forked from Charcoal-SE/SmokeDetector
-
Notifications
You must be signed in to change notification settings - Fork 0
/
findspam.py
1698 lines (1536 loc) · 96.7 KB
/
findspam.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# -*- coding: utf-8 -*-
# noinspection PyCompatibility
import math
import regex
from difflib import SequenceMatcher
from urllib.parse import urlparse, unquote_plus
from itertools import chain
from collections import Counter
from datetime import datetime
import os.path as path
# noinspection PyPackageRequirements
import tld
# noinspection PyPackageRequirements
from tld.utils import TldDomainNotFound
import phonenumbers
import dns.resolver
import requests
import chatcommunicate
from helpers import log
from globalvars import GlobalVars
from blacklists import load_blacklists
TLD_CACHE = []
LEVEN_DOMAIN_DISTANCE = 3
SIMILAR_THRESHOLD = 0.95
SIMILAR_ANSWER_THRESHOLD = 0.7
BODY_TITLE_SIMILAR_RATIO = 0.90
CHARACTER_USE_RATIO = 0.42
PUNCTUATION_RATIO = 0.42
REPEATED_CHARACTER_RATIO = 0.20
EXCEPTION_RE = r"^Domain (.*) didn't .*!$"
RE_COMPILE = regex.compile(EXCEPTION_RE)
COMMON_MALFORMED_PROTOCOLS = [
('httl://', 'http://'),
]
# These types of files frequently get caught as "misleading link"
SAFE_EXTENSIONS = {'htm', 'py', 'java', 'sh'}
SE_SITES_RE = r'(?:{sites})'.format(
sites='|'.join([
r'(?:[a-z]+\.)*stackoverflow\.com',
r'(?:{doms})\.com'.format(doms='|'.join(
[r'askubuntu', r'superuser', r'serverfault', r'stackapps', r'imgur'])),
r'mathoverflow\.net',
r'(?:[a-z]+\.)*stackexchange\.com']))
SE_SITES_DOMAINS = ['stackoverflow.com', 'askubuntu.com', 'superuser.com', 'serverfault.com',
'mathoverflow.net', 'stackapps.com', 'stackexchange.com', 'sstatic.net',
'imgur.com'] # Frequently catching FP
WHITELISTED_WEBSITES_REGEX = regex.compile(r"(?i)upload|\b(?:{})\b".format("|".join([
"yfrog", "gfycat", "tinypic", "sendvid", "ctrlv", "prntscr", "gyazo", r"youtu\.?be", "past[ie]", "dropbox",
"microsoft", "newegg", "cnet", "regex101", r"(?<!plus\.)google", "localhost", "ubuntu", "getbootstrap",
"jsfiddle\.net", "codepen\.io", "pastebin"
] + [se_dom.replace(".", r"\.") for se_dom in SE_SITES_DOMAINS])))
if GlobalVars.perspective_key:
PERSPECTIVE = "https://commentanalyzer.googleapis.com/v1alpha1/comments:analyze?key=" + GlobalVars.perspective_key
PERSPECTIVE_THRESHOLD = 0.85 # conservative
# Flee before the ugly URL validator regex!
# We are using this, instead of a nice library like BeautifulSoup, because spammers are
# stupid and don't always know how to actually *link* their web site. BeautifulSoup misses
# those plain text URLs.
# https://gist.github.com/dperini/729294#gistcomment-1296121
URL_REGEX = regex.compile(
r"""((?:(?:https?|ftp)://)(?:\S+(?::\S*)?@)?(?:(?!(?:10|127)"""
r"""(?:\.\d{1,3}){3})(?!(?:169\.254|192\.168)(?:\.\d{1,3}){2}))"""
r"""(?!172\.(?:1[6-9]|2\d|3[0-1])(?:\.\d{1,3}){2})(?:[1-9]\d?|1\d\d|2[01]\d|22[0-3])"""
r"""(?:\.(?:1?\d{1,2}|2[0-4]\d|25[0-5])){2}(?:\.(?:[1-9]\d?|1\d\d|2[0-4]\d|25[0-4]))"""
r"""|(?:(?:[A-Za-z\u00a1-\uffff0-9]-?)*[A-Za-z\u00a1-\uffff0-9]+)(?:\.(?:[A-Za-z\u00a1-\uffff0-9]-?)"""
