@aiquants/html-to-markdown 0.5.0 → 0.5.2

package/dist/{core-DJV8t_Qj.js → core-DhxxOcPm.js}

@@ -20418,6 +20418,335 @@ const rehypeAbsoluteLinks = (options) => {
     });
   };
 };
+const rehypeDecodeEntities = () => {
+  return (tree) => {
+    visit(tree, "text", (node2) => {
+      node2.value = decodeHtmlEntities(node2.value);
+    });
+    visit(tree, "element", (node2) => {
+      if (node2.properties) {
+        for (const [key2, value] of Object.entries(node2.properties)) {
+          if (typeof value === "string") {
+            node2.properties[key2] = decodeHtmlEntities(value);
+          }
+        }
+      }
+    });
+  };
+};
+const preprocessHtmlEntities = (html2) => {
+  return decodeHtmlEntities(html2);
+};
+function decodeHtmlEntities(text2) {
+  if (!text2 || typeof text2 !== "string") {
+    return text2;
+  }
+  text2 = text2.replace(/&#(\d+);/g, (match, dec) => {
+    try {
+      const codePoint = parseInt(dec, 10);
+      if (codePoint >= 0 && codePoint <= 1114111) {
+        return String.fromCodePoint(codePoint);
+      }
+      return match;
+    } catch {
+      return match;
+    }
+  });
+  text2 = text2.replace(/&#[xX]([0-9a-fA-F]+);/g, (match, hex) => {
+    try {
+      const codePoint = parseInt(hex, 16);
+      if (codePoint >= 0 && codePoint <= 1114111) {
+        return String.fromCodePoint(codePoint);
+      }
+      return match;
+    } catch {
+      return match;
+    }
+  });
+  const namedEntities = getComprehensiveNamedEntities();
+  for (const [entity, character] of Object.entries(namedEntities)) {
+    const regex2 = new RegExp(entity.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"), "g");
+    text2 = text2.replace(regex2, character);
+  }
+  return text2;
+}
+function getComprehensiveNamedEntities() {
+  return {
+    // 基本的なHTMLエンティティ（大文字小文字両対応）
+    "&amp;": "&",
+    "&AMP;": "&",
+    "&lt;": "<",
+    "&LT;": "<",
+    "&gt;": ">",
+    "&GT;": ">",
+    "&quot;": '"',
+    "&QUOT;": '"',
+    "&apos;": "'",
+    "&APOS;": "'",
+    "&nbsp;": " ",
+    "&NBSP;": " ",
+    // 数学記号
+    "&times;": "×",
+    "&divide;": "÷",
+    "&minus;": "−",
+    "&plusmn;": "±",
+    "&sup1;": "¹",
+    "&sup2;": "²",
+    "&sup3;": "³",
+    "&frac14;": "¼",
+    "&frac12;": "½",
+    "&frac34;": "¾",
+    "&infin;": "∞",
+    "&sum;": "∑",
+    "&prod;": "∏",
+    "&radic;": "√",
+    "&prop;": "∝",
+    "&part;": "∂",
+    "&int;": "∫",
+    "&ang;": "∠",
+    "&perp;": "⊥",
+    "&parallel;": "∥",
+    // ギリシャ文字
+    "&Alpha;": "Α",
+    "&alpha;": "α",
+    "&Beta;": "Β",
+    "&beta;": "β",
+    "&Gamma;": "Γ",
+    "&gamma;": "γ",
+    "&Delta;": "Δ",
+    "&delta;": "δ",
+    "&Epsilon;": "Ε",
+    "&epsilon;": "ε",
+    "&Zeta;": "Ζ",
+    "&zeta;": "ζ",
+    "&Eta;": "Η",
+    "&eta;": "η",
+    "&Theta;": "Θ",
+    "&theta;": "θ",
+    "&Iota;": "Ι",
+    "&iota;": "ι",
+    "&Kappa;": "Κ",
+    "&kappa;": "κ",
+    "&Lambda;": "Λ",
+    "&lambda;": "λ",
+    "&Mu;": "Μ",
+    "&mu;": "μ",
+    "&Nu;": "Ν",
+    "&nu;": "ν",
+    "&Xi;": "Ξ",
+    "&xi;": "ξ",
