@aiquants/html-to-markdown 0.1.7 → 0.2.0

package/dist/{mcp-server-streamable-CQD5zK30.js → mcp-server-streamable-C6-qOLdK.js}

@@ -3,6 +3,8 @@ import { a as getAugmentedNamespace, c as commonjsGlobal, b as getDefaultExportF
 import { randomUUID } from "node:crypto";
 import cors from "cors";
 import express from "express";
+import { promises } from "fs";
+import { join, resolve, dirname } from "path";
 import yargsParser from "yargs-parser";
 var bytes = { exports: {} };
 /*!
@@ -6416,10 +6418,10 @@ function requireRawBody() {
     if (done) {
       return readStream(stream, encoding, length, limit, wrap(done));
     }
-    return new Promise(function executor(resolve, reject) {
+    return new Promise(function executor(resolve2, reject) {
       readStream(stream, encoding, length, limit, function onRead(err, buf) {
         if (err) return reject(err);
-        resolve(buf);
+        resolve2(buf);
       });
     });
   }
@@ -7157,6 +7159,41 @@ class StreamableHTTPServerTransport {
     }
   }
 }
+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 StreamableHtmlToMarkdownMcpServer {
   constructor() {
     this.version = getPackageVersion();
@@ -7198,6 +7235,14 @@ class StreamableHtmlToMarkdownMcpServer {
           description: "Output format: 'markdown' (default), 'html', or 'both'",
           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."
+        },
+        save_directory: {
+          type: "string",
+          description: "Directory path where to save the converted content with auto-generated filename. Cannot be used with save_path."
         }
       },
       anyOf: [{ required: ["url"] }, { required: ["html_content"] }]
@@ -7294,10 +7339,19 @@ class StreamableHtmlToMarkdownMcpServer {
    */
   async *handleHtmlToMarkdown(args, _context) {
     try {
-      const { url, html_content, locale = "en-US", output_format = "markdown" } = args;
+      const { url, html_content, locale = "en-US", output_format = "markdown", save_path, save_directory } = args;
       if (!url && !html_content) {
         throw new Error("url または html_content のいずれかを提供する必要があります。");
       }
+      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'");
+      }
       if (url) {
         try {
           new URL(url);
@@ -7314,28 +7368,69 @@ class StreamableHtmlToMarkdownMcpServer {
         locale,
         htmlContent: html_content
       });
+      yield {
+        type: "progress",
+        progress: 50,
+        message: "変換が完了しました。ファイル保存を実行中..."
+      };
+      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;
+          }
+        }
+      }
       yield {
         type: "progress",
         progress: 100,
-        message: "変換が完了しました。"
+        message: "処理が完了しました。"
       };
-      let content;
+      let responseText = "";
       switch (output_format) {
         case "markdown":
-          content = [{ type: "text", text: result.markdown }];
+          responseText = result.markdown;
           break;
         case "html":
-          content = [{ type: "text", text: result.html }];
+          responseText = result.html;
           break;
         case "both":
-          content = [
-            { type: "text", text: result.markdown },
-            { type: "text", text: result.html }
-          ];
+          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 }];
       yield {
         type: "result",
         result: { content }

package/dist/mcp-streamable.cjs

@@ -1,2 +1,2 @@
 #!/usr/bin/env node
-"use strict";require("./mcp-server-streamable-Bu7QAr3W.cjs").runStreamableMcpServer().catch(e=>{process.exit(1)});
+"use strict";require("./mcp-server-streamable-B3R3qi85.cjs").runStreamableMcpServer().catch(e=>{process.exit(1)});

package/dist/mcp-streamable.js

@@ -1,5 +1,5 @@
 #!/usr/bin/env node
-import { r as runStreamableMcpServer } from "./mcp-server-streamable-CQD5zK30.js";
+import { r as runStreamableMcpServer } from "./mcp-server-streamable-C6-qOLdK.js";
 runStreamableMcpServer().catch((error) => {
   console.error("Failed to start streamable MCP server:", error);
   process.exit(1);

package/dist/mcp.cjs

@@ -1,2 +1,2 @@
 #!/usr/bin/env node
-const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});require("./mcp-server-D4Mj1IQ9.cjs").runMcpServer().catch(e=>{process.exit(1)});
+const e=new(require("happy-dom").Window);Object.assign(global,{document:e.document,window:e,self:e});require("./mcp-server-B5lrAMe3.cjs").runMcpServer().catch(e=>{process.exit(1)});