r"""*[A-Za-z\u00a1-\uffff0-9]+)*(?:\.(?:[A-Za-z\u00a1-\uffff]{2,})))(?::\d{2,5})?(?:/\S*)?""", regex.UNICODE)
UNIFORM = math.log(1 / 36)
UNIFORM_PRIOR = math.log(1 / 5)
ENGLISH = {
'a': -2.56940287968626,
'e': -2.6325365263400786,
'o': -2.9482912667071903,
'r': -2.9867566750238046,
'i': -3.043195438576378,
's': -3.053589802306065,
'n': -3.0696364572432233,
'1': -3.134872509228817,
't': -3.230441879550407,
'l': -3.2558408400221905,
'2': -3.4663376838336166,
'm': -3.4810979044444426,
'd': -3.5635447023561517,
'0': -3.5958227205042967,
'c': -3.6348280308631855,
'p': -3.6771505079154236,
'3': -3.7158848391017765,
'h': -3.7019152926538648,
'b': -3.74138548356748,
'u': -3.8457967842578014,
'k': -3.9048726800430713,
'4': -3.9411171656325226,
'5': -3.9708339604329925,
'g': -3.961715896933319,
'9': -4.019842096462643,
'6': -4.041864072829501,
'8': -4.096998079687665,
'7': -4.122126943234552,
'y': -4.1666976658279635,
'f': -4.351040269361279,
'w': -4.360690517108493,
'j': -4.741006747760368,
'v': -4.759276833451455,
'z': -5.036594538526155,
'x': -5.137009730369897,
'q': -5.624531280146579
}
ENGLISH_PRIOR = math.log(4 / 5)
def is_whitelisted_website(url):
# Imported from method link_at_end
return bool(WHITELISTED_WEBSITES_REGEX.search(url))
def levenshtein(s1, s2):
if len(s1) < len(s2):
return levenshtein(s2, s1)
if len(s2) == 0:
return len(s1)
previous_row = range(len(s2) + 1)
for i, c1 in enumerate(s1):
current_row = [i + 1]
for j, c2 in enumerate(s2):
insertions = previous_row[j + 1] + 1
deletions = current_row[j] + 1
substitutions = previous_row[j] + (c1 != c2)
current_row.append(min(insertions, deletions, substitutions))
previous_row = current_row
return previous_row[-1]
def contains_tld(s):
global TLD_CACHE
# Hackity hack.
if len(TLD_CACHE) == 0:
with open(path.join(tld.defaults.NAMES_LOCAL_PATH_PARENT, tld.defaults.NAMES_LOCAL_PATH), 'r') as f:
TLD_CACHE = [x.rstrip('\n') for x in f.readlines() if x.rstrip('\n') and
not x.strip().startswith('//')]
return any(('.' + x) in s for x in TLD_CACHE)
def misleading_link(s, site):
link_regex = r"<a href=\"([^\"]+)\"[^>]*>([^<]+)<\/a>"
compiled = regex.compile(link_regex)
search = compiled.search(s)
if search is None:
return False, ''
href, text = search[1], search[2]
try:
parsed_href = tld.get_tld(href, as_object=True)
log('debug', parsed_href.domain, SE_SITES_DOMAINS)
if parsed_href.fld in SE_SITES_DOMAINS:
return False, ''
if contains_tld(text) and ' ' not in text:
parsed_text = tld.get_tld(text, fix_protocol=True, as_object=True)
else:
raise tld.exceptions.TldBadUrl('Link text is not a URL')
except (tld.exceptions.TldDomainNotFound, tld.exceptions.TldBadUrl, ValueError) as err:
return False, ''
if site == 'stackoverflow.com' and parsed_text.fld.split('.')[-1] in SAFE_EXTENSIONS:
return False, ''
if levenshtein(parsed_href.domain, parsed_text.domain) <= LEVEN_DOMAIN_DISTANCE: # Preempt
return False, ''
try:
href_domain = unquote_plus(parsed_href.domain.encode("ascii").decode("idna"))
except ValueError:
href_domain = parsed_href.domain
try:
text_domain = unquote_plus(parsed_text.domain.encode("ascii").decode("idna")) # people do post this, sad
except ValueError:
text_domain = parsed_text.domain
if levenshtein(href_domain, text_domain) > LEVEN_DOMAIN_DISTANCE:
return True, 'Domain {} indicated by possible misleading text {}.'.format(
parsed_href.fld, parsed_text.fld
)
else:
return False, ''
# noinspection PyUnusedLocal,PyMissingTypeHints,PyTypeChecker
def has_repeating_words(s, site):
words = regex.split(r"[\s.,;!/\()\[\]+_-]", s)
words = [word for word in words if word != ""]
streak = 0
prev = ""
for word in words:
if word == prev and word.isalpha() and len(word) > 1:
streak += 1
else:
streak = 0
prev = word
if streak >= 5 and streak * len(word) >= 0.2 * len(s):
return True, "Repeated word: *{}*".format(word)
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_few_characters(s, site):
s = regex.sub("</?(?:p|strong|em)>", "", s).rstrip() # remove HTML paragraph tags from posts
uniques = len(set(s) - {"\n", "\t"})
if (len(s) >= 30 and uniques <= 6) or (len(s) >= 100 and uniques <= 15): # reduce if false reports appear
if uniques >= 5 and site == "math.stackexchange.com":