+    "&Omicron;": "Ο",
+    "&omicron;": "ο",
+    "&Pi;": "Π",
+    "&pi;": "π",
+    "&Rho;": "Ρ",
+    "&rho;": "ρ",
+    "&Sigma;": "Σ",
+    "&sigma;": "σ",
+    "&Tau;": "Τ",
+    "&tau;": "τ",
+    "&Upsilon;": "Υ",
+    "&upsilon;": "υ",
+    "&Phi;": "Φ",
+    "&phi;": "φ",
+    "&Chi;": "Χ",
+    "&chi;": "χ",
+    "&Psi;": "Ψ",
+    "&psi;": "ψ",
+    "&Omega;": "Ω",
+    "&omega;": "ω",
+    // 通貨記号
+    "&euro;": "€",
+    "&yen;": "¥",
+    "&pound;": "£",
+    "&cent;": "¢",
+    "&curren;": "¤",
+    // 著作権・商標（大文字小文字両対応）
+    "&copy;": "©",
+    "&COPY;": "©",
+    "&reg;": "®",
+    "&REG;": "®",
+    "&trade;": "™",
+    "&TRADE;": "™",
+    // 句読点・記号
+    "&hellip;": "…",
+    "&mdash;": "—",
+    "&ndash;": "–",
+    "&lsquo;": "‘",
+    "&rsquo;": "’",
+    "&ldquo;": "“",
+    "&rdquo;": "”",
+    "&bull;": "•",
+    "&middot;": "·",
+    "&sect;": "§",
+    "&para;": "¶",
+    "&dagger;": "†",
+    "&Dagger;": "‡",
+    "&permil;": "‰",
+    "&lsaquo;": "‹",
+    "&rsaquo;": "›",
+    "&laquo;": "«",
+    "&raquo;": "»",
+    // アクセント文字（ラテン文字）
+    "&Agrave;": "À",
+    "&agrave;": "à",
+    "&Aacute;": "Á",
+    "&aacute;": "á",
+    "&Acirc;": "Â",
+    "&acirc;": "â",
+    "&Atilde;": "Ã",
+    "&atilde;": "ã",
+    "&Auml;": "Ä",
+    "&auml;": "ä",
+    "&Aring;": "Å",
+    "&aring;": "å",
+    "&AElig;": "Æ",
+    "&aelig;": "æ",
+    "&Ccedil;": "Ç",
+    "&ccedil;": "ç",
+    "&Egrave;": "È",
+    "&egrave;": "è",
+    "&Eacute;": "É",
+    "&eacute;": "é",
+    "&Ecirc;": "Ê",
+    "&ecirc;": "ê",
+    "&Euml;": "Ë",
+    "&euml;": "ë",
+    "&Igrave;": "Ì",
+    "&igrave;": "ì",
+    "&Iacute;": "Í",
+    "&iacute;": "í",
+    "&Icirc;": "Î",
+    "&icirc;": "î",
+    "&Iuml;": "Ï",
+    "&iuml;": "ï",
+    "&ETH;": "Ð",
+    "&eth;": "ð",
+    "&Ntilde;": "Ñ",
+    "&ntilde;": "ñ",
+    "&Ograve;": "Ò",
+    "&ograve;": "ò",
+    "&Oacute;": "Ó",
+    "&oacute;": "ó",
+    "&Ocirc;": "Ô",
+    "&ocirc;": "ô",
+    "&Otilde;": "Õ",
+    "&otilde;": "õ",
+    "&Ouml;": "Ö",
+    "&ouml;": "ö",
+    "&Oslash;": "Ø",
+    "&oslash;": "ø",
+    "&Ugrave;": "Ù",
+    "&ugrave;": "ù",
+    "&Uacute;": "Ú",
+    "&uacute;": "ú",
+    "&Ucirc;": "Û",
+    "&ucirc;": "û",
+    "&Uuml;": "Ü",
+    "&uuml;": "ü",
+    "&Yacute;": "Ý",
+    "&yacute;": "ý",
+    "&THORN;": "Þ",
+    "&thorn;": "þ",
+    "&szlig;": "ß",
+    "&yuml;": "ÿ",
+    // 矢印
+    "&larr;": "←",
+    "&uarr;": "↑",
+    "&rarr;": "→",
+    "&darr;": "↓",
+    "&harr;": "↔",
+    "&crarr;": "↵",
+    "&lArr;": "⇐",
+    "&uArr;": "⇑",
+    "&rArr;": "⇒",
+    "&dArr;": "⇓",
+    "&hArr;": "⇔",
+    // 数学記号（追加）
+    "&forall;": "∀",
+    "&exist;": "∃",
+    "&empty;": "∅",
+    "&nabla;": "∇",
+    "&isin;": "∈",
+    "&notin;": "∉",
+    "&ni;": "∋",
+    "&cap;": "∩",
+    "&cup;": "∪",
+    "&sub;": "⊂",
+    "&sup;": "⊃",
+    "&nsub;": "⊄",
+    "&sube;": "⊆",
+    "&supe;": "⊇",
+    "&oplus;": "⊕",
+    "&otimes;": "⊗",
+    "&equiv;": "≡",
+    "&ne;": "≠",
+    "&le;": "≤",
+    "&ge;": "≥",
+    "&sim;": "∼",
