableHandler)({ asText: options?.tableAsText }),
       br: () => {
-      },
-      ...options?.hideImage ? { img: () => {
-      } } : {}
+      }
     }
   });
   return mdast;

package/dist/cjs/mdast-handlers/custom-a-handler.js

@@ -22,12 +22,12 @@ __export(custom_a_handler_exports, {
 });
 module.exports = __toCommonJS(custom_a_handler_exports);
 var import_hast_util_to_mdast = require("hast-util-to-mdast");
+var import_hast_util_to_string = require("hast-util-to-string");
 const customAHandler = (options) => (state, node) => {
   if (options?.asText) {
-    const link2 = import_hast_util_to_mdast.defaultHandlers.a(state, node);
-    if (link2.children?.[0]?.type === "text")
-      return link2.children[0];
-    return void 0;
+    const link2 = { type: "text", value: (0, import_hast_util_to_string.toString)(node) };
+    state.patch(node, link2);
+    return link2;
   }
   const link = import_hast_util_to_mdast.defaultHandlers.a(state, node);
   if (link.children.length > 0)

package/dist/cjs/mdast-handlers/custom-div-handler.js

@@ -21,42 +21,46 @@ __export(custom_div_handler_exports, {
   customDivHandler: () => customDivHandler
 });
 module.exports = __toCommonJS(custom_div_handler_exports);
+var import_hast_util_select = require("hast-util-select");
 var import_hast_util_to_mdast = require("hast-util-to-mdast");
+var import_hast_util_to_string = require("hast-util-to-string");
 var import_hast_util_to_text = require("hast-util-to-text");
 var import_trim_trailing_lines = require("trim-trailing-lines");
+var import_detect_code_lang = require("../utils/detect-code-lang");
+const CODE_BLOCK_REGEX = /highlight-source|language-|codegroup|codeblock|code-block/i;
+const CODE_FILENAME_SELECTORS = "[class*='fileName'],[class*='fileName'],[class*='title'],[class*='Title']";
+const LANGUAGE_MATCH_REGEX = /language-(\w+)|highlight-source-(\w+)/;
+const findRecursive = (array, condition, maxDepth = 3) => {
+  if (maxDepth <= 0)
+    return null;
+  for (const value of array) {
+    const result = condition(value);
+    if (Array.isArray(result))
+      return findRecursive(result, condition, maxDepth - 1);
+    if (result)
+      return value;
+  }
+  return null;
+};
 const customDivHandler = (state, node) => {
   const classNames = Array.isArray(node.properties.className) ? node.properties.className : [];
-  let highlightLanguage;
-  for (const className of classNames) {
-    if (String(className) === "highlight-text-html-basic") {
-      highlightLanguage = "html";
-      break;
-    }
-    if (String(className).includes("highlight-source-")) {
-      highlightLanguage = String(className).slice("highlight-source-".length);
-      break;
-    }
-    if (String(className).includes("language-")) {
-      highlightLanguage = String(className).slice("language-".length);
-      break;
-    }
-    if (String(className).includes("code-block_wrapper")) {
-      const text = (0, import_hast_util_to_text.toText)(node);
-      if (!text.includes("."))
-        break;
-      const lang = (0, import_hast_util_to_text.toText)(node).split("\n")[0].split(".").slice(-1)[0];
-      highlightLanguage = lang;
-      break;
-    }
-  }
-  if (highlightLanguage !== void 0) {
-    const code = node.children.find((child) => child.type === "element" && child.tagName === "pre") || node;
-    const result = {
-      type: "code",
-      lang: highlightLanguage,
-      meta: null,
-      value: (0, import_trim_trailing_lines.trimTrailingLines)((0, import_hast_util_to_text.toText)(code)).trim()
-    };
+  const codeBlock = findRecursive(node.children, (child) => {
+    if (child.type !== "element")
+      return false;
+    if (child.tagName === "pre")
+      return true;
+    return child.children.filter((child2) => child2.type === "element");
+  });
+  if (codeBlock && classNames.some((className) => CODE_BLOCK_REGEX.test(className))) {
+    const codeValue = (0, import_trim_trailing_lines.trimTrailingLines)((0, import_hast_util_to_text.toText)(codeBlock)).trim();
+    const filenameElement = (0, import_hast_util_select.select)(CODE_FILENAME_SELECTORS, node);
+    const fileLang = filenameElement ? (0, import_hast_util_to_string.toString)(filenameElement).match(/\.(\w+)$/)?.[1] : null;
+    const classLang = classNames.map((className) => {
+      const match = className.match(LANGUAGE_MATCH_REGEX);
+      return match?.[1] || match?.[2];
+    }).find((className) => className);
+    const lang = fileLang || classLang || (0, import_detect_code_lang.detectLanguage)(codeValue) || null;
+    const result = { type: "code", lang, meta: null, value: codeValue };
     state.patch(node, result);
     return result;
   }

