From 2afb80fc1ef63027a278569aef494a5f6d6e987f Mon Sep 17 00:00:00 2001
From: Jon Michael Aanes <jonjmaa@gmail.com>
Date: Sat, 1 Jun 2024 20:45:50 +0200
Subject: [PATCH] Avoid parsing twitter intent as username

---
 socials_util/__init__.py | 21 ++++++++++++---------
 test/test_parsing.py     | 14 +++++++++++++-
 2 files changed, 25 insertions(+), 10 deletions(-)

diff --git a/socials_util/__init__.py b/socials_util/__init__.py
index c15b7ac..7488fe9 100644
--- a/socials_util/__init__.py
+++ b/socials_util/__init__.py
@@ -188,12 +188,15 @@ WIKIDATA_PROPERTIES: dict[SocialSiteId | int, WikidataInfo] = {
 
 
 def re_social_subdomain(main_domain: str) -> str:
-    return r'^(?:https?:\/\/)?([\w_-]+)\.' + re.escape(main_domain) + r'(\/.*)?$'
+    return r'^(?:https?:\/\/)?(?:www\.)?([\w_-]+)\.' + re.escape(main_domain) + r'(\/.*)?$'
 
 
-RE_ID = r'@?([^/]+)'
-RE_DUAL_ID = r'@?([^/]+/[^/]+)'
-RE_ANY_SUBPATH = r'(|\/|\/.*)$'
+RE_ID = r'@?([^\s/]+)'
+RE_DUAL_ID = r'@?([^\s/]+/[^\s/]+)'
+RE_ANY_SUBPATH = r'(|\/|\/\S*)$'
+
+SPECIAL_REGEX_LITERALS = frozenset({RE_ID, RE_DUAL_ID, RE_ANY_SUBPATH})
+DOES_NOT_NEED_AUTO_SLASH = frozenset({RE_ANY_SUBPATH})
 
 
 def re_social_path(main_domain: str) -> str:
@@ -212,12 +215,13 @@ def re_social_path_adv(main_domain: str, *path: str) -> str:
     ]
 
     for p in path:
-        if p != RE_ANY_SUBPATH:
+        if p not in DOES_NOT_NEED_AUTO_SLASH:
             regex_builder.append(r'\/')
         regex_builder.append(
-            p if p in {RE_ID, RE_DUAL_ID, RE_ANY_SUBPATH} else re.escape(p),
+            p if p in SPECIAL_REGEX_LITERALS else re.escape(p),
         )
-    if path[-1] != RE_ANY_SUBPATH:
+        del p
+    if path[-1] not in DOES_NOT_NEED_AUTO_SLASH:
         regex_builder.append(r'\/?$')
     return ''.join(regex_builder)
 
@@ -413,8 +417,7 @@ WELL_KNOWN_MASTODON_INSTANCES: frozenset[str] = frozenset(
     },
 )
 
-DISALLOWED_IDENTIFIERS: frozenset[str] = frozenset({'www'})
-
+DISALLOWED_IDENTIFIERS: frozenset[str] = frozenset({'www', 'intent', 'user'})
 
 def determine_social_from_url_internally(
     url: str,
diff --git a/test/test_parsing.py b/test/test_parsing.py
index 6cd3f74..076ec96 100644
--- a/test/test_parsing.py
+++ b/test/test_parsing.py
@@ -1,10 +1,11 @@
 import pytest
 
-from socials_util import *
+from socials_util import determine_social_from_url, SocialSiteId, SocialLink
 
 PARSABLE_SOCIAL_IDS_COMBINED: list[tuple[str, object, str]] = [
     # Tumblr formats
     ('https://triviallytrue.tumblr.com/', SocialSiteId.TUMBLR, 'triviallytrue'),
+    ('https://www.triviallytrue.tumblr.com/', SocialSiteId.TUMBLR, 'triviallytrue'),
     ('https://tumblr.com/triviallytrue', SocialSiteId.TUMBLR, 'triviallytrue'),
     ('https://tumblr.com/blog/triviallytrue', SocialSiteId.TUMBLR, 'triviallytrue'),
     (
@@ -179,6 +180,13 @@ PARSABLE_SOCIAL_IDS_COMBINED: list[tuple[str, object, str]] = [
     ('https://solquiet.deviantart.com/', SocialSiteId.DEVIANT_ART_ACCOUNT, 'solquiet'),
 ]
 
+NOT_PARSABLE = [
+        # Twitter intents are not supported
+        'twitter.com/intent/user?user_id=123',
+        'https://twitter.com/intent/user?user_id=123',
+        'https://twitter.com/intent/user',
+        'https://twitter.com/intent',
+]
 
 @pytest.mark.parametrize(
     'url,expected_social_site_id,expected_social_id', PARSABLE_SOCIAL_IDS_COMBINED
@@ -190,3 +198,7 @@ def test_parse_social_ids(url, expected_social_site_id, expected_social_id):
         expected_social_id,
         expected_social_site_id,
     ), url
+
+@pytest.mark.parametrize('url', NOT_PARSABLE)
+def test_not_parsable(url: str):
+    assert determine_social_from_url(url) is None