+    "&cong;": "≅",
+    "&asymp;": "≈",
+    // スペード・ハート・ダイヤ・クラブ
+    "&spades;": "♠",
+    "&hearts;": "♥",
+    "&diams;": "♦",
+    "&clubs;": "♣",
+    // 追加の特殊文字
+    "&loz;": "◊",
+    "&image;": "ℑ",
+    "&real;": "ℜ",
+    "&weierp;": "℘",
+    "&alefsym;": "ℵ",
+    // 日本語・中国語・韓国語特有のエンティティ
+    "&#xff08;": "（",
+    // 全角左括弧
+    "&#xff09;": "）",
+    // 全角右括弧
+    "&#x30b7;": "シ",
+    // カタカナのシ
+    "&#x30B7;": "シ",
+    // カタカナのシ（大文字）
+    "&#x3042;": "あ",
+    // ひらがなのあ
+    "&#x3044;": "い",
+    // ひらがなのい
+    "&#x3046;": "う",
+    // ひらがなのう
+    "&#x3048;": "え",
+    // ひらがなのえ
+    "&#x304a;": "お",
+    // ひらがなのお
+    // 中国語の文字
+    "&#x4e2d;": "中",
+    // 中
+    "&#x6587;": "文",
+    // 文
+    "&#x56fd;": "国",
+    // 国
+    // 韓国語の文字
+    "&#xd55c;": "한",
+    // 한
+    "&#xad6d;": "국",
+    // 국
+    "&#xc5b4;": "어",
+    // 어
+    // その他のUnicode文字
+    "&ensp;": " ",
+    // En space
+    "&emsp;": " ",
+    // Em space
+    "&thinsp;": " ",
+    // Thin space
+    "&zwnj;": "‌",
+    // Zero width non-joiner
+    "&zwj;": "‍",
+    // Zero width joiner
+    "&lrm;": "‎",
+    // Left-to-right mark
+    "&rlm;": "‏"
+    // Right-to-left mark
+  };
+}
 const blockLevelTags = /* @__PURE__ */ new Set([
   "address",
   "article",
@@ -20687,7 +21016,8 @@ const handleTableCell = (_state, node2, _parent) => {
 const convertHtmlToMarkdown = async (html2, baseUrl, locale) => {
   console.info(getMessage(locale, "convert_start"));
   const slugger2 = new BananaSlug();
-  const file = await unified().use(rehypeParse, { fragment: true }).use(rehypeRaw).use(rehypeSanitizeHtml).use(rehypeWikipediaFootnotes).use(rehypeNamedAnchors).use(rehypeSlug).use(rehypeParagraphWrapper).use(rehypeAbsoluteLinks, { baseUrl }).use(rehypeRemark, {
+  const preprocessedHtml = preprocessHtmlEntities(html2);
+  const file = await unified().use(rehypeParse, { fragment: true }).use(rehypeRaw).use(rehypeDecodeEntities).use(rehypeSanitizeHtml).use(rehypeWikipediaFootnotes).use(rehypeNamedAnchors).use(rehypeSlug).use(rehypeParagraphWrapper).use(rehypeAbsoluteLinks, { baseUrl }).use(rehypeRemark, {
     handlers: {
       span(state, node2) {
         const id = node2.properties?.id;
@@ -20784,11 +21114,12 @@ const convertHtmlToMarkdown = async (html2, baseUrl, locale) => {
         return value.replace(/\n/g, "").replace(/\|/g, "\\|");
       }
     }
-  }).process(html2);
+  }).process(preprocessedHtml);
   const unescapedMarkdown = String(file).replace(/\\\[\^(.+?)\\?\](?!\s*:)/g, "[^$1]").replace(/\]\(\\#/g, "](#");
   const correctedMarkdown = unescapedMarkdown.replace(/(\s*)- - /g, "$1- ").replace(/(\s*)\* \* /g, "$1* ");
+  const finalMarkdown = preprocessHtmlEntities(correctedMarkdown);
   console.info(getMessage(locale, "convert_success"));
-  return correctedMarkdown;
+  return finalMarkdown;
 };
 const htmlToMarkdown = async (urlOrHtml, options = {}) => {
   const { locale = "en-US", htmlContent } = options;

package/dist/index.cjs

@@ -1 +1 @@