# Special case for Math.SE: Uniques case may trigger false-positives.
return False, ""
return True, "Contains {} unique character{}".format(uniques, "s" if uniques >= 2 else "")
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_repeating_characters(s, site):
s = s.strip().replace("\u200B", "").replace("\u200C", "") # Strip leading and trailing spaces
if "\n" in s or "<code>" in s or "<pre>" in s:
return False, ""
s = regex.sub(URL_REGEX, "", s) # Strip URLs for this check
if not s:
return False, ""
matches = regex.compile(r"([^\s_.,?!=~*/0-9-])(\1{9,})", regex.UNICODE).findall(s)
match = "".join(["".join(match) for match in matches])
if len(match) / len(s) >= REPEATED_CHARACTER_RATIO: # Repeating characters make up >= 20 percent
return True, "Repeated character: {}".format(", ".join(
["{}*{}".format(repr(match[0]), len(''.join(match))) for match in matches]))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def link_at_end(s, site): # link at end of question, on selected sites
s = regex.sub("</?(?:strong|em|p)>", "", s)
match = regex.compile(r"(?i)https?://(?:[.A-Za-z0-9-]*/?[.A-Za-z0-9-]*/?|plus\.google\.com/"
r"[\w/]*|www\.pinterest\.com/pin/[\d/]*)(?=</a>\s*$)").search(s)
if match and not is_whitelisted_website(match.group(0)):
return True, u"Link at end: {}".format(match.group(0))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints,PyTypeChecker
def non_english_link(s, site): # non-english link in short answer
if len(s) < 600:
links = regex.compile(r'nofollow(?: noreferrer)?">([^<]*)(?=</a>)', regex.UNICODE).findall(s)
for link_text in links:
word_chars = regex.sub(r"(?u)\W", "", link_text)
non_latin_chars = regex.sub(r"\w", "", word_chars)
if len(word_chars) >= 1 and ((len(word_chars) <= 20 and len(non_latin_chars) >= 1) or
(len(non_latin_chars) >= 0.05 * len(word_chars))):
return True, u"Non-English link text: *{}*".format(link_text)
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints,PyTypeChecker
def mostly_non_latin(s, site): # majority of post is in non-Latin, non-Cyrillic characters
word_chars = regex.sub(r'(?u)[\W0-9]|http\S*', "", s)
non_latin_chars = regex.sub(r"(?u)\p{script=Latin}|\p{script=Cyrillic}", "", word_chars)
if len(non_latin_chars) > 0.4 * len(word_chars):
return True, u"Text contains {} non-Latin characters out of {}".format(len(non_latin_chars), len(word_chars))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_phone_number(s, site):
if regex.compile(r"(?i)\b(address(es)?|run[- ]?time|error|value|server|hostname|timestamp|warning|code|"
r"(sp)?exception|version|chrome|1234567)\b", regex.UNICODE).search(s):
return False, "" # not a phone number
s = regex.sub("[^A-Za-z0-9\\s\"',|]", "", s) # deobfuscate
s = regex.sub("[Oo]", "0", s)
s = regex.sub("[Ss]", "5", s)
s = regex.sub("[Iil|]", "1", s)
matched = regex.compile(r"(?<!\d)(?:\d{2}\s?\d{8,11}|\d\s{0,2}\d{3}\s{0,2}\d{3}\s{0,2}\d{4}|8\d{2}"
r"\s{0,2}\d{3}\s{0,2}\d{4})(?!\d)", regex.UNICODE).findall(s)
test_formats = ["IN", "US", "NG", None] # ^ don't match parts of too long strings of digits
for phone_number in matched:
if regex.compile(r"^21474(672[56]|8364)|^192168|^3221225").search(phone_number):
return False, "" # error code or limit of int size, or 192.168 IP, or 0xC000000_ error code
for testf in test_formats:
try:
z = phonenumbers.parse(phone_number, testf)
if phonenumbers.is_possible_number(z) and phonenumbers.is_valid_number(z):
log('debug', "Possible {}, Valid {}, Explain: {}".format(phonenumbers.is_possible_number(z),
phonenumbers.is_valid_number(z), z))
return True, u"Phone number: {}".format(phone_number)
except phonenumbers.phonenumberutil.NumberParseException:
pass
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_customer_service(s, site): # flexible detection of customer service in titles