package/dist/mcp.js

@@ -6,7 +6,7 @@ Object.assign(global, {
   window,
   self: window
 });
-import { r as runMcpServer } from "./mcp-server-CpmnW1p8.js";
+import { r as runMcpServer } from "./mcp-server-Ct2TYHcU.js";
 runMcpServer().catch((error) => {
   console.error("Failed to start MCP server:", error);
   process.exit(1);

package/dist/src/core.d.ts

@@ -1,4 +1,4 @@
-import { HtmlToMarkdownOptions } from './types.js';
+import { HtmlToMarkdownOptions } from './types.ts';
 /**
  * Converts the HTML content of a given URL or HTML string to Markdown.
  * 指定されたURLまたはHTML文字列のHTMLコンテンツをMarkdownに変換します。

package/dist/src/i18n.d.ts

@@ -1,4 +1,4 @@
-import { MessageKey } from './types.js';
+import { MessageKey } from './types.ts';
 /**
  * Gets a localized message.
  * ローカライズされたメッセージを取得します。

package/dist/src/index.d.ts

@@ -2,8 +2,8 @@
  * Main export module for htmlToMarkdown library.
  * htmlToMarkdown ライブラリのメインエクスポートモジュール
  */
-export { htmlToMarkdown } from './core.js';
-export { getMessage } from './i18n.js';
-export type { HtmlToMarkdownOptions } from './types.js';
-export { createMcpServer, runMcpServer } from './mcp-server.js';
+export { htmlToMarkdown } from './core.ts';
+export { getMessage } from './i18n.ts';
+export type { HtmlToMarkdownOptions } from './types.ts';
+export { createMcpServer, runMcpServer } from './mcp-server.ts';
 export { createStreamableMcpServer, runStreamableMcpServer } from './mcp-server-streamable.ts';

package/dist/src/mcp-server-streamable.d.ts

@@ -16,6 +16,8 @@ interface HtmlToMarkdownArgs {
     html_content?: string;
     locale?: "en-US" | "ja-JP";
     output_format?: "markdown" | "html" | "both";
+    save_path?: string;
+    save_directory?: string;
 }
 /**
  * Extended Tool interface with handler function.

package/dist/src/mcp-server.d.ts

@@ -8,6 +8,8 @@ interface HtmlToMarkdownArgs {
     html_content?: string;
     locale?: "en-US" | "ja-JP";
     output_format?: "markdown" | "html" | "both";
+    save_path?: string;
+    save_directory?: string;
 }
 /**
  * Validate and parse arguments for HTML to Markdown conversion.

package/dist/src/plugins/index.d.ts

@@ -2,8 +2,8 @@
  * Rehype plugins index.
  * Rehype プラグインのインデックス
  */
-export { rehypeAbsoluteLinks } from './rehype-absolute-links.js';
-export { rehypeNamedAnchors } from './rehype-named-anchors.js';
-export { rehypeParagraphWrapper } from './rehype-paragraph-wrapper.js';
-export { rehypeSanitizeHtml } from './rehype-sanitize-html.js';
-export { rehypeWikipediaFootnotes } from './rehype-wikipedia-footnotes.js';
+export { rehypeAbsoluteLinks } from './rehype-absolute-links.ts';
+export { rehypeNamedAnchors } from './rehype-named-anchors.ts';
+export { rehypeParagraphWrapper } from './rehype-paragraph-wrapper.ts';
+export { rehypeSanitizeHtml } from './rehype-sanitize-html.ts';
+export { rehypeWikipediaFootnotes } from './rehype-wikipedia-footnotes.ts';

