From c149e8d3cafba9d0ce350dcf1dd2ba672bd88c0b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?J=C3=A9r=C3=B4me=20Deuchnord?= <jerome@deuchnord.fr>
Date: Sun, 7 Nov 2021 17:36:58 +0100
Subject: [PATCH] Add link moderator

---
 Pipfile                 |  1 +
 Pipfile.lock            | 41 +++++++++++++++++++++++++++++++-
 _twitchbot/config.py    | 10 ++++++++
 _twitchbot/moderator.py | 52 +++++++++++++++++++++++++++++++++++++++++
 4 files changed, 103 insertions(+), 1 deletion(-)

diff --git a/Pipfile b/Pipfile
index 7daa0cf..35b05a5 100644
--- a/Pipfile
+++ b/Pipfile
@@ -5,6 +5,7 @@ name = "pypi"
 
 [packages]
 irc3 = "*"
+urlextract = "*"
 
 [dev-packages]
 
diff --git a/Pipfile.lock b/Pipfile.lock
index eee9531..937da08 100644
--- a/Pipfile.lock
+++ b/Pipfile.lock
@@ -1,7 +1,7 @@
 {
     "_meta": {
         "hash": {
-            "sha256": "a2d7b3880bcd5122a42c9d5a0d2e5ade937786b9b36437bfefa0eba5ef0359fd"
+            "sha256": "3cd5a528e307bfbf3392d0a088d77f1bc83dc37db66b63cc7b3205a6a0df2bfa"
         },
         "pipfile-spec": 6,
         "requires": {
@@ -16,12 +16,35 @@
         ]
     },
     "default": {
+        "appdirs": {
+            "hashes": [
+                "sha256:7d5d0167b2b1ba821647616af46a749d1c653740dd0d2415100fe26e27afdf41",
+                "sha256:a841dacd6b99318a741b166adb07e19ee71a274450e68237b4650ca1055ab128"
+            ],
+            "version": "==1.4.4"
+        },
         "docopt": {
             "hashes": [
                 "sha256:49b3a825280bd66b3aa83585ef59c4a8c82f2c8a522dbe754a8bc8d08c85c491"
             ],
             "version": "==0.6.2"
         },
+        "filelock": {
+            "hashes": [
+                "sha256:7afc856f74fa7006a289fd10fa840e1eebd8bbff6bffb69c26c54a0512ea8cf8",
+                "sha256:bb2a1c717df74c48a2d00ed625e5a66f8572a3a30baacb7657add1d7bac4097b"
+            ],
+            "markers": "python_version >= '3.6'",
+            "version": "==3.3.2"
+        },
+        "idna": {
+            "hashes": [
+                "sha256:84d9dd047ffa80596e0f246e2eab0b391788b0503584e8945f2368256d2735ff",
+                "sha256:9d643ff0a55b762d5cdb124b8eaa99c66322e2157b69160bc32796e824360e6d"
+            ],
+            "markers": "python_version >= '3.5'",
+            "version": "==3.3"
+        },
         "irc3": {
             "hashes": [
                 "sha256:088b7be88817dee5d02e7e06c9dbf503c8e807a8cccc9b36b5be419918574cf1",
@@ -30,6 +53,22 @@
             "index": "pypi",
             "version": "==1.1.7"
         },
+        "uritools": {
+            "hashes": [
+                "sha256:28ffef82ce3b2793237d36e45aa7cde28dae6502f6a93fdbd05ede401520e279",
+                "sha256:576737664f51f82d5c2a98e25f6c5da73a57cc88326dbb686fd6c5d06ebd6c29"
+            ],
+            "markers": "python_version ~= '3.5'",
+            "version": "==3.0.2"
+        },
+        "urlextract": {
+            "hashes": [
+                "sha256:55c401dddf12aa65c09f0fcc02eff2be6a28f6e305797ed6ff4b8ed26100e860",
+                "sha256:669f07192584b841b49ba8868fbd6b00e7ddc28367d36a3d8ca8c8e429420748"
+            ],
+            "index": "pypi",
+            "version": "==1.4.0"
+        },
         "venusian": {
             "hashes": [
                 "sha256:06e7385786ad3a15c70740b2af8d30dfb063a946a851dcb4159f9e2a2302578f",
diff --git a/_twitchbot/config.py b/_twitchbot/config.py
index ba1773e..0cab307 100644
--- a/_twitchbot/config.py
+++ b/_twitchbot/config.py
@@ -158,6 +158,16 @@ class Config:
                     moderator_config.get("max-msg-occurrences", None),
                     moderator_config.get("min-time-between-occurrence", None)
                 ))
+            if mod == 'links':
+                moderators.append(moderator.LinksModerator(
+                    moderator_config.get(
+                        "message",
+                        "{author}, your message contained forbidden links, it had to be removed for safety."
+                    ),
+                    cls.parse_decision(moderator_config.get("decision", "delete")),
+                    moderator_config.get("duration", None),
+                    moderator_config.get("authorized", [])
+                ))
 
         # Generate help command
         if params.get('help', True):
diff --git a/_twitchbot/moderator.py b/_twitchbot/moderator.py
index 219edc7..40523b0 100644
--- a/_twitchbot/moderator.py
+++ b/_twitchbot/moderator.py
@@ -19,6 +19,9 @@ from abc import ABC, abstractmethod
 from enum import Enum
 from typing import Union
 from datetime import datetime, timedelta
+from urlextract import URLExtract
+from fnmatch import fnmatch
+import re
 
 EPOCH = datetime(1970, 1, 1)
 
@@ -148,3 +151,52 @@ class FloodModerator(Moderator):
 
     def declare_raid(self):
         self.last_raid = datetime.now()
+
+
+class LinksModerator(Moderator):
+    def __init__(
+            self,
+            message: str,
+            decision: ModerationDecision,
+            timeout_duration: Union[None, int],
+            authorized_urls: [str]
+    ):
+        super().__init__(message, decision, timeout_duration)
+        self.authorized_urls = authorized_urls
+
+    def get_name(self) -> str:
+        return 'Link'
+
+    def vote(self, msg: str, author: str) -> ModerationDecision:
+        url_extractor = URLExtract()
+        links = url_extractor.find_urls(msg)
+
+        if len(links) == 0:
+            return ModerationDecision.ABSTAIN
+
+        if not self.are_urls_authorized(links):
+            return self.decision
+
+        return ModerationDecision.ABSTAIN
+
+    def are_urls_authorized(self, links: [str]) -> bool:
+
+        for link in links:
+            is_link_authorized = False
+            print(link)
+
+            for pattern in self.authorized_urls:
+                print(pattern)
+                if dump(fnmatch(link, pattern)) or dump(fnmatch(f"http://{link}", pattern)) or dump(fnmatch(f"https://{link}", pattern)):
+                    is_link_authorized = True
+                    break
+
+            if not is_link_authorized:
+                return False
+
+        return True
+
+
+def dump(what):
+    print(what)
+    return what
\ No newline at end of file