s = s[0:300].lower() # if applied to body, the beginning should be enough: otherwise many false positives
s = regex.sub(r"[^A-Za-z0-9\s]", "", s) # deobfuscate
phrase = regex.compile(r"(tech(nical)? support)|((support|service|contact|help(line)?) (telephone|phone|"
r"number))").search(s)
if phrase and site in ["askubuntu.com", "webapps.stackexchange.com", "webmasters.stackexchange.com"]:
return True, u"Key phrase: *{}*".format(phrase.group(0))
business = regex.compile(
r"(?i)\b(airlines?|apple|AVG|BT|netflix|dell|Delta|epson|facebook|gmail|google|hotmail|hp|"
r"lexmark|mcafee|microsoft|norton|out[l1]ook|quickbooks|sage|windows?|yahoo)\b").search(s)
digits = len(regex.compile(r"\d").findall(s))
if business and digits >= 5:
keywords = regex.compile(r"(?i)\b(customer|help|care|helpline|reservation|phone|recovery|service|support|"
r"contact|tech|technical|telephone|number)\b").findall(s)
if len(set(keywords)) >= 2:
matches = ", ".join(["".join(match) for match in keywords])
return True, u"Scam aimed at *{}* customers. Keywords: *{}*".format(business.group(0), matches)
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_health(s, site): # flexible detection of health spam in titles
s = s[0:200] # if applied to body, the beginning should be enough: otherwise many false positives
capitalized = len(regex.compile(r"\b[A-Z][a-z]").findall(s)) >= 5 # words beginning with uppercase letter
organ = regex.compile(r"(?i)\b(colon|skin|muscle|bicep|fac(e|ial)|eye|brain|IQ|mind|head|hair|peni(s|le)|"
r"breast|body|joint|belly|digest\w*)s?\b").search(s)
condition = regex.compile(r"(?i)\b(weight|constipat(ed|ion)|dysfunction|swollen|sensitive|wrinkle|aging|"
r"suffer|acne|pimple|dry|clog(ged)?|inflam(ed|mation)|fat|age|pound)s?\b").search(s)
goal = regex.compile(r"(?i)\b(supple|build|los[es]|power|burn|erection|tone(d)|rip(ped)?|bulk|get rid|mood)s?\b|"
r"\b(diminish|look|reduc|beaut|renew|young|youth|lift|eliminat|enhance|energ|shred|"
r"health(?!kit)|improve|enlarge|remov|vital|slim|lean|boost|str[oe]ng)").search(s)
remedy = regex.compile(r"(?i)\b(remed(y|ie)|serum|cleans?(e|er|ing)|care|(pro)?biotic|herbal|lotion|cream|"
r"gel|cure|drug|formula|recipe|regimen|solution|therapy|hydration|soap|treatment|supplement|"
r"diet|moist\w*|injection|potion|ingredient|aid|exercise|eat(ing)?)s?\b").search(s)
boast = regex.compile(r"(?i)\b(most|best|simple|top|pro|real|mirac(le|ulous)|secrets?|organic|natural|perfect|"
r"ideal|fantastic|incredible|ultimate|important|reliable|critical|amazing|fast|good)\b|"
r"\b(super|hyper|advantag|benefi|effect|great|valu|eas[iy])").search(s)
other = regex.compile(r"(?i)\b(product|thing|item|review|advi[cs]e|myth|make use|your?|really|work|tip|shop|"
r"store|method|expert|instant|buy|fact|consum(e|ption)|baby|male|female|men|women|grow|"
r"idea|suggest\w*|issue)s?\b").search(s)
score = 4 * bool(organ) + 2 * bool(condition) + 2 * bool(goal) + 2 * bool(remedy) + bool(boast) + \
bool(other) + capitalized
if score >= 8:
match_objects = [organ, condition, goal, remedy, boast, other]
words = [match.group(0) for match in match_objects if match]
return True, u"Health-themed spam (score {}). Keywords: *{}*".format(score, ", ".join(words).lower())
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def pattern_product_name(s, site):
# Always use (?: non-capturing groups ) in the keywords list
keywords = [
"Testo", "Derma?(?:pholia)?", "Garcinia", "Cambogia", "Aurora", "Diet", "Slim", "Premier", "(?:Pure)?Fit",
"Junivive", "Gain", "Allure", "Nuvella", "Blast", "Burn", "Perfect",
"Elite", "Force", "Exceptional", "Enhance(?:ment)?", "Nitro", "Max+", "Boost", "E?xtreme", "Grow",
"Deep", "Male", "Pro", "Advanced", "Monster", "Divine", "Royale", "Angele*", "Trinity", "Andro",
"Pure", "Skin", "Sea", "Muscle", "Ascend", "Youth", "Hyper(?:tone)?", "Boost(?:er)?",
