Double-escape HTML when ansi is used

When using ansi_up for ansi conversion as documented in the README file, the HTML output is escaped twice. The cell contains:

...
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'NoneType' object has no attribute 'call'",
     "output_type": "error",
     "traceback": [
      "\u001b[31m---------------------------------------------------------------------------\u001b[39m",
      "\u001b[31mAttributeError\u001b[39m                            Traceback (most recent call last)",
      "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[3]\u001b[39m\u001b[32m, line 1\u001b[39m\n\u001b[32m----> \u001b[39m\u001b[32m1\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m.call()\n",
      "\u001b[31mAttributeError\u001b[39m: 'NoneType' object has no attribute 'call'"
     ]
    }
   ],
   "source": [
    "None.call()"
   ]
  },
...

In particular note the use of ----->, which is created by the Python traceback. When I set nb.ansi = (code) => ansiUp.ansi_to_html(code), the resulting output has the > character escaped twice, once from notebookjs itself, and once from the nb.ansi() call (which also escapes HTML text). This seems to come from this code:
https://github.com/jsvine/notebookjs/blob/fb05a45a80f30951d7e03b70bb01d9260a0223a4/notebook.js#L194

One possible solution would be to replace the default for nb.ansi from ident to escapeHTML, then remove all calls to escapeHTML inside calls to nb.ansi. I'm happy to submit a PR if there is hope this would be reviewed/merged/released.

Start at notebook.js line 194 and compare the default nb.ansi behavior with the README's ansi_up usage. Trace the escaping applied by notebookjs and nb.ansi, then confirm that traceback characters such as > are escaped only once while other HTML output remains safe.