Convert-to-markdown plugin added

README.md

@@ -2,14 +2,14 @@
 
 This repository comprises plugins which allow extending the functionality of [ONLYOFFICE DocSpace](https://www.onlyoffice.com/docspace.aspx).
 
+* [ZIP Archives](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/archives)
 * [Codemirror](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/codemirror)
-* [ConvertToMarkdown](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/convert-to-markdown)
 * [draw.io](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/draw.io)
 * [Markdown](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/markdown)
 * [ImageEditor](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/image-editor)
 * [PDFConverter](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/pdf-converter)
 * [SpeechToText](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/speech-to-text)
-* [ZIP Archives](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/archives)
+* [ConvertToMarkdown](https://github.com/ONLYOFFICE/docspace-plugins/tree/master/convert-to-markdown)
 
 ## User feedback and support
 

convert-to-markdown/README.md

@@ -1,13 +1,13 @@
 # Markdown Converter - ONLYOFFICE DocSpace Plugin
 
-Convert DOCX and TXT files to Markdown format directly in DocSpace with a single click.
+Convert DOCX, HTML and TXT files to Markdown format directly in DocSpace with a single click.
 
 ## Features
 
-- **Easy Conversion**: Right-click any DOCX or TXT file and select "Convert to Markdown"
+- **Easy Conversion**: Right-click any DOCX, HTML or TXT file and select "Convert to Markdown"
 - **Client-Side Processing**: All conversions happen in your browser - no external services required
 - **Supported Formats**:
-  - `.docx` - Microsoft Word documents
+  - `.docx` - Word documents
   - `.txt` - Plain text files
   - `.html` - HTML files
 - **Fast & Reliable**: Uses industry-standard libraries (mammoth.js for DOCX parsing, Turndown for HTML-to-Markdown conversion)
@@ -16,7 +16,7 @@ Convert DOCX and TXT files to Markdown format directly in DocSpace with a single
 ## How to Use
 
 1. Navigate to a folder in DocSpace
-2. Right-click on a DOCX, TXT, or HTML file
+2. Right-click on a DOCX, HTML or TXT file
 3. Select **"Convert to Markdown"** from the context menu
 4. The converted `.md` file will be created in the same folder
 
@@ -27,7 +27,7 @@ Convert DOCX and TXT files to Markdown format directly in DocSpace with a single
 
 - **DOCX Files**: Parsed using [mammoth.js](https://github.com/mwilliamson/mammoth.js) to extract content as HTML, then converted to Markdown
 - **TXT Files**: Converted directly to Markdown format
-- **HTML Files**: Converted directly to Markdown format
+- **HTML Files**: Converted directly to Markdown format using [Turndown](https://github.com/mixmark-io/turndown)
 
 ### Technologies Used
 
@@ -43,7 +43,7 @@ Convert DOCX and TXT files to Markdown format directly in DocSpace with a single
 ## Troubleshooting
 
 ### Conversion Failed
-- Check that the file is a valid DOCX, TXT or HTML file
+- Check that the file is a valid DOCX, HTML or TXT file
 - Ensure you have permission to create files in the folder
 - Try refreshing the page and attempting again
 

convert-to-markdown/package.json

@@ -3,7 +3,6 @@
   "version": "1.0.0",
   "description": "Convert DOCX, TXT and HTML files to Markdown format directly in DocSpace",
   "license": "Apache-2.0",
-  "homepage": "https://github.com/ONLYOFFICE/docspace-plugins/tree/master/convert-to-markdown",
   "author": "ONLYOFFICE",
   "maintainers": [
     "Ascensio System SIA <integration@onlyoffice.com> (https://www.onlyoffice.com)"
@@ -13,6 +12,7 @@
   },
   "pluginName": "MarkdownConverter",
   "logo": "icon-md.svg",
+  "homepage": "",
   "main": "index.ts",
   "private": true,
   "scopes": [
@@ -28,12 +28,14 @@
     "prettier": "2.8.6",
     "ts-loader": "^9.3.1",
     "typescript": "^4.7.4",
-    "webpack": "^5.105.0",
+    "webpack": "^5.74.0",
     "webpack-cli": "^4.10.0"
   },
   "dependencies": {
     "@onlyoffice/docspace-plugin-sdk": "^2.0.0",
+    "@truto/turndown-plugin-gfm": "^1.0.2",
     "mammoth": "^1.11.0",
-    "turndown": "^7.2.2"
-  }
+    "turndown": "^7.2.2",
+    "turndown-plugin-gfm": "^1.0.2"
+  },
 }

convert-to-markdown/src/ConvertFile.ts

@@ -20,10 +20,38 @@ import {
   ToastType,
 } from "@onlyoffice/docspace-plugin-sdk";
 import TurndownService from "turndown";
+import { tables } from "turndown-plugin-gfm";
 import mammoth from "mammoth";
 
 import plugin from ".";
 