-const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e}),Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const r=require("./core-Dq4LfiYy.cjs"),t=require("./mcp-server-CMBUHOmc.cjs"),c=require("./mcp-server-streamable-IQlimU2H.cjs");exports.getMessage=r.getMessage,exports.htmlToMarkdown=r.htmlToMarkdown,exports.createMcpServer=t.createMcpServer,exports.runMcpServer=t.runMcpServer,exports.createStreamableMcpServer=c.createStreamableMcpServer,exports.runStreamableMcpServer=c.runStreamableMcpServer;
+const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e}),Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const r=require("./core--XIBQvpW.cjs"),t=require("./mcp-server-Dn3EHzcf.cjs"),c=require("./mcp-server-streamable-DUB_ixZe.cjs");exports.getMessage=r.getMessage,exports.htmlToMarkdown=r.htmlToMarkdown,exports.createMcpServer=t.createMcpServer,exports.runMcpServer=t.runMcpServer,exports.createStreamableMcpServer=c.createStreamableMcpServer,exports.runStreamableMcpServer=c.runStreamableMcpServer;

package/dist/index.js

@@ -5,9 +5,9 @@ Object.assign(global, {
   window,
   self: window
 });
-import { g, h } from "./core-DJV8t_Qj.js";
-import { c, r } from "./mcp-server-DD8CqdNw.js";
-import { c as c2, r as r2 } from "./mcp-server-streamable-DODolSbj.js";
+import { g, h } from "./core-DhxxOcPm.js";
+import { c, r } from "./mcp-server-DOe8W9IZ.js";
+import { c as c2, r as r2 } from "./mcp-server-streamable-pcpU_lDl.js";
 export {
   c as createMcpServer,
   c2 as createStreamableMcpServer,

package/dist/main.cjs

@@ -1,2 +1,2 @@
 #!/usr/bin/env node
-Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});const t=require("./core-Dq4LfiYy.cjs");var r="undefined"!=typeof document?document.currentScript:null;(async()=>{(("undefined"==typeof document?require("url").pathToFileURL(__filename).href:r&&"SCRIPT"===r.tagName.toUpperCase()&&r.src||new URL("main.cjs",document.baseURI).href).endsWith(process.argv[1])||process.argv[1]&&process.argv[1].includes("vite-node"))&&(async()=>{const e=(await import("yargs-parser")).default,r=await import("node:fs"),o=await import("node:path"),c=e(process.argv.slice(2),{string:["locale","output","html-content"],alias:{output:["o"],"html-content":["h"]},default:{locale:"en-US"}}),n=c._[0],s=c.locale,a=c.output,i=c["html-content"];if(n||i||process.exit(1),n&&!i)try{new URL(n)}catch(p){process.exit(1)}try{const{markdown:e,html:c}=await t.htmlToMarkdown(n||"direct-html-input",{locale:s,htmlContent:i});let p;if(a){p=o.resolve(process.cwd(),a);const e=o.dirname(p);r.existsSync(e)||r.mkdirSync(e,{recursive:!0})}else{let e;e=i?"html_content":n.replace(/https?:\/\//,"").replace(/[^a-zA-Z0-9]/g,"_");const t=`${e}_${Date.now()}.md`,c=o.join(process.cwd(),".outputs","raw");r.existsSync(c)||r.mkdirSync(c,{recursive:!0}),p=o.join(c,t)}r.writeFileSync(p,e);const l=o.parse(p),d=o.join(l.dir,`${l.name}.html`);r.writeFileSync(d,c)}catch(p){process.exit(1)}})()})();