"Serum", "Supplement", "Fuel", "Cream", "Keto", "Rapid", "Tone", "Forskolin", "Neuro", "Luma"
"(?:Anti-)?Ag(?:ed?|ing)", "Trim", "Premi(?:um|er)", "Vital", "Master", "Ultra", "Radiant(?:ly)?",
]
if site not in {"math.stackexchange.com", "mathoverflow.net"}:
keywords += [r"X[\dLOST]?", "Alpha", "Plus", "Prime", "Formula"]
keywords = "|".join(keywords)
three_words = regex.compile(r"(?i)\b(({0})[ -]({0})[ -]({0}))\b".format(keywords)).findall(s)
two_words = regex.compile(r"(?i)\b(({0})[ -]({0}))\b".format(keywords)).findall(s)
unique_three_words = sum([len(m[1:]) == len(set([regex.sub(r"(?i)X\d", "X0", w) for w in m[1:]]))
for m in three_words])
unique_two_words = sum([len(m[1:]) == len(set([regex.sub(r"(?i)X\d", "X0", w) for w in m[1:]]))
for m in two_words])
if unique_three_words >= 1:
return True, u"Pattern-matching product name *{}*".format(", ".join([match[0] for match in set(three_words)]))
elif unique_two_words >= 2:
return True, u"Pattern-matching product name *{}*".format(", ".join([match[0] for match in set(two_words)]))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def what_is_this_pharma_title(s, site): # title "what is this Xxxx?"
if regex.compile(r'^what is this (?:[A-Z]|https?://)').match(s):
return True, u'Title starts with "what is this"'
else:
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def keyword_email(s, site): # a keyword and an email in the same post
if regex.compile("<pre>|<code>").search(s) and site == "stackoverflow.com": # Avoid false positives on SO
return False, ""
keyword = regex.compile(r"(?i)\b(training|we (will )?(offer|develop|provide)|sell|invest(or|ing|ment)|credit|"
r"money|quality|legit|interest(ed)?|guarantee|rent|crack|opportunity|fundraising|campaign|"
r"career|employment|candidate|loan|lover|husband|wife|marriage|illuminati|brotherhood|"
r"(join|contact) (me|us|him)|reach (us|him)|spell(caster)?|doctor|cancer|krebs|"
r"(cheat|hack)(er|ing)?|spying|passport|seaman|scam|pics|vampire|bless(ed)?|atm|miracle|"
r"cure|testimony|kidney|hospital|wetting)s?\b| Dr\.? |\$ ?[0-9,.]{4}|@qq\.com|"
r"\b(герпес|муж|жена|доктор|болезн)").search(s)
email = regex.compile(r"(?<![=#/])\b[A-z0-9_.%+-]+\b(?:@|\(?at\)?)\b(?!(example|domain|site|foo|\dx)"
r"(?:\.|\(?dot\)?)[A-z]{2,4})\b(?:[A-z0-9_.%+-]|\(?dot\)?)+\b"
r"(?:\.|\(?dot\)?)[A-z]{2,4}\b").search(s)
if keyword and email:
return True, u"Keyword *{}* with email {}".format(keyword.group(0), email.group(0))
obfuscated_email = regex.compile(r"(?<![=#/])\b[A-z0-9_.%+-]+ *@ *(g *mail|yahoo) *\. *com\b").search(s)
if obfuscated_email and not email:
return True, u"Obfuscated email {}".format(obfuscated_email.group(0))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def pattern_email(s, site):
pattern = regex.compile(r"(?i)(?<![=#/])\b(dr|[A-z0-9_.%+-]*"
r"(loan|hack|financ|fund|spell|temple|herbal|spiritual|atm|heal|priest|classes|"
r"investment))[A-z0-9_.%+-]*"
r"@(?!(example|domain|site|foo|\dx)\.[A-z]{2,4})[A-z0-9_.%+-]+\.[A-z]{2,4}\b"
).search(s)
if pattern:
return True, u"Pattern-matching email {}".format(pattern.group(0))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def keyword_link(s, site): # thanking keyword and a link in the same short answer
if len(s) > 400:
return False, ""
link = regex.compile(r'(?i)<a href="https?://\S+').search(s)
if not link or is_whitelisted_website(link.group(0)):
return False, ""
praise = regex.compile(r"(?i)\b(nice|good|interesting|helpful|great|amazing) (article|blog|post|information)\b|"
r"very useful").search(s)
thanks = regex.compile(r"(?i)\b(appreciate|than(k|ks|x))\b").search(s)
keyword = regex.compile(r"(?i)\b(I really appreciate|many thanks|thanks a lot|thank you (very|for)|"
r"than(ks|x) for (sharing|this|your)|dear forum members|(very (informative|useful)|"
