@aiquants/html-to-markdown 0.2.6 → 0.3.0

package/dist/mcp-server-B8hz3lXI.js

@@ -1,364 +0,0 @@
-import { Window } from "happy-dom";
-const window = new Window();
-Object.assign(global, {
-  document: window.document,
-  window,
-  self: window
-});
-import { J as JSONRPCMessageSchema, g as getPackageVersion, S as Server, L as ListToolsRequestSchema, C as CallToolRequestSchema } from "./version-PCzxchJe.js";
-import process from "node:process";
-import { promises } from "fs";
-import { join, resolve, dirname } from "path";
-import { h as htmlToMarkdown } from "./core-CouDTni-.js";
-class ReadBuffer {
-  append(chunk) {
-    this._buffer = this._buffer ? Buffer.concat([this._buffer, chunk]) : chunk;
-  }
-  readMessage() {
-    if (!this._buffer) {
-      return null;
-    }
-    const index = this._buffer.indexOf("\n");
-    if (index === -1) {
-      return null;
-    }
-    const line = this._buffer.toString("utf8", 0, index).replace(/\r$/, "");
-    this._buffer = this._buffer.subarray(index + 1);
-    return deserializeMessage(line);
-  }
-  clear() {
-    this._buffer = void 0;
-  }
-}
-function deserializeMessage(line) {
-  return JSONRPCMessageSchema.parse(JSON.parse(line));
-}
-function serializeMessage(message) {
-  return JSON.stringify(message) + "\n";
-}
-class StdioServerTransport {
-  constructor(_stdin = process.stdin, _stdout = process.stdout) {
-    this._stdin = _stdin;
-    this._stdout = _stdout;
-    this._readBuffer = new ReadBuffer();
-    this._started = false;
-    this._ondata = (chunk) => {
-      this._readBuffer.append(chunk);
-      this.processReadBuffer();
-    };
-    this._onerror = (error) => {
-      var _a;
-      (_a = this.onerror) === null || _a === void 0 ? void 0 : _a.call(this, error);
-    };
-  }
-  /**
-   * Starts listening for messages on stdin.
-   */
-  async start() {
-    if (this._started) {
-      throw new Error("StdioServerTransport already started! If using Server class, note that connect() calls start() automatically.");
-    }
-    this._started = true;
-    this._stdin.on("data", this._ondata);
-    this._stdin.on("error", this._onerror);
-  }
-  processReadBuffer() {
-    var _a, _b;
-    while (true) {
-      try {
-        const message = this._readBuffer.readMessage();
-        if (message === null) {
-          break;
-        }
-        (_a = this.onmessage) === null || _a === void 0 ? void 0 : _a.call(this, message);
-      } catch (error) {
-        (_b = this.onerror) === null || _b === void 0 ? void 0 : _b.call(this, error);
-      }
-    }
-  }
-  async close() {
-    var _a;
-    this._stdin.off("data", this._ondata);
-    this._stdin.off("error", this._onerror);
-    const remainingDataListeners = this._stdin.listenerCount("data");
-    if (remainingDataListeners === 0) {
-      this._stdin.pause();
-    }
-    this._readBuffer.clear();
-    (_a = this.onclose) === null || _a === void 0 ? void 0 : _a.call(this);
-  }
-  send(message) {
-    return new Promise((resolve2) => {
-      const json = serializeMessage(message);
-      if (this._stdout.write(json)) {
-        resolve2();
-      } else {
-        this._stdout.once("drain", resolve2);
-      }
-    });
-  }
-}
-function parseHtmlToMarkdownArgs(args) {
-  if (!args || typeof args !== "object") {
-    throw new Error("Arguments are required");
-  }
-  const { url, html_content, locale = "en-US", output_format = "markdown", save_path, save_directory } = args;
-  if (!url && !html_content) {
-    throw new Error("Either 'url' or 'html_content' is required");
-  }
-  if (url && typeof url !== "string") {
-    throw new Error("'url' must be a string");
-  }
-  if (html_content && typeof html_content !== "string") {
-    throw new Error("'html_content' must be a string");
-  }
-  if (locale && !["en-US", "ja-JP"].includes(locale)) {
-    throw new Error("'locale' must be 'en-US' or 'ja-JP'");
-  }
-  if (output_format && !["markdown", "html", "both"].includes(output_format)) {
-    throw new Error("'output_format' must be 'markdown', 'html', or 'both'");