package/dist/cjs/mdast-handlers/custom-img-handler.js

@@ -0,0 +1,33 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var custom_img_handler_exports = {};
+__export(custom_img_handler_exports, {
+  customImgHandler: () => customImgHandler
+});
+module.exports = __toCommonJS(custom_img_handler_exports);
+var import_hast_util_to_mdast = require("hast-util-to-mdast");
+const customImgHandler = (options) => (state, node) => {
+  if (options?.hideImage)
+    return void 0;
+  return import_hast_util_to_mdast.defaultHandlers.image(state, node);
+};
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  customImgHandler
+});

package/dist/cjs/mdast-handlers/custom-table-handler.js

@@ -0,0 +1,37 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var custom_table_handler_exports = {};
+__export(custom_table_handler_exports, {
+  customTableHandler: () => customTableHandler
+});
+module.exports = __toCommonJS(custom_table_handler_exports);
+var import_hast_util_to_mdast = require("hast-util-to-mdast");
+var import_hast_util_to_text = require("hast-util-to-text");
+const customTableHandler = (options) => (state, node) => {
+  if (options?.asText) {
+    const paragraph = { type: "paragraph", children: [{ type: "text", value: (0, import_hast_util_to_text.toText)(node) }] };
+    state.patch(node, paragraph);
+    return paragraph;
+  }
+  return import_hast_util_to_mdast.defaultHandlers.table(state, node);
+};
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  customTableHandler
+});

package/dist/cjs/utils/detect-code-lang.js

@@ -0,0 +1,73 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var detect_code_lang_exports = {};
+__export(detect_code_lang_exports, {
+  detectLanguage: () => detectLanguage
+});
+module.exports = __toCommonJS(detect_code_lang_exports);
+const languages = [
+  ["bash", [/#!(\/usr)?\/bin\/bash/g, 500], [/\b(if|elif|then|fi|echo)\b|\$/g, 10]],
+  ["html", [/<\/?[a-z-]+[^\n>]*>/g, 10], [/^\s+<!DOCTYPE\s+html/g, 500]],
+  ["http", [/^(GET|HEAD|POST|PUT|DELETE|PATCH|HTTP)\b/g, 500]],
+  [
+    "ts",
+    [/(import .* from|export \*|export const|const .* = await)/g, 400],
+    [
+      /\b(console|await|async|function|undefined|export|import|this|class|for|let|const|map|join|require|implements|interface|namespace)\b/g,
+      10
+    ]
+  ],
+  [
+    "tsx",
+    [/(import .* from|export \*|export const)/g, 400],
+    [/\b(react|next|FC)\b/g, 200],
+    [
+      /\b(console|await|async|function|export|undefined|import|this|class|for|let|const|map|join|require|implements|interface)\b/g,
+      10
+    ]
+  ],
+  ["py", [/\b(def|print|class|and|or|lambda)\b/g, 10]],
+  ["sql", [/\b(SELECT|INSERT|FROM)\b/g, 50]],
+  ["pl", [/#!(\/usr)?\/bin\/perl/g, 500], [/\b(use|print)\b|\$/g, 10]],
+  ["lua", [/#!(\/usr)?\/bin\/lua/g, 500]],
+  ["make", [/\b(ifneq|endif|if|elif|then|fi|echo|.PHONY|^[a-z]+ ?:$)\b|\$/gm, 10]],
+  ["uri", [/https?:|mailto:|tel:|ftp:/g, 30]],
+  ["css", [/^(@import|@page|@media|(\.|#)[a-z]+)/gm, 20]],
+  ["diff", [/^[+><-]/gm, 10], [/^@@ ?[-+,0-9 ]+ ?@@/gm, 25]],
+  ["md", [/^(>|\t\*|\t\d+.)/gm, 10], [/\[.*\](.*)/g, 10]],
+  ["docker", [/^(FROM|ENTRYPOINT|RUN)/gm, 500]],
+  ["xml", [/<\/?[a-z-]+[^\n>]*>/g, 10], [/^<\?xml/g, 500]],
+  ["c", [/#include\b|\bprintf\s+\(/g, 100]],
+  ["rs", [/^\s+(use|fn|mut|match)\b/gm, 100]],
+  ["go", [/\b(func|fmt|package)\b/g, 100]],
+  ["java", [/^import\s+java/gm, 500]],
+  ["asm", [/^(section|global main|extern|\t(call|mov|ret))/gm, 100]],
+  ["css", [/^(@import|@page|@media|(\.|#)[a-z]+)/gm, 20]],
+  ["json", [/\b(true|false|null|\{})\b|\"[^"]+\":/g, 10]],
+  ["yaml", [/^(\s+)?[a-z][a-z0-9]*:/gim, 10]]
+];
+const detectLanguage = (code) => {
+  return languages.map(
+    ([lang, ...features]) => [lang, features.reduce((acc, [match, score]) => acc + [...code.matchAll(match)].length * score, 0)]
+  ).filter(([_, score]) => score > 20).sort((a, b) => b[1] - a[1])[0]?.[0] || "plain";
+};
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  detectLanguage
+});