r"stumbled upon (your|this)|wonderful|visit my) (blog|site|website))\b").search(s)
if link and keyword:
return True, u"Keyword *{}* with link {}".format(keyword.group(0), link.group(0))
if link and thanks and praise:
return True, u"Keywords *{}*, *{}* with link {}".format(thanks.group(0), praise.group(0), link.group(0))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def bad_link_text(s, site): # suspicious text of a hyperlink
s = regex.sub("</?strong>|</?em>", "", s) # remove font tags
keywords = regex.compile(
r"(?isu)"
r"\b(buy|cheap) |live[ -]?stream|"
r"\bmake (money|\$)|"
r"\b(porno?|(whole)?sale|coins|replica|luxury|coupons?|essays?|in \L<city>)\b|"
r"\b\L<city>(?:\b.{1,20}\b)?(service|escort|call girls?)|"
r"(best|make|full|hd|software|cell|data)[\w ]{1,20}(online|service|company|repair|recovery|school|university)|"
r"\b(writing (service|help)|essay (writing|tips))", city=FindSpam.city_list)
links = regex.compile(r'nofollow(?: noreferrer)?">([^<]*)(?=</a>)', regex.UNICODE).findall(s)
business = regex.compile(
r"(?i)(^| )(airlines?|apple|AVG|BT|netflix|dell|Delta|epson|facebook|gmail|google|hotmail|hp|"
r"lexmark|mcafee|microsoft|norton|out[l1]ook|quickbooks|sage|windows?|yahoo)($| )")
support = regex.compile(r"(?i)(^| )(customer|care|helpline|reservation|phone|recovery|service|support|contact|"
r"tech|technical|telephone|number)($| )")
for link_text in links:
keywords_match = keywords.search(link_text)
if keywords_match:
return True, u"Bad keyword *{}* in link text".format(keywords_match.group(0).strip())
business_match = business.search(link_text)
support_match = support.search(link_text)
if business_match and support_match:
return True, u"Bad keywords *{}*, *{}* in link text".format(business_match.group(0).strip(),
support_match.group(0).strip())
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def bad_pattern_in_url(s, site):
patterns = [
r'[^"]*-reviews?(?:-(?:canada|(?:and|or)-scam))?/?',
r'[^"]*-support/?',
]
matches = regex.compile(
r'<a href="(?P<frag>{0})"|<a href="[^"]*"(?:\s+"[^"]*")*>(?P<frag>{0})</a>'.format(
'|'.join(patterns)), regex.UNICODE).findall(s)
matches = [x for x in matches if not regex.match(
r'^https?://{0}'.format(SE_SITES_RE), x[0])]
if matches:
return True, u"Bad fragment in link {}".format(
", ".join(["".join(match) for match in matches]))
else:
return False, ""
def ns_for_url_domain(s, site, nslist):
invalid_tld_count = 0
for nsentry in nslist:
if isinstance(nsentry, set):
for ns in nsentry:
assert ns.endswith('.'),\
"Missing final dot on NS entry {0}".format(ns)
else:
assert nsentry.endswith('.'),\
"Missing final dot on NS entry {0}".format(nsentry)
for domain in set([get_domain(link, full=True) for link in post_links(s)]):
if not tld.get_tld(domain, fix_protocol=True, fail_silently=True):
log('debug', '{0} has no valid tld; skipping'.format(domain))
invalid_tld_count += 1
if invalid_tld_count > 3:
log('debug', 'too many invalid TLDs; abandoning post')
return False, ""
continue
try:
starttime = datetime.now()
ns = dns.resolver.query(domain, 'ns')
except dns.exception.DNSException as exc:
if str(exc).startswith('None of DNS query names exist:'):
log('debug', 'domain {0} not found; skipping'.format(domain))
continue
endtime = datetime.now()
log('warning', 'DNS error {0} (duration: {1})'.format(
exc, endtime - starttime))
continue
endtime = datetime.now()
log('debug', 'NS query duration {0}'.format(endtime - starttime))
nameservers = set([server.target.to_text() for server in ns])
for ns_candidate in nslist:
if (type(ns_candidate) is set and nameservers == ns_candidate) or \
any(ns.endswith('.{0}'.format(ns_candidate))
for ns in nameservers):
return True, '{domain} NS suspicious {ns}'.format(
domain=domain, ns=','.join(nameservers))
return False, ""
def bad_ns_for_url_domain(s, site):
