Better handle file vs urls

torchbox · Jan 21, 2022 · 240514a · 240514a
1 parent a8c2ab4
commit 240514a
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 6 deletions.
diff --git a/wagtail_wordpress_import/block_builder_defaults.py b/wagtail_wordpress_import/block_builder_defaults.py
@@ -329,12 +329,17 @@ def document_exists(name):
 def get_absolute_src(src, domain_prefix=None):
     src = src.lstrip("/")
     url = urlparse(src)
+    known_file_extensions = ["jpg", "jpeg", "gif", "png", "webp"]
+
+    if not url.scheme and domain_prefix:
+        first_url_part = url.path.split("/")[0]
+        if (
+            "." in first_url_part
+            and first_url_part.split(".")[1].lower() not in known_file_extensions
+        ):
+            return "{}://{}".format(urlparse(domain_prefix).scheme, url.path)
+        return "{}/{}".format(domain_prefix, src)
 
-    if not url.netloc and domain_prefix:
-        domain_prefix_scheme = urlparse(domain_prefix).scheme
-        return "{}://{}".format(domain_prefix_scheme, url.path)
-    elif not url.scheme and domain_prefix:
-        return domain_prefix + "/" + src
     return src
 
 

diff --git a/wagtail_wordpress_import/test/tests/test_block_builder.py b/wagtail_wordpress_import/test/tests/test_block_builder.py
@@ -338,7 +338,7 @@ def test_get_absolute_src_without_base_url(self):
     def test_get_abolute_src_slashes_at_start(self):
         self.assertEqual(
             get_absolute_src("//folder/fakeimage.jpg", "http://www.example.com"),
-            "http://example.com/folder/fakeimage.jpg",
+            "http://www.example.com/folder/fakeimage.jpg",
         )
 
     def test_get_abolute_src_slashes_at_start_external_domain(self):