+// Patch mammoths internal Element.prototype.text which throws "Not implemented"
+// when an element has multiple or non text children 
+// Both convertToHtml and extractRawText hit the same reader path, so patching
+// at the source is the only reliable fix for the issue
+const patchMammothNodes = (): void => {
+  try {
+    // eslint-disable-next-line @typescript-eslint/no-var-requires
+    const nodes = require("mammoth/lib/xml/nodes");
+    const collectText = (children: any[]): string =>
+      children
+        .map((child: any) => {
+          if (child.type === "text") return child.value;
+          if (Array.isArray(child.children)) return collectText(child.children);
+          return "";
+        })
+        .join("");
+
+    nodes.Element.prototype.text = function (): string {
+      if (this.children.length === 0) return "";
+      return collectText(this.children);
+    };
+  } catch (_) {
+    // If patching fails mammoth will still work for docs that dont hit the bug
+  }
+};
+patchMammothNodes();
+
 // Supported file extensions
 const SUPPORTED_EXTENSIONS = {
   docx: ".docx",
@@ -37,12 +65,52 @@ class ConvertFile {
   private apiURL = "";
   private createLock = false;
 
-  private turndownService = new TurndownService({
-    headingStyle: "atx",
-    codeBlockStyle: "fenced",
-    emDelimiter: "*",
-    bulletListMarker: "-",
-  });
+  private turndownService = (() => {
+    const service = new TurndownService({
+      headingStyle: "atx",
+      codeBlockStyle: "fenced",
+      emDelimiter: "*",
+      bulletListMarker: "-",
+    });
+    service.use(tables);
+
+    // Handle tables that have no <thead> (all rows in <tbody>)
+    // The GFM tables plugin only produces pipe-tables when it finds <th> / <thead>
+    // so we need a separate rule that treats the first <tr> as the header
+    service.addRule("tableWithoutHeader", {
+      filter: (node: HTMLElement): boolean => {
+        return node.nodeName === "TABLE" && !node.querySelector("thead");
+      },
+      replacement: (_content: string, node: Node): string => {
+        const el = node as HTMLElement;
+        const rows = Array.from(el.querySelectorAll("tr"));
+        if (rows.length === 0) return _content;
+
+        const getCells = (row: Element): string[] =>
+          Array.from(row.querySelectorAll("td, th")).map((cell) =>
+            (cell.textContent || "")
+              .trim()
+              .replace(/\s+/g, " ")
+              .replace(/\|/g, "\\|")
+          );
+
+        const allRows = rows.map(getCells);
+        const header = allRows[0];
+        const separator = header.map(() => "---");
+        const body = allRows.slice(1);
+
+        const fmt = (cells: string[]): string => `| ${cells.join(" | ")} |`;
+
+        return (
+          "\n\n" +
+          [fmt(header), fmt(separator), ...body.map(fmt)].join("\n") +
+          "\n\n"
+        );
+      },
+    });
+
+    return service;
+  })();
 
   private createAPIUrl = (): void => {
     const api = plugin.getAPI();
@@ -123,6 +191,13 @@ class ConvertFile {
     if (!response.ok) {
       throw new Error(`Failed to download file: ${response.statusText}`);
     }
+
+    const contentType = response.headers.get("Content-Type") || "";
+    if (contentType.includes("application/pdf")) {
+      throw new Error(
+        "This file is watermark-protected. Disable the \"Add watermarks to documents\" room setting and try again."
+      );
+    }
 
     const buffer = await response.arrayBuffer();
     if (!buffer || buffer.byteLength === 0) {

convert-to-markdown/src/turndown-plugin-gfm.d.ts

@@ -0,0 +1,8 @@
+declare module "turndown-plugin-gfm" {
+  import TurndownService from "turndown";
+
+  export function gfm(service: TurndownService): void;
+  export function tables(service: TurndownService): void;
+  export function strikethrough(service: TurndownService): void;
+  export function taskListItems(service: TurndownService): void;
+}