converting HTML to Markdown - Emojis

[markus-96]

Hi,

When I have a page in HTML with embedded emojis and I convert them into Markdown, the emojis are always embedded as a picture, filling in the whole width of the page, what breaks the entire page.

For example:

HTML:

 <p><img src="/_assets/svg/twemoji/1f34e.svg" alt="🍎" draggable="false" class="emoji"></p>

gets converted to:

![🍎](/_assets/svg/twemoji/1f34e.svg)

instead of:

:apple:

Maybe I could implement a turndown rule to compensate this?

Something like (completely untested):

---a/server/models/pages.js
+++b/server/models/pages.js
@@ -12,6 +12,7 @@
 const TurndownService = require('turndown')
 const turndownPluginGfm = require('@joplin/turndown-plugin-gfm').gfm
 const cheerio = require('cheerio')
+const emojiData = require('markdown-it-emoji/lib/data/full.js')
 
 /* global WIKI */
 
@@ -621,6 +622,25 @@
           replacement: c => ''
         })
 
+        const emojiToMarkdown = {};
+        for (const [shortcode, emoji] of Object.entries(emojiData)) {
+          emojiToMarkdown[emoji.codePointAt(0).toString(16)] = shortcode;
+        }
+
+        td.addRule('keepEmojis', {
+          filter: (n, o) => {
+            return n.nodeName === 'IMG' && n.hasAttribute('src') && n.getAttribute('src').startsWith('/_assets/svg/twemoji/');
+          },
+          replacement: (c, n) => {
+            const src = n.getAttribute('src');
+            const utf8Code = src.match(/\/([0-9a-f]+)\.svg$/i)[1];
+
+            const markdownCode = emojiToMarkdown[utf8Code.toLowerCase()];
+
+            return markdownCode ? `:${markdownCode}:` : c;
+          }
+        });
+
         convertedContent = td.turndown(ogPage.content)
       // -> Unsupported
       } else {

If you like the idea, I would start making a PR for it.