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aiquants/html-to-markdown",
-  "version": "0.1.7",
+  "version": "0.2.0",
   "description": "HTML to Markdown converter",
   "type": "module",
   "main": "./dist/index.js",
@@ -29,7 +29,12 @@
     "debug": "NODE_OPTIONS='--enable-source-maps' vite-node --inspect-brk=9229 main.ts --",
     "start": "vite-node main.ts",
     "build": "vite build",
-    "prepublishOnly": "npm run build",
+    "typecheck": "tsc --noEmit",
+    "clean": "rm -rf dist",
+    "prepublishOnly": "npm run clean && npm run typecheck && npm run build",
+    "publish:patch": "npm version patch && npm publish",
+    "publish:minor": "npm version minor && npm publish",
+    "publish:major": "npm version major && npm publish",
     "lint": "biome lint app/",
     "format": "biome format --write",
     "check": "biome check --fix src/",

package/dist/mcp-server-D4Mj1IQ9.cjs

@@ -1 +0,0 @@
-const t=new(require("happy-dom").Window);Object.assign(global,{document:t.document,window:t,self:t});const r=require("./version-DTaUKNBS.cjs"),e=require("node:process"),o=require("./core-D2dlEt1l.cjs");class n{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 e=this._buffer.toString("utf8",0,t).replace(/\r$/,"");return this._buffer=this._buffer.subarray(t+1),function(t){return r.JSONRPCMessageSchema.parse(JSON.parse(t))}(e)}clear(){this._buffer=void 0}}class s{constructor(t=e.stdin,r=e.stdout){this._stdin=t,this._stdout=r,this._readBuffer=new n,this._started=!1,this._ondata=t=>{this._readBuffer.append(t),this.processReadBuffer()},this._onerror=t=>{var r;null===(r=this.onerror)||void 0===r||r.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,r;;)try{const r=this._readBuffer.readMessage();if(null===r)break;null===(t=this.onmessage)||void 0===t||t.call(this,r)}catch(e){null===(r=this.onerror)||void 0===r||r.call(this,e)}}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(r=>{const e=function(t){return JSON.stringify(t)+"\n"}(t);this._stdout.write(e)?r():this._stdout.once("drain",r)})}}class a{constructor(){this.version=r.getPackageVersion(),this.server=new r.Server({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."},{capabilities:{tools:{}}}),this.setupToolHandlers()}setupToolHandlers(){this.server.setRequestHandler(r.ListToolsRequestSchema,async()=>({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.",inputSchema:{type:"object",properties:{url:{type:"string",description:"The URL of the web page to convert to Markdown"},html_content:{type:"string",description:"HTML content as a string to convert to Markdown (alternative to URL)"},locale:{type:"string",description:"The locale to use for browser context (en-US or ja-JP)",enum:["en-US","ja-JP"],default:"en-US"},output_format:{type:"string",description:"Output format: 'markdown' (default), 'html', or 'both'",enum:["markdown","html","both"],default:"markdown"}},anyOf:[{required:["url"]},{required:["html_content"]}]}}]})),this.server.setRequestHandler(r.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 r=function(t){if(!t||"object"!=typeof t)throw new Error("Arguments are required");const{url:r,html_content:e,locale:o="en-US",output_format:n="markdown"}=t;if(!r&&!e)throw new Error("Either 'url' or 'html_content' is required");if(r&&"string"!=typeof r)throw new Error("'url' must be a string");if(e&&"string"!=typeof e)throw new Error("'html_content' must be a string");if(o&&!["en-US","ja-JP"].includes(o))throw new Error("'locale' must be 'en-US' or 'ja-JP'");if(n&&!["markdown","html","both"].includes(n))throw new Error("'output_format' must be 'markdown', 'html', or 'both'");return{url:r,html_content:e,locale:o,output_format:n}}(t),{url:e,html_content:n,locale:s="en-US",output_format:a="markdown"}=r;if(e)try{new URL(e)}catch{throw new Error(`Invalid URL: ${e}`)}const i=await o.htmlToMarkdown(e||"direct-html-input",{locale:s,htmlContent:n});let c;switch(a){case"markdown":c=[{type:"text",text:i.markdown}];break;case"html":c=[{type:"text",text:i.html}];break;case"both":c=[{type:"text",text:i.markdown},{type:"text",text:i.html}];break;default:throw new Error(`Invalid output format: ${a}`)}return{content:c,isError:!1}}catch(r){return{content:[{type:"text",text:`Error: ${r instanceof Error?r.message:String(r)}`}],isError:!0}}}async start(){const t=new s;await this.server.connect(t)}}const i=()=>new a;exports.createMcpServer=i,exports.runMcpServer=async()=>{const t=i();await t.start()};