michael/pyfedi
1
0
Fork 0
mirror of https://codeberg.org/rimu/pyfedi synced 2025-01-23 19:36:56 -08:00
Code Issues Projects Releases Packages Wiki Activity Actions

Exclude already-discovered communities from admin-preload function

Browse source 
Also - pls see note about using the admin's alt profile for this
This commit is contained in:
freamon 2024-10-31 21:29:17 +00:00
parent 830ef632d9
commit 2e9720539c
2 changed files with 81 additions and 74 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

117
app/admin/routes.py
View file

@ -210,65 +210,60 @@ def admin_federation():
community_json = resp.json()
resp.close()
# sort out the nsfw communities
safe_for_work_communities = []
for community in community_json:
if community['nsfw']:
continue
else:
safe_for_work_communities.append(community)
# sort out any that have less than 100 posts
communities_with_lots_of_content = []
for community in safe_for_work_communities:
if community['counts']['posts'] < 100:
continue
else:
communities_with_lots_of_content.append(community)
# sort out any that do not have greater than 500 active users over the past week
communities_with_lots_of_activity = []
for community in communities_with_lots_of_content:
if community['counts']['users_active_week'] < 500:
continue
else:
communities_with_lots_of_activity.append(community)
# sort out any instances we have already banned
banned_instances = BannedInstances.query.all()
banned_urls = []
communities_not_banned = []
for bi in banned_instances:
banned_urls.append(bi.domain)
for community in communities_with_lots_of_activity:
if community['baseurl'] in banned_urls:
continue
else:
communities_not_banned.append(community)
# sort out the 'seven things you can't say on tv' names (cursewords, ie sh*t), plus some
# "low effort" communities
# I dont know why, but some of them slip through on the first pass, so I just
# ran the list again and filter out more
#
# TO-DO: fix the need for the double filter
already_known = list(db.session.execute(text('SELECT ap_public_url FROM "community"')).scalars())
banned_urls = list(db.session.execute(text('SELECT domain FROM "banned_instances"')).scalars())
seven_things_plus = [
'shit', 'piss', 'fuck',
'cunt', 'cocksucker', 'motherfucker', 'tits',
'memes', 'piracy', '196', 'greentext', 'usauthoritarianism',
'enoughmuskspam', 'political_weirdos', '4chan'
]
for community in communities_not_banned:
for word in seven_things_plus:
if word in community['name']:
communities_not_banned.remove(community)
for community in communities_not_banned:
for word in seven_things_plus:
if word in community['name']:
communities_not_banned.remove(community)
total_count = already_known_count = nsfw_count = low_content_count = low_active_users_count = banned_count = bad_words_count = 0
candidate_communities = []
for community in community_json:
total_count += 1
# sort out already known communities
if community['url'] in already_known:
already_known_count += 1
continue
# sort out the nsfw communities
elif community['nsfw']:
nsfw_count += 1
continue
# sort out any that have less than 100 posts
elif community['counts']['posts'] < 100:
low_content_count += 1
continue
# sort out any that do not have greater than 500 active users over the past week
elif community['counts']['users_active_week'] < 500:
low_active_users_count += 1
continue
# sort out any instances we have already banned
elif community['baseurl'] in banned_urls:
banned_count += 1
continue
# sort out the 'seven things you can't say on tv' names (cursewords), plus some
# "low effort" communities
if any(badword in community['name'].lower() for badword in seven_things_plus):
bad_words_count += 1
continue
else:
candidate_communities.append(community)
filtered_count = already_known_count + nsfw_count + low_content_count + low_active_users_count + banned_count + bad_words_count
flash(_('%d out of %d communities were excluded using current filters' % (filtered_count, total_count)))
# sort the list based on the users_active_week key
parsed_communities_sorted = sorted(communities_not_banned, key=lambda c: c['counts']['users_active_week'], reverse=True)
parsed_communities_sorted = sorted(candidate_communities, key=lambda c: c['counts']['users_active_week'], reverse=True)
# get the community urls to join
community_urls_to_join = []
@ -279,25 +274,37 @@ def admin_federation():
# make the list of urls
for i in range(communities_to_add):
community_urls_to_join.append(parsed_communities_sorted[i]['url'])
community_urls_to_join.append(parsed_communities_sorted[i]['url'].lower())
# loop through the list and send off the follow requests
# use User #1, the first instance admin
# NOTE: Subscribing using the admin's alt_profile causes problems:
# 1. 'Leave' will use the main user name so unsubscribe won't succeed.
# 2. De-selecting and re-selecting 'vote privately' generates a new alt_user_name every time,
# so the username needed for a successful unsubscribe might get lost
# 3. If the admin doesn't have 'vote privately' selected, the federation JSON will populate
# with a blank space for the name, so the subscription won't succeed.
# 4. Membership is based on user id, so using the alt_profile doesn't decrease the admin's joined communities
#
# Therefore, 'main_user_name=False' has been changed to 'admin_preload=True' below
user = User.query.get(1)
pre_load_messages = []
for community in community_urls_to_join:
# get the relevant url bits
server, community = extract_domain_and_actor(community)
# find the community
new_community = search_for_community('!' + community + '@' + server)
# subscribe to the community using alt_profile
# subscribe to the community
# capture the messages returned by do_subscibe
# and show to user if instance is in debug mode
if current_app.debug:
message = do_subscribe(new_community.ap_id, user.id, main_user_name=False)
message = do_subscribe(new_community.ap_id, user.id, admin_preload=True)
pre_load_messages.append(message)
else:
message_we_wont_do_anything_with = do_subscribe.delay(new_community.ap_id, user.id, main_user_name=False)
message_we_wont_do_anything_with = do_subscribe.delay(new_community.ap_id, user.id, admin_preload=True)
if current_app.debug:
flash(_('Results: %(results)s', results=str(pre_load_messages)))

