Make relative URLs in e-* properties absolute (#201)

microformats · Nov 30, 2023 · a0edfdb · a0edfdb
1 parent ab580f6
commit a0edfdb
Show file tree

Hide file tree

Showing 4 changed files with 28 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,7 @@
 All notable changes to this project will be documented in this file.
 
 ## 2.0 - unreleased
+- make relative URLs in e-* properties absolute (#201)
 
 ## 1.1.3 - 2023-06-28
 - reduce instances where photo is implied (#135)

diff --git a/mf2py/parse_property.py b/mf2py/parse_property.py
@@ -96,6 +96,10 @@ def datetime(el, default_date=None):
 
 def embedded(el, base_url=""):
     """Process e-* properties"""
+    for tag in el.find_all():
+        for attr in ("href", "src", "cite", "data", "poster"):
+            if attr in tag.attrs:
+                tag.attrs[attr] = try_urljoin(base_url, tag.attrs[attr])
     return {
         "html": el.decode_contents().strip(),  # secret bs4 method to get innerHTML
         "value": get_textContent(el, replace_img=True, base_url=base_url),

diff --git a/test/examples/relative_url_in_e.html b/test/examples/relative_url_in_e.html
@@ -0,0 +1,13 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta http-equiv="content-type" content="text/html; charset=utf-8">
+  <title>Relative URLs in e-content</title>
+  <base href="http://example.com/" />
+</head>
+<body>
+<div class="h-entry">
+<div class="e-content"><p><a href=/cat.html>Cat <img src=cat.jpg></a></p></div>
+</div>
+</body>
+</html>
diff --git a/test/test_parser.py b/test/test_parser.py
@@ -389,6 +389,16 @@ def test_complex_e_content():
     } == result["items"][0]
 
 
+def test_relative_url_in_e():
+    """When parsing e-* properties, make relative URLs absolute."""
+    result = parse_fixture("relative_url_in_e.html")
+
+    assert (
+        '<p><a href="http://example.com/cat.html">Cat '
+        '<img src="http://example.com/cat.jpg"/></a></p>'
+    ) == result["items"][0]["properties"]["content"][0]["html"]
+
+
 def test_nested_values():
     """When parsing nested microformats, check that value is the value of
     the simple property element"""