-  }
-  if (save_path && typeof save_path !== "string") {
-    throw new Error("'save_path' must be a string");
-  }
-  if (save_directory && typeof save_directory !== "string") {
-    throw new Error("'save_directory' must be a string");
-  }
-  if (save_path && save_directory) {
-    throw new Error("Cannot specify both 'save_path' and 'save_directory'");
-  }
-  return {
-    url,
-    html_content,
-    locale,
-    output_format,
-    save_path,
-    save_directory
-  };
-}
-function generateFilename(url, outputFormat) {
-  if (url && url !== "direct-html-input") {
-    try {
-      const urlObj = new URL(url);
-      let filename = urlObj.pathname.split("/").pop() || "page";
-      filename = filename.replace(/\.[^.]*$/, "");
-      filename = filename.replace(/[<>:"/\\|?*]/g, "_");
-      if (!filename || filename.match(/^\.+$/)) {
-        filename = "page";
-      }
-      return filename;
-    } catch {
-      return "page";
-    }
-  }
-  return "page";
-}
-async function saveContentToFile(content, savePath, saveDirectory, url, outputFormat) {
-  if (!savePath && !saveDirectory) {
-    return null;
-  }
-  let finalPath;
-  if (savePath) {
-    finalPath = resolve(savePath);
-  } else if (saveDirectory) {
-    const filename = generateFilename(url);
-    const extension = outputFormat === "html" ? ".html" : ".md";
-    finalPath = resolve(saveDirectory, `${filename}${extension}`);
-  } else {
-    return null;
-  }
-  await promises.mkdir(dirname(finalPath), { recursive: true });
-  await promises.writeFile(finalPath, content, "utf-8");
-  return finalPath;
-}
-class HtmlToMarkdownMcpServer {
-  constructor() {
-    this.version = getPackageVersion();
-    this.server = new Server(
-      {
-        name: "@aiquants/html-to-markdown",
-        version: this.version,
-        description: "Convert web pages and HTML to Markdown with JavaScript dynamic content support, table structure preservation, and file saving capabilities."
-      },
-      {
-        capabilities: {
-          tools: {}
-        }
-      }
-    );
-    this.setupToolHandlers();
-  }
-  /**
-   * Setup tool handlers for the MCP server.
-   * MCP サーバーのツールハンドラーを設定
-   */
-  setupToolHandlers() {
-    this.server.setRequestHandler(ListToolsRequestSchema, async () => {
-      return {
-        tools: [
-          {
-            name: "html_to_markdown",
-            description: "Convert web pages or HTML strings to Markdown format. Handles JavaScript-rendered dynamic content and preserves table structures with file saving capabilities.",
-            inputSchema: {
-              type: "object",
-              properties: {
-                url: {
-                  type: "string",
-                  description: "URL of the web page to convert (e.g., https://example.com). Handles JavaScript-rendered dynamic content effectively."
-                },
-                html_content: {
-                  type: "string",
-                  description: "HTML string to convert directly. Use instead of URL for local files or API response HTML content."
-                },
-                locale: {
-                  type: "string",
-                  description: "Browser locale setting. Use 'ja-JP' for Japanese sites, 'en-US' for English sites. Affects content rendering and language detection.",
-                  enum: ["en-US", "ja-JP"],
-                  default: "en-US"
-                },
-                output_format: {
-                  type: "string",
-                  description: "Output format: 'markdown' (Markdown only), 'html' (original HTML only), 'both' (returns both formats).",
-                  enum: ["markdown", "html", "both"],
-                  default: "markdown"
-                },
-                save_path: {
-                  type: "string",
-                  description: "Complete file path including filename to save converted content (e.g., '/home/user/output.md'). Cannot be used with save_directory."
