Allow bookmarks to have empty title and description (#843)

* add migration for merging fields

* remove usage of website title and description

* keep empty website title and description in API for compatibility

* restore scraping in API and add option for disabling it

* document API scraping behavior

* remove deprecated fields from API docs

* improve form layout

* cleanup migration

* cleanup website loader

* update tests

bookmarks/api/routes.py

@@ -56,7 +56,12 @@ class BookmarkViewSet(
         return Bookmark.objects.all().filter(owner=user)
 
     def get_serializer_context(self):
-        return {"request": self.request, "user": self.request.user}
+        disable_scraping = "disable_scraping" in self.request.GET
+        return {
+            "request": self.request,
+            "user": self.request.user,
+            "disable_scraping": disable_scraping,
+        }
 
     @action(methods=["get"], detail=False)
     def archived(self, request):
@@ -101,16 +106,7 @@ class BookmarkViewSet(
             self.get_serializer(bookmark).data if bookmark else None
         )
 
-        # Either return metadata from existing bookmark, or scrape from URL
-        if bookmark:
-            metadata = WebsiteMetadata(
-                url,
-                bookmark.website_title,
-                bookmark.website_description,
-                None,
-            )
-        else:
-            metadata = website_loader.load_website_metadata(url)
+        metadata = website_loader.load_website_metadata(url)
 
         # Return tags that would be automatically applied to the bookmark
         profile = request.user.profile
@@ -120,7 +116,7 @@ class BookmarkViewSet(
                 auto_tags = auto_tagging.get_tags(profile.auto_tagging_rules, url)
             except Exception as e:
                 logger.error(
-                    f"Failed to auto-tag bookmark. url={bookmark.url}",
+                    f"Failed to auto-tag bookmark. url={url}",
                     exc_info=e,
                 )
 

bookmarks/api/serializers.py

@@ -4,7 +4,11 @@ from rest_framework import serializers
 from rest_framework.serializers import ListSerializer
 
 from bookmarks.models import Bookmark, Tag, build_tag_string, UserProfile
-from bookmarks.services.bookmarks import create_bookmark, update_bookmark
+from bookmarks.services.bookmarks import (
+    create_bookmark,
+    update_bookmark,
+    enhance_with_website_metadata,
+)
 from bookmarks.services.tags import get_or_create_tag
 
 
@@ -29,8 +33,6 @@ class BookmarkSerializer(serializers.ModelSerializer):
             "title",
             "description",
             "notes",
-            "website_title",
-            "website_description",
             "web_archive_snapshot_url",
             "favicon_url",
             "preview_image_url",
@@ -40,15 +42,17 @@ class BookmarkSerializer(serializers.ModelSerializer):
             "tag_names",
             "date_added",
             "date_modified",
-        ]
-        read_only_fields = [
             "website_title",
             "website_description",
+        ]
+        read_only_fields = [
             "web_archive_snapshot_url",
             "favicon_url",
             "preview_image_url",
             "date_added",
             "date_modified",
+            "website_title",
+            "website_description",
         ]
         list_serializer_class = BookmarkListSerializer
 
@@ -63,6 +67,9 @@ class BookmarkSerializer(serializers.ModelSerializer):
     tag_names = TagListField(required=False, default=[])
     favicon_url = serializers.SerializerMethodField()
     preview_image_url = serializers.SerializerMethodField()
+    # Add dummy website title and description fields for backwards compatibility but keep them empty
+    website_title = serializers.SerializerMethodField()
+    website_description = serializers.SerializerMethodField()
 
     def get_favicon_url(self, obj: Bookmark):
         if not obj.favicon_file:
@@ -80,6 +87,12 @@ class BookmarkSerializer(serializers.ModelSerializer):
         preview_image_url = request.build_absolute_uri(preview_image_file_path)
         return preview_image_url
 
+    def get_website_title(self, obj: Bookmark):
+        return None
+
+    def get_website_description(self, obj: Bookmark):
+        return None
+
     def create(self, validated_data):
         bookmark = Bookmark()
         bookmark.url = validated_data["url"]
@@ -90,7 +103,14 @@ class BookmarkSerializer(serializers.ModelSerializer):
         bookmark.unread = validated_data["unread"]
         bookmark.shared = validated_data["shared"]
         tag_string = build_tag_string(validated_data["tag_names"])
-        return create_bookmark(bookmark, tag_string, self.context["user"])
+
+        saved_bookmark = create_bookmark(bookmark, tag_string, self.context["user"])
+        # Unless scraping is explicitly disabled, enhance bookmark with website
+        # metadata to preserve backwards compatibility with clients that expect
+        # title and description to be populated automatically when left empty
+        if not self.context.get("disable_scraping", False):
+            enhance_with_website_metadata(saved_bookmark)
+        return saved_bookmark
 
     def update(self, instance: Bookmark, validated_data):
         # Update fields if they were provided in the payload