From 75fe0bd29c9a93076b6d8b3db887ce6d11f58fe5 Mon Sep 17 00:00:00 2001
From: user12986714 <65436504+user12986714@users.noreply.github.com>
Date: Thu, 16 Jul 2020 23:31:21 -0400
Subject: [PATCH] Strip code block in mostly-img reason
This should be reworked once https://chat.stackexchange.com/transcript/message/54842978 get implemented.
---
findspam.py | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/findspam.py b/findspam.py
index 7420ac0c21..5cfabc9b1f 100644
--- a/findspam.py
+++ b/findspam.py
@@ -608,12 +608,19 @@ def len_img_block(string):
# max_score=2 to prevent voting fraud
@create_rule("post is mostly images", title=False, max_rep=201, max_score=2)
def mostly_img(s, site):
- if len(s) == 0:
+ s_len_orig = len(s)
+ if s_len_orig == 0:
return False, ""
+ # Strip code blocks manually. This should be removed once feature
+ # https://chat.stackexchange.com/transcript/message/54842978
+ # get implemented.
+ s = regex.sub("(?s).*?
", "\nstripped pre\n", s)
+ s = regex.sub("(?s).*?", "\nstripped code\n", s)
+
s_len_img = len_img_block(s)
- if s_len_img / len(s) > IMG_TXT_R_THRES:
- return True, "{:.4f} of the post is html image blocks".format(s_len_img / len(s))
+ if s_len_img / s_len_orig > IMG_TXT_R_THRES:
+ return True, "{:.4f} of the post is html image blocks".format(s_len_img / s_len_orig)
return False, ""