-                },
-                save_directory: {
-                  type: "string",
-                  description: "Directory path to save converted content with auto-generated filename (e.g., '/home/user/downloads'). Cannot be used with save_path."
-                }
-              },
-              anyOf: [{ required: ["url"] }, { required: ["html_content"] }]
-            }
-          }
-        ]
-      };
-    });
-    this.server.setRequestHandler(CallToolRequestSchema, async (request) => {
-      switch (request.params.name) {
-        case "html_to_markdown": {
-          return await this.handleHtmlToMarkdown(request.params.arguments);
-        }
-        default:
-          throw new Error(`Unknown tool: ${request.params.name}`);
-      }
-    });
-  }
-  /**
-   * Handle HTML to Markdown conversion tool call.
-   * HTML から Markdown への変換ツール呼び出しを処理
-   */
-  async handleHtmlToMarkdown(args) {
-    try {
-      const validatedArgs = parseHtmlToMarkdownArgs(args);
-      const { url, html_content, locale = "en-US", output_format = "markdown", save_path, save_directory } = validatedArgs;
-      if (url) {
-        try {
-          new URL(url);
-        } catch {
-          throw new Error(`Invalid URL: ${url}`);
-        }
-      }
-      const result = await htmlToMarkdown(url || "direct-html-input", {
-        locale,
-        htmlContent: html_content
-      });
-      const savedPaths = [];
-      if (save_path || save_directory) {
-        switch (output_format) {
-          case "markdown": {
-            const savedPath = await saveContentToFile(result.markdown, save_path, save_directory, url, "markdown");
-            if (savedPath) savedPaths.push(savedPath);
-            break;
-          }
-          case "html": {
-            const savedPath = await saveContentToFile(result.html, save_path, save_directory, url, "html");
-            if (savedPath) savedPaths.push(savedPath);
-            break;
-          }
-          case "both": {
-            const markdownPath = save_path ? save_path.endsWith(".md") ? save_path : `${save_path}.md` : save_directory ? join(save_directory, `${generateFilename(url)}.md`) : void 0;
-            if (markdownPath) {
-              const savedMarkdownPath = await saveContentToFile(result.markdown, markdownPath, void 0, url, "markdown");
-              if (savedMarkdownPath) savedPaths.push(savedMarkdownPath);
-            }
-            const htmlPath = save_path ? save_path.endsWith(".html") ? save_path : `${save_path}.html` : save_directory ? join(save_directory, `${generateFilename(url)}.html`) : void 0;
-            if (htmlPath) {
-              const savedHtmlPath = await saveContentToFile(result.html, htmlPath, void 0, url, "html");
-              if (savedHtmlPath) savedPaths.push(savedHtmlPath);
-            }
-            break;
-          }
-        }
-      }
-      let responseText = "";
-      switch (output_format) {
-        case "markdown":
-          responseText = result.markdown;
-          break;
-        case "html":
-          responseText = result.html;
-          break;
-        case "both":
-          responseText = `=== Markdown ===
-${result.markdown}
-
-=== HTML ===
-${result.html}`;
-          break;
-        default:
-          throw new Error(`Invalid output format: ${output_format}`);
-      }
-      if (savedPaths.length > 0) {
-        responseText += `
-
-=== Files Saved ===
-${savedPaths.map((path) => `- ${path}`).join("\n")}`;
-      }
-      const content = [
-        {
-          type: "text",
-          text: responseText
-        }
-      ];
-      return {
-        content,
-        isError: false
-      };
-    } catch (error) {
-      const errorMessage = error instanceof Error ? error.message : String(error);
-      return {
-        content: [
-          {
-            type: "text",
-            text: `Error: ${errorMessage}`
-          }
-        ],
-        isError: true
-      };
-    }
-  }
-  /**
-   * Start the MCP server.