package/dist/extract-hast/readability.js

@@ -0,0 +1,127 @@
+// src/extract-hast/readability.ts
+import { select, selectAll } from "hast-util-select";
+import { toString } from "hast-util-to-string";
+import { filter } from "unist-util-filter";
+import { parents } from "unist-util-parents";
+import { classnames, hasAncestors, isStrInclude, matchString } from "./utils.js";
+var UNLIKELY_ROLES = ["menu", "menubar", "complementary", "navigation", "alert", "alertdialog", "dialog"];
+var REGEXPS = {
+  hidden: /hidden|invisible|fallback-image/i,
+  byline: /byline|author|dateline|writtenby|p-author/i,
+  unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote|speechify-ignore/i,
+  okMaybeItsACandidate: /and|article|body|column|content|main|shadow|code/i
+};
+var BODY_SELECTORS = ["article", "#article", ".article_body", ".article-body", "#content", ".entry", "main"];
+var PARAGRAPH_TAGS = ["p", "div", "section", "article", "main", "ul", "ol", "li"];
+var BASE_MINIMAL_LENGTH = {
+  ja: 200,
+  en: 500
+};
+var tryGetLang = (node) => {
+  if (node.type !== "element")
+    return;
+  const element = node;
+  if (element.tagName !== "html")
+    return;
+  const langAttr = element.properties.lang || element.properties["xml:lang"];
+  if (langAttr)
+    return langAttr;
+};
+var metadataFilter = (node) => !["comment", "doctype"].includes(node.type);
+var universalElementFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  if (["script", "aside"].includes(element.tagName))
+    return false;
+  if (["hidden", "aria-hidden"].some((key) => element.properties[key]))
+    return false;
+  if (classnames(element).some((classname) => REGEXPS.hidden.test(classname)))
+    return false;
+  if (element.tagName === "dialog")
+    return false;
+  if (element.properties.role === "dialog" && element.properties["aria-modal"])
+    return false;
+  if (element.properties.rel === "author" && isStrInclude(element.properties.itemprop, "author"))
+    return false;
+  if (REGEXPS.byline.test(matchString(element)))
+    return false;
+  if (element.properties.role && UNLIKELY_ROLES.includes(element.properties.role))
+    return false;
+  return true;
+};
+var unlikelyElementFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  const match = matchString(element);
+  if (["body", "article", "main", "section", "a"].includes(element.tagName))
+    return true;
+  if (hasAncestors(element, ["table", "code"], 3))
+    return true;
+  if (REGEXPS.unlikelyCandidates.test(match) && !REGEXPS.okMaybeItsACandidate.test(match))
+    return false;
+  return true;
+};
+var removeEmptyFilter = (node) => {
+  if (node.type !== "element")
+    return true;
+  const element = node;
+  if (!PARAGRAPH_TAGS.includes(element.tagName))
+    return true;
+  const text = toString(element);
+  if (text.length < 10)
+    return false;
+  return true;
+};
+var readabilityExtractHast = (hast) => {
+  const lang = String(select("html", hast)?.properties.lang || tryGetLang(hast) || "en");
+  const body = select("body", hast) ?? hast;
+  const proxiedHast = parents(body);
+  let baseFilterd = filter(proxiedHast, (node) => {
+    if (!metadataFilter(node))
+      return false;
+    if (!universalElementFilter(node))
+      return false;
+    return true;
+  });
+  if (!baseFilterd)
+    return { type: "root", children: [] };
+  const baseText = toString(baseFilterd);
+  let minimalLength = lang in BASE_MINIMAL_LENGTH ? BASE_MINIMAL_LENGTH[lang] : 500;
+  if (baseText.length > minimalLength) {
+    const filterd = filter(baseFilterd, (node) => {
+      if (!unlikelyElementFilter(node))
+        return false;
+      return true;
+    });
+    if (filterd)
+      baseFilterd = filterd;
+  } else
+    minimalLength = Math.max(0, baseText.length - 200);
+  let bodyTree = baseFilterd;
+  for (const selector of BODY_SELECTORS) {
+    const body2 = { type: "root", children: selectAll(selector, baseFilterd) };
+    const bodyText = toString(body2);
+    if (bodyText.length < 25)
+      continue;
+    const links = selectAll("a", body2);
+    const linkText = links.map((link) => toString(link)).join("");
+    const linkDensity = linkText.length / bodyText.length;
+    if (linkDensity > 0.4)
+      continue;
+    if (bodyText.length > minimalLength) {
+      bodyTree = body2;
+      break;
+    }
+  }
+  const finalTree = filter(bodyTree, (node) => {
+    if (!removeEmptyFilter(node))
+      return false;
+    return true;
+  });
+  return finalTree;
+};
+export {
+  readabilityExtractHast
+};