@aiquants/html-to-markdown 0.2.4 → 0.2.6

package/dist/{core-CFnkfU5N.js → core-CouDTni-.js}

@@ -61,12 +61,12 @@ const getMessage = (locale, key2, replacements) => {
 const getHtmlWithPlaywright = async (url, locale) => {
   let browser;
   try {
-    console.error(getMessage(locale, "playwright_launch"));
+    console.info(getMessage(locale, "playwright_launch"));
     browser = await chromium.launch();
     const page = await browser.newPage({ locale });
-    console.error(getMessage(locale, "playwright_goto", { url, locale }));
+    console.info(getMessage(locale, "playwright_goto", { url, locale }));
     await page.goto(url, { waitUntil: "networkidle" });
-    console.error(getMessage(locale, "playwright_get_content"));
+    console.info(getMessage(locale, "playwright_get_content"));
     const html2 = await page.content();
     return html2;
   } catch (error) {
@@ -75,7 +75,7 @@ const getHtmlWithPlaywright = async (url, locale) => {
   } finally {
     if (browser) {
       await browser.close();
-      console.error(getMessage(locale, "playwright_close"));
+      console.info(getMessage(locale, "playwright_close"));
     }
   }
 };
@@ -20681,7 +20681,7 @@ const handleTableCell = (_state, node2, _parent) => {
   return { type: "tableCell", children: [{ type: "html", value }] };
 };
 const convertHtmlToMarkdown = async (html2, baseUrl, locale) => {
-  console.error(getMessage(locale, "convert_start"));
+  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, {
     handlers: {
@@ -20783,7 +20783,7 @@ const convertHtmlToMarkdown = async (html2, baseUrl, locale) => {
   }).process(html2);
   const unescapedMarkdown = String(file).replace(/\\\[\^(.+?)\\?\](?!\s*:)/g, "[^$1]").replace(/\]\(\\#/g, "](#");
   const correctedMarkdown = unescapedMarkdown.replace(/(\s*)- - /g, "$1- ").replace(/(\s*)\* \* /g, "$1* ");
-  console.error(getMessage(locale, "convert_success"));
+  console.info(getMessage(locale, "convert_success"));
   return correctedMarkdown;
 };
 const htmlToMarkdown = async (urlOrHtml, 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-D2dlEt1l.cjs"),t=require("./mcp-server-B5lrAMe3.cjs"),c=require("./mcp-server-streamable-B3R3qi85.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-D2dlEt1l.cjs"),t=require("./mcp-server-Bo_mCvzz.cjs"),c=require("./mcp-server-streamable-Cc4QsZQe.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-CFnkfU5N.js";
-import { c, r } from "./mcp-server-BEggrdWH.js";
-import { c as c2, r as r2 } from "./mcp-server-streamable-CuBJeQTj.js";
+import { g, h } from "./core-CouDTni-.js";
+import { c, r } from "./mcp-server-B8hz3lXI.js";
+import { c as c2, r as r2 } from "./mcp-server-streamable-3B3B1fuS.js";
 export {
   c as createMcpServer,
   c2 as createStreamableMcpServer,

package/dist/main.js

@@ -6,7 +6,7 @@ Object.assign(global, {
   window,
   self: window
 });
-import { g as getMessage, h as htmlToMarkdown } from "./core-CFnkfU5N.js";
+import { g as getMessage, h as htmlToMarkdown } from "./core-CouDTni-.js";
 const runCli = async () => {
   const yargsParser = (await import("yargs-parser")).default;
   const fs = await import("node:fs");

package/dist/{mcp-server-BEggrdWH.js → mcp-server-B8hz3lXI.js}

@@ -5,11 +5,11 @@ Object.assign(global, {
   window,
   self: window
 });
-import { J as JSONRPCMessageSchema, g as getPackageVersion, S as Server, L as ListToolsRequestSchema, C as CallToolRequestSchema } from "./version-BXLLJx3Z.js";
+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-CFnkfU5N.js";
+import { h as htmlToMarkdown } from "./core-CouDTni-.js";
 class ReadBuffer {
   append(chunk) {
     this._buffer = this._buffer ? Buffer.concat([this._buffer, chunk]) : chunk;
@@ -178,7 +178,7 @@ class HtmlToMarkdownMcpServer {
       {
         name: "@aiquants/html-to-markdown",
         version: this.version,
-        description: "A tool to dynamically fetch a web page from a given URL and convert it to Markdown."
+        description: "Convert web pages and HTML to Markdown with JavaScript dynamic content support, table structure preservation, and file saving capabilities."
       },
       {
         capabilities: {
@@ -198,37 +198,37 @@ class HtmlToMarkdownMcpServer {
         tools: [
           {
             name: "html_to_markdown",
-            description: "Convert HTML content from a URL or HTML string to Markdown format. Supports dynamic content loading and preserves table structures.",
+            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: "The URL of the web page to convert to Markdown"
+                  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 content as a string to convert to Markdown (alternative to URL)"
+                  description: "HTML string to convert directly. Use instead of URL for local files or API response HTML content."
                 },
                 locale: {
                   type: "string",
-                  description: "The locale to use for browser context (en-US or ja-JP)",
+                  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' (default), 'html', or 'both'",
+                  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: "Full path (including filename) where to save the converted content. Cannot be used with save_directory."
+                  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 where to save the converted content with auto-generated filename. Cannot be used with save_path."
+                  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"] }]
@@ -348,7 +348,7 @@ ${savedPaths.map((path) => `- ${path}`).join("\n")}`;
   async start() {
     const transport = new StdioServerTransport();
     await this.server.connect(transport);
-    console.error("HTML to Markdown MCP server started");
+    console.info("HTML to Markdown MCP server started");
   }
 }
 const createMcpServer = () => {

package/dist/mcp-server-Bo_mCvzz.cjs

@@ -0,0 +1 @@
+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()};

package/dist/{mcp-server-streamable-CuBJeQTj.js → mcp-server-streamable-3B3B1fuS.js}