return ns_for_url_domain(s, site, [
# Don't forget the trailing dot on the resolved name!
{'ns1.md-95.bigrockservers.com.', 'ns2.md-95.bigrockservers.com.'},
{'ns1.md-99.bigrockservers.com.', 'ns2.md-99.bigrockservers.com.'},
{'apollo.ns.cloudflare.com.', 'liz.ns.cloudflare.com.'},
{'ara.ns.cloudflare.com.', 'greg.ns.cloudflare.com.'},
{'brenda.ns.cloudflare.com.', 'merlin.ns.cloudflare.com.'},
{'chip.ns.cloudflare.com.', 'lola.ns.cloudflare.com.'},
{'lee.ns.cloudflare.com.', 'ulla.ns.cloudflare.com.'},
{'lloyd.ns.cloudflare.com.', 'reza.ns.cloudflare.com.'},
'247support-number.com.',
'promoocodes.com.',
'myassignmenthelp.co.uk.',
'socialmonkee.com.',
'aapkeaajanese.website.',
'healthymum.org.',
'escortdomain.net.',
])
def watched_ns_for_url_domain(s, site):
return ns_for_url_domain(s, site, [
# Don't forget the trailing dot on the resolved name here either!
# {'dns1.namecheaphosting.com.', 'dns2.namecheaphosting.com.'},
# {'dns11.namecheaphosting.com.', 'dns12.namecheaphosting.com.'},
'namecheaphosting.com.', # has FPs, don't blacklist again
# 'domaincontrol.com.',
# {'dns1.registrar-servers.com.', 'dns2.registrar-servers.com.'},
{'adi.ns.cloudflare.com.', 'miles.ns.cloudflare.com.'},
{'aida.ns.cloudflare.com.', 'lloyd.ns.cloudflare.com.'},
{'ajay.ns.cloudflare.com.', 'lia.ns.cloudflare.com.'},
{'betty.ns.cloudflare.com.', 'kai.ns.cloudflare.com.'},
{'bonnie.ns.cloudflare.com.', 'guss.ns.cloudflare.com.'},
{'chip.ns.cloudflare.com.', 'cruz.ns.cloudflare.com.'},
{'chris.ns.cloudflare.com.', 'tess.ns.cloudflare.com.'},
{'dana.ns.cloudflare.com.', 'piotr.ns.cloudflare.com.'},
{'duke.ns.cloudflare.com.', 'lola.ns.cloudflare.com.'},
{'ernest.ns.cloudflare.com.', 'pat.ns.cloudflare.com.'},
{'greg.ns.cloudflare.com.', 'kia.ns.cloudflare.com.'},
{'jay.ns.cloudflare.com.', 'jule.ns.cloudflare.com.'},
{'kia.ns.cloudflare.com.', 'noah.cs.cloudflare.com.'},
{'mark.ns.cloudflare.com.', 'wanda.ns.cloudflare.com.'},
{'naomi.ns.cloudflare.com.', 'tim.ns.cloudflare.com.'},
{'norm.ns.cloudflare.com.', 'olga.ns.cloudflare.com.'},
{'pablo.ns.cloudflare.com.', 'pola.ns.cloudflare.com.'},
'mihanwebhost.com.', # FPs, don't blacklist
'offshoreracks.com.',
'sathyats.net.',
'shared-host.org.',
'web.com.ph.',
{'ns09.domaincontrol.com.', 'ns10.domaincontrol.com.'}, # FPs, don't blacklist
{'ns43.domaincontrol.com.', 'ns44.domaincontrol.com.'}, # FPs, don't blacklist
'supercloudapps.com.',
'vultr.com.', # has FPs, don't move to blacklist
'directory92.com.',
'offshoric.com.',
'freehostia.com.',
'hawkhost.com.', # has FPs, don't move to blacklist
'greengeeks.com.',
'supportaus.com.',
'utecho.com.',
'syrahost.com.',
'256gbserver.com.',
'solutionsinfini.org.',
'dnsdomen.com.',
'ownmyserver.com.',
'websitewelcome.com.',
'fatcow.com.',
'vedigitize.us.',
'serverpars.com.',
])
# noinspection PyUnusedLocal,PyMissingTypeHints
def is_offensive_post(s, site):
if not s:
return False, ""
offensive = regex.compile(
r"(?is)\b((?:ur\Wm[ou]m|(yo)?u suck|[8B]={3,}[D>)]|nigg[aeu][rh]?|(ass\W?|a|a-)hole|(?:fur)?fa+g+(ot)?s?\b|"
r"daf[au][qk]|(?<!brain)(mother|mutha)?f\W*u\W*c?\W*k+(a|ing?|e?[rd]| *off+| *(you|ye|u)(rself)?|"
r" u+|tard)?|(bul+)?shit(t?er|head)?|(yo)?u(r|'?re)? (gay|scum)|dickhead|"
r"pedo(?!bapt|dont|log|mete?r|troph)|cocksuck(e?[rd])?|"
r"whore|cunt|jerk(ing)?\W?off|cumm(y|ie)|butthurt|queef|lesbo|"
r"bitche?|(eat|suck|throbbing|sw[oe]ll(en|ing)?)\b.{0,20}\b(cock|dick)|dee[sz]e? nut[sz]|"
r"dumb\W?ass|wet\W?puss(y|ie)?|slut+y?|shot\W?my\W?(hot\W?)?load)s?)\b")
matches = offensive.finditer(s)
len_of_match = 0
text_matched = []
for match in matches:
len_of_match += match.end() - match.start()
text_matched.append(match.group(0))
if len_of_match / len(s) >= 0.015: # currently at 1.5%, this can change if it needs to