+Object.create,Object.defineProperty,Object.getOwnPropertyDescriptor,Object.getOwnPropertyNames,Object.getPrototypeOf,Object.prototype.hasOwnProperty;const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});const t=require("./core--XIBQvpW.cjs");var r="undefined"!=typeof document?document.currentScript:null;(async()=>{(("undefined"==typeof document?require("url").pathToFileURL(__filename).href:r&&"SCRIPT"===r.tagName.toUpperCase()&&r.src||new URL("main.cjs",document.baseURI).href).endsWith(process.argv[1])||process.argv[1]&&process.argv[1].includes("vite-node"))&&(async()=>{const e=(await import("yargs-parser")).default,r=await import("node:fs"),o=await import("node:path"),c=e(process.argv.slice(2),{string:["locale","output","html-content"],alias:{output:["o"],"html-content":["h"]},default:{locale:"en-US"}}),n=c._[0],s=c.locale,a=c.output,i=c["html-content"];if(n||i||process.exit(1),n&&!i)try{new URL(n)}catch(p){process.exit(1)}try{const{markdown:e,html:c}=await t.htmlToMarkdown(n||"direct-html-input",{locale:s,htmlContent:i});let p;if(a){p=o.resolve(process.cwd(),a);const e=o.dirname(p);r.existsSync(e)||r.mkdirSync(e,{recursive:!0})}else{let e;e=i?"html_content":n.replace(/https?:\/\//,"").replace(/[^a-zA-Z0-9]/g,"_");const t=`${e}_${Date.now()}.md`,c=o.join(process.cwd(),".outputs","raw");r.existsSync(c)||r.mkdirSync(c,{recursive:!0}),p=o.join(c,t)}r.writeFileSync(p,e);const l=o.parse(p),d=o.join(l.dir,`${l.name}.html`);r.writeFileSync(d,c)}catch(p){process.exit(1)}})()})();

package/dist/main.js

@@ -6,7 +6,7 @@ Object.assign(global, {
   window,
   self: window
 });
-import { g as getMessage, h as htmlToMarkdown } from "./core-DJV8t_Qj.js";
+import { g as getMessage, h as htmlToMarkdown } from "./core-DhxxOcPm.js";
 const runCli = async () => {
   const yargsParser = (await import("yargs-parser")).default;
   const fs = await import("node:fs");

package/dist/{mcp-server-DD8CqdNw.js → mcp-server-DOe8W9IZ.js}

@@ -5,9 +5,9 @@ Object.assign(global, {
   window,
   self: window
 });
-import { J as JSONRPCMessageSchema, g as getPackageVersion, S as Server, L as ListToolsRequestSchema, M as MCP_TOOL_SCHEMAS, C as CallToolRequestSchema, p as parseHtmlToMarkdownArgs, c as createSuccessResult, a as createErrorResult, b as parseSaveContentArgs, s as saveContentToFile, d as parseUrlToMarkdownFileArgs } from "./version-CTdrYd1G.js";
+import { J as JSONRPCMessageSchema, g as getPackageVersion, S as Server, L as ListToolsRequestSchema, M as MCP_TOOL_SCHEMAS, C as CallToolRequestSchema, p as parseHtmlToMarkdownArgs, c as createSuccessResult, a as createErrorResult, b as parseSaveContentArgs, s as saveContentToFile, d as parseUrlToMarkdownFileArgs } from "./version-BNwQMbri.js";
 import process from "node:process";
-import { h as htmlToMarkdown } from "./core-DJV8t_Qj.js";
+import { h as htmlToMarkdown } from "./core-DhxxOcPm.js";
 class ReadBuffer {
   append(chunk) {
     this._buffer = this._buffer ? Buffer.concat([this._buffer, chunk]) : chunk;

package/dist/{mcp-server-CMBUHOmc.cjs → mcp-server-Dn3EHzcf.cjs}

@@ -1 +1 @@