-   * MCP サーバーを開始
-   */
-  async start() {
-    const transport = new StdioServerTransport();
-    await this.server.connect(transport);
-    console.info("HTML to Markdown MCP server started");
-  }
-}
-const createMcpServer = () => {
-  return new HtmlToMarkdownMcpServer();
-};
-const runMcpServer = async () => {
-  const server = createMcpServer();
-  await server.start();
-};
-export {
-  createMcpServer as c,
-  runMcpServer as r
-};

package/dist/mcp-server-Bo_mCvzz.cjs

@@ -1 +0,0 @@
-const t=new(require("happy-dom").Window);Object.assign(global,{document:t.document,window:t,self:t});const e=require("./version-DTaUKNBS.cjs"),r=require("node:process"),n=require("fs"),o=require("path"),s=require("./core-D2dlEt1l.cjs");class a{append(t){this._buffer=this._buffer?Buffer.concat([this._buffer,t]):t}readMessage(){if(!this._buffer)return null;const t=this._buffer.indexOf("\n");if(-1===t)return null;const r=this._buffer.toString("utf8",0,t).replace(/\r$/,"");return this._buffer=this._buffer.subarray(t+1),function(t){return e.JSONRPCMessageSchema.parse(JSON.parse(t))}(r)}clear(){this._buffer=void 0}}class i{constructor(t=r.stdin,e=r.stdout){this._stdin=t,this._stdout=e,this._readBuffer=new a,this._started=!1,this._ondata=t=>{this._readBuffer.append(t),this.processReadBuffer()},this._onerror=t=>{var e;null===(e=this.onerror)||void 0===e||e.call(this,t)}}async start(){if(this._started)throw new Error("StdioServerTransport already started! If using Server class, note that connect() calls start() automatically.");this._started=!0,this._stdin.on("data",this._ondata),this._stdin.on("error",this._onerror)}processReadBuffer(){for(var t,e;;)try{const e=this._readBuffer.readMessage();if(null===e)break;null===(t=this.onmessage)||void 0===t||t.call(this,e)}catch(r){null===(e=this.onerror)||void 0===e||e.call(this,r)}}async close(){var t;this._stdin.off("data",this._ondata),this._stdin.off("error",this._onerror);0===this._stdin.listenerCount("data")&&this._stdin.pause(),this._readBuffer.clear(),null===(t=this.onclose)||void 0===t||t.call(this)}send(t){return new Promise(e=>{const r=function(t){return JSON.stringify(t)+"\n"}(t);this._stdout.write(r)?e():this._stdout.once("drain",e)})}}function c(t,e){if(t&&"direct-html-input"!==t)try{let e=new URL(t).pathname.split("/").pop()||"page";return e=e.replace(/\.[^.]*$/,""),e=e.replace(/[<>:"/\\|?*]/g,"_"),e&&!e.match(/^\.+$/)||(e="page"),e}catch{return"page"}return"page"}async function l(t,e,r,s,a){if(!e&&!r)return null;let i;if(e)i=o.resolve(e);else{if(!r)return null;{const t=c(s),e="html"===a?".html":".md";i=o.resolve(r,`${t}${e}`)}}return await n.promises.mkdir(o.dirname(i),{recursive:!0}),await n.promises.writeFile(i,t,"utf-8"),i}class d{constructor(){this.version=e.getPackageVersion(),this.server=new e.Server({name:"@aiquants/html-to-markdown",version:this.version,description:"Convert web pages and HTML to Markdown with JavaScript dynamic content support, table structure preservation, and file saving capabilities."},{capabilities:{tools:{}}}),this.setupToolHandlers()}setupToolHandlers(){this.server.setRequestHandler(e.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 with file saving capabilities.",inputSchema:{type:"object",properties:{url:{type:"string",description:"URL of the web page to convert (e.g., https://example.com). Handles JavaScript-rendered dynamic content effectively."