parse mastodon urls in extra fields

app/activitypub/util.py

@@ -33,7 +33,8 @@ from app.utils import get_request, allowlist_html, get_setting, ap_datetime, mar
     microblog_content_to_title, is_video_url, \
     notification_subscribers, communities_banned_from, actor_contains_blocked_words, \
     html_to_text, add_to_modlog_activitypub, joined_communities, \
-    moderating_communities, get_task_session, is_video_hosting_site, opengraph_parse, instance_banned
+    moderating_communities, get_task_session, is_video_hosting_site, opengraph_parse, instance_banned, \
+    mastodon_extra_field_link
 
 from sqlalchemy import or_
 
@@ -517,6 +518,8 @@ def refresh_user_profile_task(user_id):
                 user.extra_fields = []
                 for field_data in activity_json['attachment']:
                     if field_data['type'] == 'PropertyValue':
+                        if '<a ' in field_data['value']:
+                            field_data['value'] = mastodon_extra_field_link(field_data['value'])
                         user.extra_fields.append(UserExtraField(label=field_data['name'].strip(), text=field_data['value'].strip()))
             if 'type' in activity_json:
                 user.bot = True if activity_json['type'] == 'Service' else False
@@ -769,6 +772,8 @@ def actor_json_to_model(activity_json, address, server):
             user.extra_fields = []
             for field_data in activity_json['attachment']:
                 if field_data['type'] == 'PropertyValue':
+                    if '<a ' in field_data['value']:
+                        field_data['value'] = mastodon_extra_field_link(field_data['value'])
                     user.extra_fields.append(UserExtraField(label=field_data['name'].strip(), text=field_data['value'].strip()))
         try:
             db.session.add(user)

app/main/routes.py

@@ -21,7 +21,7 @@ from app.utils import render_template, get_setting, request_etag_matches, return
     ap_datetime, shorten_string, markdown_to_text, user_filters_home, \
     joined_communities, moderating_communities, markdown_to_html, allowlist_html, \
     blocked_instances, communities_banned_from, topic_tree, recently_upvoted_posts, recently_downvoted_posts, \
-    blocked_users, menu_topics, blocked_communities, get_request
+    blocked_users, menu_topics, blocked_communities, get_request, mastodon_extra_field_link
 from app.models import Community, CommunityMember, Post, Site, User, utcnow, Topic, Instance, \
     Notification, Language, community_language, ModLog, read_posts
 
@@ -509,6 +509,8 @@ def replay_inbox():
 @bp.route('/test')
 def test():
 
+    return mastodon_extra_field_link('<a href="https://www.brammeehan.com" target="_blank" rel="nofollow noopener me" translate="no"><span class="invisible">https://www.</span><span class="">brammeehan.com</span><span class="invisible"></span></a>')
+
     response = get_request('https://rimu.geek.nz')
     x = ''
     if response.status_code == 200:

app/models.py

@@ -2067,7 +2067,7 @@ class UserExtraField(db.Model):
     id = db.Column(db.Integer, primary_key=True)
     user_id = db.Column(db.Integer, db.ForeignKey('user.id'), index=True)
     label = db.Column(db.String(50))
-    text = db.Column(db.String(256))
+    text = db.Column(db.String(1024))
 
 
 class UserBlock(db.Model):

app/utils.py

@@ -384,6 +384,12 @@ def html_to_text(html) -> str:
     return soup.get_text()
 
 
+def mastodon_extra_field_link(extra_field: str) -> str:
+    soup = BeautifulSoup(extra_field, 'html.parser')
+    for tag in soup.find_all('a'):
+        return tag['href']
+
+
 def microblog_content_to_title(html: str) -> str:
     title = ''
     if '<p>' in html:

migrations/versions/9df13396fd54_extra_field_length.py

@@ -0,0 +1,38 @@
+"""extra field length
+
+Revision ID: 9df13396fd54
+Revises: 8758c0a94f82
+Create Date: 2025-01-01 15:30:42.165900
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = '9df13396fd54'
+down_revision = '8758c0a94f82'
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table('user_extra_field', schema=None) as batch_op:
+        batch_op.alter_column('text',
+               existing_type=sa.VARCHAR(length=256),
+               type_=sa.String(length=1024),
+               existing_nullable=True)
+
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table('user_extra_field', schema=None) as batch_op:
+        batch_op.alter_column('text',
+               existing_type=sa.String(length=1024),
+               type_=sa.VARCHAR(length=256),
+               existing_nullable=True)
+
+    # ### end Alembic commands ###