-const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});const t=require("./version-BkgGdAYt.cjs"),r=require("node:process"),a=require("./core-Dq4LfiYy.cjs");class n{append(e){this._buffer=this._buffer?Buffer.concat([this._buffer,e]):e}readMessage(){if(!this._buffer)return null;const e=this._buffer.indexOf("\n");if(-1===e)return null;const r=this._buffer.toString("utf8",0,e).replace(/\r$/,"");return this._buffer=this._buffer.subarray(e+1),function(e){return t.JSONRPCMessageSchema.parse(JSON.parse(e))}(r)}clear(){this._buffer=void 0}}class s{constructor(e=r.stdin,t=r.stdout){this._stdin=e,this._stdout=t,this._readBuffer=new n,this._started=!1,this._ondata=e=>{this._readBuffer.append(e),this.processReadBuffer()},this._onerror=e=>{var t;null===(t=this.onerror)||void 0===t||t.call(this,e)}}async start(){if(this._started)throw new Error("StdioServerTransport already started! If using Server class, note that connect() calls start() automatically.");this._started=!0,this._stdin.on("data",this._ondata),this._stdin.on("error",this._onerror)}processReadBuffer(){for(var e,t;;)try{const t=this._readBuffer.readMessage();if(null===t)break;null===(e=this.onmessage)||void 0===e||e.call(this,t)}catch(r){null===(t=this.onerror)||void 0===t||t.call(this,r)}}async close(){var e;this._stdin.off("data",this._ondata),this._stdin.off("error",this._onerror);0===this._stdin.listenerCount("data")&&this._stdin.pause(),this._readBuffer.clear(),null===(e=this.onclose)||void 0===e||e.call(this)}send(e){return new Promise(t=>{const r=function(e){return JSON.stringify(e)+"\n"}(e);this._stdout.write(r)?t():this._stdout.once("drain",t)})}}class o{constructor(){this.version=t.getPackageVersion(),this.server=new t.Server({name:"@aiquants/html-to-markdown",version:this.version,description:"Convert web pages and HTML to Markdown with JavaScript dynamic content support and table structure preservation. Includes separate file saving tool."},{capabilities:{tools:{}}}),this.setupToolHandlers()}setupToolHandlers(){this.server.setRequestHandler(t.ListToolsRequestSchema,async()=>({tools:[{name:"html_to_markdown",description:"Convert web pages or HTML strings to Markdown format. Handles JavaScript-rendered dynamic content and preserves table structures.",inputSchema:t.MCP_TOOL_SCHEMAS.htmlToMarkdown},{name:"save_content_to_file",description:"Save text content to a file with specified path or directory. Supports automatic filename generation and format detection.",inputSchema:t.MCP_TOOL_SCHEMAS.saveContentToFile},{name:"url_to_markdown_file",description:"Convert web pages or HTML strings directly to Markdown files. Combines HTML-to-Markdown conversion and file saving in one step. Handles JavaScript-rendered dynamic content.",inputSchema:t.MCP_TOOL_SCHEMAS.urlToMarkdownFile}]})),this.server.setRequestHandler(t.CallToolRequestSchema,async e=>{switch(e.params.name){case"html_to_markdown":return await this.handleHtmlToMarkdown(e.params.arguments);case"save_content_to_file":return await this.handleSaveContentToFile(e.params.arguments);case"url_to_markdown_file":return await this.handleUrlToMarkdownFile(e.params.arguments);default:throw new Error(`Unknown tool: ${e.params.name}`)}})}async handleHtmlToMarkdown(e){try{const