fix: improve docgen

tools/docgen/README.md

@@ -31,6 +31,15 @@ channels:
           ms.author: authors' Microsoft alias
 ```
 
+## Run the tool
+
+```bash
+cd tools/docgen
+pip install -e .
+
+python -m docgen --manifest docgen-manifest.yaml
+```
+
 ## Modify input file
 
 ### Image alt text

tools/docgen/docgen/__main__.py

@@ -7,6 +7,8 @@
 def instantiate_channel(channel_yml):
     name = channel_yml["name"]
     module_name, class_name = name.rsplit(".", 1)
+
+    module_name = f"{module_name}"
     print(f"Instantiating {class_name} from module {module_name}")
     clazz = getattr(importlib.import_module(module_name), class_name)
     channel_yml.pop("name")

tools/docgen/docgen/channels.py

@@ -258,7 +258,7 @@ def convert_soup_to_md(soup, **options):
                     "markdown.extensions.fenced_code",
                 ],
             )
-            parsed_html = BeautifulSoup(parsed_html)
+            parsed_html = BeautifulSoup(parsed_html, features="html.parser")
             parsed_html = self._download_and_replace_images(
                 parsed_html,
                 None,

tools/docgen/setup.py

@@ -34,6 +34,7 @@
         "argparse",
         "pypandoc",
         "markdownify",
+        "markdown",
         "traitlets",
     ],
 )