@@ -1,5 +1,5 @@
-import { J as JSONRPCMessageSchema, i as isInitializeRequest, a as isJSONRPCRequest, D as DEFAULT_NEGOTIATED_PROTOCOL_VERSION, b as SUPPORTED_PROTOCOL_VERSIONS, c as isJSONRPCResponse, d as isJSONRPCError, g as getPackageVersion, S as Server, I as InitializeRequestSchema, L as ListToolsRequestSchema, C as CallToolRequestSchema } from "./version-BXLLJx3Z.js";
-import { a as getAugmentedNamespace, c as commonjsGlobal, b as getDefaultExportFromCjs, h as htmlToMarkdown } from "./core-CFnkfU5N.js";
+import { J as JSONRPCMessageSchema, i as isInitializeRequest, a as isJSONRPCRequest, D as DEFAULT_NEGOTIATED_PROTOCOL_VERSION, b as SUPPORTED_PROTOCOL_VERSIONS, c as isJSONRPCResponse, d as isJSONRPCError, g as getPackageVersion, S as Server, I as InitializeRequestSchema, L as ListToolsRequestSchema, C as CallToolRequestSchema } from "./version-PCzxchJe.js";
+import { a as getAugmentedNamespace, c as commonjsGlobal, b as getDefaultExportFromCjs, h as htmlToMarkdown } from "./core-CouDTni-.js";
 import { randomUUID } from "node:crypto";
 import cors from "cors";
 import express from "express";
@@ -7201,7 +7201,7 @@ class StreamableHtmlToMarkdownMcpServer {
       {
         name: "@aiquants/html-to-markdown-streamable",
         version: this.version,
-        description: "A streamable tool to dynamically fetch a web page from a given URL and convert it to Markdown."
+        description: "Streamable HTML to Markdown converter with real-time progress updates and JavaScript dynamic content support."
       },
       {
         capabilities: {
@@ -7218,31 +7218,31 @@ class StreamableHtmlToMarkdownMcpServer {
       properties: {
         url: {
           type: "string",
-          description: "The URL of the web page to convert to Markdown"
+          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 content as a string to convert to Markdown (alternative to URL)"
+          description: "HTML string to convert directly. Use instead of URL for local files or API response HTML content."
         },
         locale: {
           type: "string",
-          description: "The locale to use for browser context (en-US or ja-JP)",
+          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' (default), 'html', or 'both'",
+          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: "Full path (including filename) where to save the converted content. Cannot be used with save_directory."
+          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 where to save the converted content with auto-generated filename. Cannot be used with save_path."
+          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"] }]
@@ -7250,7 +7250,7 @@ class StreamableHtmlToMarkdownMcpServer {
     this.tools = [
       {
         name: "html_to_markdown_streamable",
-        description: "Convert HTML content from a URL or HTML string to Markdown format with streamable support. Supports dynamic content loading and preserves table structures.",
+        description: "Streamable HTML to Markdown converter with real-time progress updates. Ideal for large pages and time-consuming conversions with JavaScript dynamic content support.",
         inputSchema: toolParameters,
         handler: (args, context) => this.handleHtmlToMarkdown(args, context)
       }
@@ -7466,7 +7466,7 @@ ${savedPaths.map((path) => `- ${path}`).join("\n")}`;
     app.use(express.json());
     await this.server.connect(this.transport);
     app.post("/mcp", async (req, res) => {
-      console.error("Received MCP request:", req.body);
+      console.info("Received MCP request:", req.body);
       try {
         res.setHeader("Content-Type", "application/json");
         const request = req.body;
@@ -7573,7 +7573,7 @@ ${savedPaths.map((path) => `- ${path}`).join("\n")}`;
       }
     });
     app.get("/mcp", async (_req, res) => {
-      console.error("Received GET MCP request");
+      console.info("Received GET MCP request");
       res.writeHead(405).end(
         JSON.stringify({
           jsonrpc: "2.0",
@@ -7586,7 +7586,7 @@ ${savedPaths.map((path) => `- ${path}`).join("\n")}`;
       );
     });
     app.delete("/mcp", async (_req, res) => {
-      console.error("Received DELETE MCP request");
+      console.info("Received DELETE MCP request");
       res.writeHead(405).end(
         JSON.stringify({
           jsonrpc: "2.0",
@@ -7599,19 +7599,19 @@ ${savedPaths.map((path) => `- ${path}`).join("\n")}`;
       );
     });
     const server = app.listen(port, host, () => {
-      console.error(`HTML to Markdown Streamable MCP server started on http://${host}:${port}/mcp`);
+      console.info(`HTML to Markdown Streamable MCP server started on http://${host}:${port}/mcp`);
     });
     process.on("SIGINT", async () => {
-      console.error("Shutting down server...");
+      console.info("Shutting down server...");
       try {
-        console.error("Closing transport");
+        console.info("Closing transport");
         await this.transport.close();
       } catch (error) {
         console.error("Error closing transport:", error);
       }
       await this.server.close();
       server.close(() => {
-        console.error("Server shutdown complete");
+        console.info("Server shutdown complete");
         process.exit(0);
       });
     });