r=t.parseHtmlToMarkdownArgs(e),{url:n,html_content:s,locale:o="en-US"}=r;if(n)try{new URL(n)}catch{throw new Error(`Invalid URL: ${n}`)}const i=(await a.htmlToMarkdown(n||"direct-html-input",{locale:o,htmlContent:s})).markdown;return t.createSuccessResult(i)}catch(r){return t.createErrorResult(r)}}async handleSaveContentToFile(e){try{const r=t.parseSaveContentArgs(e),{content:a,save_path:n,save_directory:s,filename:o}=r,i=`File saved successfully to: ${await t.saveContentToFile(a,n,s,o)}`;return t.createSuccessResult(i)}catch(r){return t.createErrorResult(r)}}async handleUrlToMarkdownFile(e){try{const r=t.parseUrlToMarkdownFileArgs(e),{url:n,html_content:s,locale:o="en-US",save_path:i,save_directory:c,filename:l}=r;if(n)try{new URL(n)}catch{throw new Error(`Invalid URL: ${n}`)}const d=(await a.htmlToMarkdown(n||"direct-html-input",{locale:o,htmlContent:s})).markdown;let u=l;!u&&n&&(u=this.generateFilenameFromUrl(n));const h=`URL successfully converted to Markdown and saved to: ${await t.saveContentToFile(d,i,c,u)}`;return t.createSuccessResult(h)}catch(r){return t.createErrorResult(r)}}generateFilenameFromUrl(e){try{let t=new URL(e).pathname.split("/").pop()||"page";return t=t.replace(/\.[^.]*$/,""),t=t.replace(/[<>:"/\\|?*]/g,"_"),t&&!t.match(/^\.+$/)||(t="page"),t}catch{return"page"}}async start(){const e=new s;await this.server.connect(e)}}const i=()=>new o;exports.createMcpServer=i,exports.runMcpServer=async()=>{const e=i();await e.start()};
+const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});const t=require("./version-BsWiDlbg.cjs"),r=require("node:process"),a=require("./core--XIBQvpW.cjs");class n{append(e){this._buffer=this._buffer?Buffer.concat([this._buffer,e]):e}readMessage(){if(!this._buffer)return null;const e=this._buffer.indexOf("\n");if(-1===e)return null;const r=this._buffer.toString("utf8",0,e).replace(/\r$/,"");return this._buffer=this._buffer.subarray(e+1),function(e){return t.JSONRPCMessageSchema.parse(JSON.parse(e))}(r)}clear(){this._buffer=void 0}}class s{constructor(e=r.stdin,t=r.stdout){this._stdin=e,this._stdout=t,this._readBuffer=new n,this._started=!1,this._ondata=e=>{this._readBuffer.append(e),this.processReadBuffer()},this._onerror=e=>{var t;null===(t=this.onerror)||void 0===t||t.call(this,e)}}async start(){if(this._started)throw new Error("StdioServerTransport already started! If using Server class, note that connect() calls start() automatically.");this._started=!0,this._stdin.on("data",this._ondata),this._stdin.on("error",this._onerror)}processReadBuffer(){for(var e,t;;)try{const t=this._readBuffer.readMessage();if(null===t)break;null===(e=this.onmessage)||void 0===e||e.call(this,t)}catch(r){null===(t=this.onerror)||void 0===t||t.call(this,r)}}async close(){var e;this._stdin.off("data",this._ondata),this._stdin.off("error",this._onerror);0===this._stdin.listenerCount("data")&&this._stdin.pause(),this._readBuffer.clear(),null===(e=this.onclose)||void 0===e||e.call(this)}send(e){return new Promise(t=>{const r=function(e){return JSON.stringify(e)+"\n"}(e);this._stdout.write(r)?t():this._stdout.once("drain",t)})}}class o{constructor(){this.version=t.getPackageVersion(),this.server=new t.Server({name:"@aiquants/html-to-markdown",version:this.version,description:"Convert web pages and HTML to Markdown with JavaScript dynamic content support and table structure preservation. Includes separate file saving tool."