From 699cd16ae146dec1f075282b27bf443f55cc94c4 Mon Sep 17 00:00:00 2001
From: duogenesis <136373989+duogenesis@users.noreply.github.com>
Date: Tue, 24 Dec 2024 10:21:42 +1100
Subject: [PATCH] Moderation (#566)

---
 banned-club.sql                        | 2 ++
 service/chat/__init__.py               | 2 +-
 service/chat/offensive/__init__.py     | 9 ++++++++-
 service/chat/offensive/test_init.py    | 3 +++
 test/functionality4/xmpp-rate-limit.sh | 2 +-
 5 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/banned-club.sql b/banned-club.sql
index 7a23b29..6d96c36 100644
--- a/banned-club.sql
+++ b/banned-club.sql
@@ -2514,6 +2514,8 @@ VALUES
     ('shedtwt'),
     ('she''s mine now bitch'),
     ('she took the fucking kids'),
+    ('shit skin'),
+    ('shitskin'),
     ('shooter'),
     ('shoot me'),
     ('shota'),
diff --git a/service/chat/__init__.py b/service/chat/__init__.py
index 1520664..076f021 100644
--- a/service/chat/__init__.py
+++ b/service/chat/__init__.py
@@ -33,7 +33,7 @@
 
 PORT = sys.argv[1] if len(sys.argv) >= 2 else 5443
 
-MAX_INTROS_PER_DAY = 35
+MAX_INTROS_PER_DAY = 50
 
 # TODO: Tables to migrate to monolithic DB:
 #
diff --git a/service/chat/offensive/__init__.py b/service/chat/offensive/__init__.py
index a29933d..4d13f81 100644
--- a/service/chat/offensive/__init__.py
+++ b/service/chat/offensive/__init__.py
@@ -21,6 +21,7 @@
     "c+v+m+": "cum",
     "c0ck": "cock",
     "c0cksucker": "cocksucker",
+    "cok": "cock",
     "cvmming": "cumming",
     "cvms": "cums",
     "cvmshot": "cumshot",
@@ -43,11 +44,13 @@
     "p0rn": "porn",
     "pissin": "pissing",
     "r": "are",
+    "sxy": "sexy",
     "u": "you",
     "ur": "your",
     "urself": "yourself",
     "wh0re": "whore",
     "wh0res": "whores",
+    "a[s5$]{2}": "ass",
 }
 
 
@@ -204,6 +207,7 @@
     "foot job",
     "footjob",
     "fuck you",
+    "fuck your ass",
     "fuck your asshole",
     "fuck your brains out",
     "fuck your face",
@@ -244,6 +248,7 @@
     "hit it raw",
     "hoe",
     "hoes",
+    "homo",
     "horniest",
     "horny",
     "hotsex",
@@ -376,8 +381,10 @@
     "shag",
     "shagging",
     "shemale",
+    "shit skin",
     "shitdick",
     "shitfuck",
+    "shitskin",
     "shota",
     "shotacon",
     "skank",
@@ -516,7 +523,7 @@ def _normalize_spelling(haystack: str):
     for needle, replacement in _normalization_map.items():
         # Apparently compiled regexes are cached between invocations of
         # re.compile.
-        pattern = re.compile(f"\\b{needle}\\b", re.IGNORECASE)
+        pattern = re.compile(f"(?:(?<=^)|(?<=\s)){needle}(?=\s|$)", re.IGNORECASE)
 
         haystack = pattern.sub(replacement, haystack)
 
diff --git a/service/chat/offensive/test_init.py b/service/chat/offensive/test_init.py
index 19c09ca..c1c60ee 100644
--- a/service/chat/offensive/test_init.py
+++ b/service/chat/offensive/test_init.py
@@ -4,6 +4,9 @@
 class TestIsOffensive(unittest.TestCase):
 
     def test_offensive_strings(self):
+        self.assertTrue(
+                is_offensive("fuk your as$"))
+
         self.assertTrue(
                 is_offensive("go fk urself please"))
 
diff --git a/test/functionality4/xmpp-rate-limit.sh b/test/functionality4/xmpp-rate-limit.sh
index 31d3c20..4c74d41 100755
--- a/test/functionality4/xmpp-rate-limit.sh
+++ b/test/functionality4/xmpp-rate-limit.sh
@@ -6,7 +6,7 @@ cd "$script_dir"
 source ../util/setup.sh
 mapfile -t greetings < ../fixtures/greetings.txt
 
-max_intros_per_day=35
+max_intros_per_day=50
 
 set -xe