convert relative links to .md files to .html

this uses a markdown.treeprocessor as extension to get all html links
from the resulting html document

sg/sg.py

@@ -13,8 +13,11 @@ import os
 import shutil
 import logging
 from datetime import datetime
+from urllib.parse import urlsplit, urlunsplit
 
-import markdown
+from markdown import Markdown
+from markdown.extensions import Extension
+from markdown.treeprocessors import Treeprocessor
 from jinja2 import Environment, ChoiceLoader, FileSystemLoader, PackageLoader
 import feedgenerator
 
@@ -96,8 +99,11 @@ def convert_to_html(convertibles):
             ])
     )
 
-    md = markdown.Markdown(
-            extensions=['meta', 'fenced_code', 'codehilite'],
+    md = Markdown(
+        extensions=[
+            'meta', 'fenced_code', 'codehilite',
+            MarkdownLinkExtension()
+        ],
         output_format='html5',
     )
 
@@ -172,5 +178,38 @@ def convert_to_html(convertibles):
 
     # generate tags
 
+
+class MarkdownLinkTreeprocessor(Treeprocessor):
+    """Converts relative links to .md files to .html
+
+    """
+
+    def run(self, root):
+        for element in root.iter():
+            if element.tag == 'a':
+                url = element.get('href')
+                converted = self.convert(url)
+                element.set('href', converted)
+        return root
+
+    def convert(self, url):
+        scheme, netloc, path, query, fragment = urlsplit(url)
+        logger.debug(f'{url} -> scheme: {scheme} netloc: {netloc} path: {path} query: {query} fragment: {fragment}')
+        if (scheme or netloc or not path):
+            return url
+        if path.endswith('.md'):
+            path = path[:-3] + '.html'
+
+        url = urlunsplit((scheme, netloc, path, query, fragment))
+        return url
+
+
+class MarkdownLinkExtension(Extension):
+    def extendMarkdown(self, md):
+        md.treeprocessors.register(
+                MarkdownLinkTreeprocessor(md), 'mdlink', 0,
+        )
+
+
 if __name__ == '__main__':
     main()