},{capabilities:{tools:{}}}),this.setupToolHandlers()}setupToolHandlers(){this.server.setRequestHandler(t.ListToolsRequestSchema,async()=>({tools:[{name:"html_to_markdown",description:"Convert web pages or HTML strings to Markdown format. Handles JavaScript-rendered dynamic content and preserves table structures.",inputSchema:t.MCP_TOOL_SCHEMAS.htmlToMarkdown},{name:"save_content_to_file",description:"Save text content to a file with specified path or directory. Supports automatic filename generation and format detection.",inputSchema:t.MCP_TOOL_SCHEMAS.saveContentToFile},{name:"url_to_markdown_file",description:"Convert web pages or HTML strings directly to Markdown files. Combines HTML-to-Markdown conversion and file saving in one step. Handles JavaScript-rendered dynamic content.",inputSchema:t.MCP_TOOL_SCHEMAS.urlToMarkdownFile}]})),this.server.setRequestHandler(t.CallToolRequestSchema,async e=>{switch(e.params.name){case"html_to_markdown":return await this.handleHtmlToMarkdown(e.params.arguments);case"save_content_to_file":return await this.handleSaveContentToFile(e.params.arguments);case"url_to_markdown_file":return await this.handleUrlToMarkdownFile(e.params.arguments);default:throw new Error(`Unknown tool: ${e.params.name}`)}})}async handleHtmlToMarkdown(e){try{const r=t.parseHtmlToMarkdownArgs(e),{url:n,html_content:s,locale:o="en-US"}=r;if(n)try{new URL(n)}catch{throw new Error(`Invalid URL: ${n}`)}const i=(await a.htmlToMarkdown(n||"direct-html-input",{locale:o,htmlContent:s})).markdown;return t.createSuccessResult(i)}catch(r){return t.createErrorResult(r)}}async handleSaveContentToFile(e){try{const r=t.parseSaveContentArgs(e),{content:a,save_path:n,save_directory:s,filename:o}=r,i=`File saved successfully to: ${await t.saveContentToFile(a,n,s,o)}`;return t.createSuccessResult(i)}catch(r){return t.createErrorResult(r)}}async handleUrlToMarkdownFile(e){try{const r=t.parseUrlToMarkdownFileArgs(e),{url:n,html_content:s,locale:o="en-US",save_path:i,save_directory:c,filename:l}=r;if(n)try{new URL(n)}catch{throw new Error(`Invalid URL: ${n}`)}const d=(await a.htmlToMarkdown(n||"direct-html-input",{locale:o,htmlContent:s})).markdown;let u=l;!u&&n&&(u=this.generateFilenameFromUrl(n));const h=`URL successfully converted to Markdown and saved to: ${await t.saveContentToFile(d,i,c,u)}`;return t.createSuccessResult(h)}catch(r){return t.createErrorResult(r)}}generateFilenameFromUrl(e){try{let t=new URL(e).pathname.split("/").pop()||"page";return t=t.replace(/\.[^.]*$/,""),t=t.replace(/[<>:"/\\|?*]/g,"_"),t&&!t.match(/^\.+$/)||(t="page"),t}catch{return"page"}}async start(){const e=new s;await this.server.connect(e)}}const i=()=>new o;exports.createMcpServer=i,exports.runMcpServer=async()=>{const e=i();await e.start()};