},html_content:{type:"string",description:"HTML string to convert directly. Use instead of URL for local files or API response HTML content."},locale:{type:"string",description:"Browser locale setting. Use 'ja-JP' for Japanese sites, 'en-US' for English sites. Affects content rendering and language detection.",enum:["en-US","ja-JP"],default:"en-US"},output_format:{type:"string",description:"Output format: 'markdown' (Markdown only), 'html' (original HTML only), 'both' (returns both formats).",enum:["markdown","html","both"],default:"markdown"},save_path:{type:"string",description:"Complete file path including filename to save converted content (e.g., '/home/user/output.md'). Cannot be used with save_directory."},save_directory:{type:"string",description:"Directory path to save converted content with auto-generated filename (e.g., '/home/user/downloads'). Cannot be used with save_path."}},anyOf:[{required:["url"]},{required:["html_content"]}]}}]})),this.server.setRequestHandler(e.CallToolRequestSchema,async t=>{if("html_to_markdown"===t.params.name)return await this.handleHtmlToMarkdown(t.params.arguments);throw new Error(`Unknown tool: ${t.params.name}`)})}async handleHtmlToMarkdown(t){try{const e=function(t){if(!t||"object"!=typeof t)throw new Error("Arguments are required");const{url:e,html_content:r,locale:n="en-US",output_format:o="markdown",save_path:s,save_directory:a}=t;if(!e&&!r)throw new Error("Either 'url' or 'html_content' is required");if(e&&"string"!=typeof e)throw new Error("'url' must be a string");if(r&&"string"!=typeof r)throw new Error("'html_content' must be a string");if(n&&!["en-US","ja-JP"].includes(n))throw new Error("'locale' must be 'en-US' or 'ja-JP'");if(o&&!["markdown","html","both"].includes(o))throw new Error("'output_format' must be 'markdown', 'html', or 'both'");if(s&&"string"!=typeof s)throw new Error("'save_path' must be a string");if(a&&"string"!=typeof a)throw new Error("'save_directory' must be a string");if(s&&a)throw new Error("Cannot specify both 'save_path' and 'save_directory'");return{url:e,html_content:r,locale:n,output_format:o,save_path:s,save_directory:a}}(t),{url:r,html_content:n,locale:a="en-US",output_format:i="markdown",save_path:d,save_directory:h}=e;if(r)try{new URL(r)}catch{throw new Error(`Invalid URL: ${r}`)}const u=await s.htmlToMarkdown(r||"direct-html-input",{locale:a,htmlContent:n}),m=[];if(d||h)switch(i){case"markdown":{const t=await l(u.markdown,d,h,r,"markdown");t&&m.push(t);break}case"html":{const t=await l(u.html,d,h,r,"html");t&&m.push(t);break}case"both":{const t=d?d.endsWith(".md")?d:`${d}.md`:h?o.join(h,`${c(r)}.md`):void 0;if(t){const e=await l(u.markdown,t,void 0,r,"markdown");e&&m.push(e)}const e=d?d.endsWith(".html")?d:`${d}.html`:h?o.join(h,`${c(r)}.html`):void 0;if(e){const t=await l(u.html,e,void 0,r,"html");t&&m.push(t)}break}}let f="";switch(i){case"markdown":f=u.markdown;break;case"html":f=u.html;break;case"both":f=`=== Markdown ===\n${u.markdown}\n\n=== HTML ===\n${u.html}`;break;default:throw new Error(`Invalid output format: ${i}`)}m.length>0&&(f+=`\n\n=== Files Saved ===\n${m.map(t=>`- ${t}`).join("\n")}`);return{content:[{type:"text",text:f}],isError:!1}}catch(e){return{content:[{type:"text",text:`Error: ${e instanceof Error?e.message:String(e)}`}],isError:!0}}}async start(){const t=new i;await this.server.connect(t)}}const h=()=>new d;exports.createMcpServer=h,exports.runMcpServer=async()=>{const t=h();await t.start()};