indymedia
/
epicyon
mirror of https://gitlab.com/bashrc2/epicyon
7
4
Fork 1
Code Issues 14 Releases Wiki Activity

Check that text line resembles url

main
bashrc 2026-07-19 11:15:00 +01:00
parent 03ff9496de
commit 3e3ffc51c1
1 changed files with 11 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
src/collections.py
View File

@ -20,6 +20,7 @@ from src.utils import get_nickname_from_actor
from src.utils import get_domain_from_actor from src.utils import get_domain_from_actor
from src.utils import url_text_to_number from src.utils import url_text_to_number
from src.utils import save_json from src.utils import save_json
from src.utils import resembles_url
from src.data import is_a_file from src.data import is_a_file
from src.data import load_list from src.data import load_list
from src.data import is_a_dir from src.data import is_a_dir
@ -249,18 +250,19 @@ def get_featured_collections_feed(base_dir: str,
continue continue
if not collection_name: if not collection_name:
continue continue
# add a collection item # add a collection item which is typically an actor or hashtag url
item_url = '' item_url = ''
text = remove_html(text) text = remove_html(text)
if '/tags/' in text: if resembles_url(text):
# hashtag url if '/tags/' in text:
item_url = text # hashtag url
else:
# check that this is an actor url
item_nickname = get_nickname_from_actor(text)
item_domain, _ = get_domain_from_actor(text)
if item_nickname and item_domain:
item_url = text item_url = text
else:
# check that this is an actor url
item_nickname = get_nickname_from_actor(text)
item_domain, _ = get_domain_from_actor(text)
if item_nickname and item_domain:
item_url = text
if item_url: if item_url:
if authorizations.get(item_url): if authorizations.get(item_url):
collection_items.append(authorizations[item_url]) collection_items.append(authorizations[item_url])