From 61b22077952c5bfdf227eed326bed098468f0abc Mon Sep 17 00:00:00 2001
From: Bob Mottram <bob@libreserver.org>
Date: Wed, 27 Jul 2022 10:13:30 +0100
Subject: [PATCH] Check for twitter content within announces

---
 content.py | 18 ++++++++++++++++++
 inbox.py   | 30 +++++++-----------------------
 posts.py   | 12 +++++++++---
 3 files changed, 34 insertions(+), 26 deletions(-)

diff --git a/content.py b/content.py
index 9ef638d44..e184df2aa 100644
--- a/content.py
+++ b/content.py
@@ -1941,3 +1941,21 @@ def remove_script(content: str, log_filename: str,
                         print('EX: cannot append to svg script log')
             content = content.replace(text, '')
     return content
+
+
+def reject_twitter_summary(base_dir: str, nickname: str, domain: str,
+                           summary: str) -> bool:
+    """Returns true if the post should be rejected due to twitter
+    existing within the summary
+    """
+    if not summary:
+        return False
+    remove_twitter = \
+        acct_dir(base_dir, nickname, domain) + '/.removeTwitter'
+    if not os.path.isfile(remove_twitter):
+        return False
+    summary_lower = summary.lower()
+    if 'twitter' in summary_lower or \
+       'birdsite' in summary_lower:
+        return True
+    return False
diff --git a/inbox.py b/inbox.py
index db56a03e9..ea521a2cc 100644
--- a/inbox.py
+++ b/inbox.py
@@ -131,6 +131,7 @@ from conversation import update_conversation
 from webapp_hashtagswarm import html_hash_tag_swarm
 from person import valid_sending_actor
 from fitnessFunctions import fitness_performance
+from content import reject_twitter_summary
 from content import load_dogwhistles
 from content import valid_url_lengths
 from content import remove_script
@@ -567,22 +568,6 @@ def inbox_permitted_message(domain: str, message_json: {},
     return True
 
 
-def _reject_twitter_summary(base_dir: str, nickname: str, domain: str,
-                            summary: str) -> bool:
-    """Returns true if the post should be rejected due to twitter
-    existing within the summary
-    """
-    remove_twitter = \
-        acct_dir(base_dir, nickname, domain) + '/.removeTwitter'
-    if not os.path.isfile(remove_twitter):
-        return False
-    summary_lower = summary.lower()
-    if 'twitter' in summary_lower or \
-       'birdsite' in summary_lower:
-        return True
-    return False
-
-
 def save_post_to_inbox_queue(base_dir: str, http_prefix: str,
                              nickname: str, domain: str,
                              post_json_object: {},
@@ -671,13 +656,12 @@ def save_post_to_inbox_queue(base_dir: str, http_prefix: str,
             if debug:
                 print('WARN: post was filtered out due to content')
             return None
-        if summary_str:
-            if _reject_twitter_summary(base_dir, nickname, domain,
-                                       summary_str):
-                if debug:
-                    print('WARN: post was filtered out due to ' +
-                          'twitter content')
-                return None
+        if reject_twitter_summary(base_dir, nickname, domain,
+                                  summary_str):
+            if debug:
+                print('WARN: post was filtered out due to ' +
+                      'twitter summary')
+            return None
 
     original_post_id = None
     if post_json_object.get('id'):
diff --git a/posts.py b/posts.py
index 58dbbca83..f65e97d1f 100644
--- a/posts.py
+++ b/posts.py
@@ -79,6 +79,7 @@ from media import get_music_metadata
 from media import attach_media
 from media import replace_you_tube
 from media import replace_twitter
+from content import reject_twitter_summary
 from content import words_similarity
 from content import limit_repeated_words
 from content import post_tag_exists
@@ -5026,15 +5027,20 @@ def download_announce(session, base_dir: str, http_prefix: str,
                              base_dir, nickname, domain, post_id,
                              recent_posts_cache)
             return None
-
+        if reject_twitter_summary(base_dir, nickname, domain,
+                                  summary_str):
+            print('WARN: announced post has twitter summary ' +
+                  str(announced_json))
+            _reject_announce(announce_filename,
+                             base_dir, nickname, domain, post_id,
+                             recent_posts_cache)
+            return None
         if invalid_ciphertext(content_str):
             print('WARN: announced post contains invalid ciphertext ' +
                   str(announced_json))
             _reject_announce(announce_filename,
                              base_dir, nickname, domain, post_id,
                              recent_posts_cache)
-            print('WARN: Invalid ciphertext within announce ' +
-                  str(announced_json))
             return None
 
         # remove any long words