22
app/community/routes.py
View file

@ -409,7 +409,7 @@ def subscribe(actor):
# this is separated out from the subscribe route so it can be used by the
# admin.admin_federation.preload_form as well
@celery.task
def do_subscribe(actor, user_id, main_user_name=True):
def do_subscribe(actor, user_id, admin_preload=False):
remote = False
actor = actor.strip()
user = User.query.get(user_id)
@ -423,14 +423,14 @@ def do_subscribe(actor, user_id, main_user_name=True):
if community is not None:
pre_load_message['community'] = community.ap_id
if community.id in communities_banned_from(user.id):
if main_user_name:
if not admin_preload:
abort(401)
else:
pre_load_message['user_banned'] = True
if community_membership(user, community) != SUBSCRIPTION_MEMBER and community_membership(user, community) != SUBSCRIPTION_PENDING:
banned = CommunityBan.query.filter_by(user_id=user.id, community_id=community.id).first()
if banned:
if main_user_name:
if not admin_preload:
flash(_('You cannot join this community'))
else:
pre_load_message['community_banned_by_local_instance'] = True
@ -442,24 +442,24 @@ def do_subscribe(actor, user_id, main_user_name=True):
db.session.commit()
if community.instance.online():
follow = {
"actor": user.public_url(main_user_name=main_user_name),
"actor": user.public_url(),
"to": [community.public_url()],
"object": community.public_url(),
"type": "Follow",
"id": f"https://{current_app.config['SERVER_NAME']}/activities/follow/{join_request.id}"
}
success = post_request(community.ap_inbox_url, follow, user.private_key,
user.public_url(main_user_name=main_user_name) + '#main-key', timeout=10)
user.public_url() + '#main-key', timeout=10)
if success is False or isinstance(success, str):
if 'is not in allowlist' in success:
msg_to_user = f'{community.instance.domain} does not allow us to join their communities.'
if main_user_name:
if not admin_preload:
flash(_(msg_to_user), 'error')
else:
pre_load_message['status'] = msg_to_user
else:
msg_to_user = "There was a problem while trying to communicate with remote server. If other people have already joined this community it won't matter."
if main_user_name:
if not admin_preload:
flash(_(msg_to_user), 'error')
else:
pre_load_message['status'] = msg_to_user
@ -469,20 +469,20 @@ def do_subscribe(actor, user_id, main_user_name=True):
db.session.add(member)
db.session.commit()
if success is True:
if main_user_name:
if not admin_preload:
flash('You joined ' + community.title)
else:
pre_load_message['status'] = 'joined'
else:
if not main_user_name:
if admin_preload:
pre_load_message['status'] = 'already subscribed, or subsciption pending'
cache.delete_memoized(community_membership, user, community)
cache.delete_memoized(joined_communities, user.id)
if not main_user_name:
if admin_preload:
return pre_load_message
else:
if main_user_name:
if not admin_preload:
abort(404)
else:
pre_load_message['community'] = actor