return True, "Offensive keyword{}: *{}*".format("s" if len(text_matched) > 1 else "", ", ".join(text_matched))
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def has_eltima(s, site):
reg = regex.compile(r"(?is)\beltima")
if reg.search(s) and len(s) <= 750:
return True, u"Bad keyword *eltima* and body length under 750 chars"
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints,PyTypeChecker
def username_similar_website(post):
s, username = post.body, post.user_name
sim_ratio, sim_webs = perform_similarity_checks(s, username)
if sim_ratio >= SIMILAR_THRESHOLD:
return False, False, True, "Username `{}` similar to {}, ratio={}".format(
username,
', '.join(['*{}* at position {}-{}'.format(w, s.index(w), s.index(w) + len(w)) for w in sim_webs]),
sim_ratio)
else:
return False, False, False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints,PyTypeChecker
def character_utilization_ratio(s, site):
s = strip_urls_and_tags(s)
counter = Counter(s)
total_chars = len(s)
highest_ratio = 0.0
highest_char = ""
for key, value in counter.items():
char_ratio = value / float(total_chars)
# key, value, char_ratio
if char_ratio > highest_ratio:
highest_ratio = char_ratio
highest_char = key
if highest_ratio > CHARACTER_USE_RATIO:
return True, "The `{}` character appears in a high percentage of the post".format(highest_char)
else:
return False, ""
def post_links(post):
"""
Helper function to extract URLs from a piece of HTML.
"""
# Fix stupid spammer tricks
for p in COMMON_MALFORMED_PROTOCOLS:
post = post.replace(p[0], p[1])
links = []
for l in regex.findall(URL_REGEX, post):
if l[-1].isalnum():
links.append(l)
else:
links.append(l[:-1])
return set(links)
# noinspection PyMissingTypeHints
def perform_similarity_checks(post, name):
"""
Performs 4 tests to determine similarity between links in the post and the user name
:param post: Test of the post
:param name: Username to compare against
:return: Float ratio of similarity
"""
max_similarity, similar_links = 0.0, []
# Keep checking links until one is deemed "similar"
for link in post_links(post):
domain = get_domain(link)
# Straight comparison
s1 = similar_ratio(domain, name)
# Strip all spaces
s2 = similar_ratio(domain, name.replace(" ", ""))
# Strip all hyphens
s3 = similar_ratio(domain.replace("-", ""), name.replace("-", ""))
# Strip all hyphens and all spaces
s4 = similar_ratio(domain.replace("-", "").replace(" ", ""), name.replace("-", "").replace(" ", ""))
similarity = max(s1, s2, s3, s4)
max_similarity = max(max_similarity, similarity)
if similarity >= SIMILAR_THRESHOLD:
similar_links.append(domain)
return max_similarity, similar_links
# noinspection PyMissingTypeHints
def similar_ratio(a, b):
return SequenceMatcher(None, a.lower(), b.lower()).ratio()
# noinspection PyMissingTypeHints
def get_domain(s, full=False):
"""
Extract the domain name; with full=True, keep the TLD tacked on.
"""
try:
extract = tld.get_tld(s, fix_protocol=True, as_object=True, )
if full:
domain = extract.fld
else:
domain = extract.domain
except TldDomainNotFound as e:
invalid_tld = RE_COMPILE.match(str(e)).group(1)
# Attempt to replace the invalid protocol
s1 = s.replace(invalid_tld, 'http', 1)
try:
extract = tld.get_tld(s1, fix_protocol=True, as_object=True, )
if full:
domain = extract.fld
else:
domain = extract.domain
except TldDomainNotFound:
# Assume bad TLD and try one last fall back, just strip the trailing TLD and leading subdomain
parsed_uri = urlparse(s)
if len(parsed_uri.path.split(".")) >= 3:
if full:
domain = '.'.join(parsed_uri.path.split(".")[1:])
else:
domain = parsed_uri.path.split(".")[1]
else:
if full:
domain = parsed_uri.path
else:
domain = parsed_uri.path.split(".")[0]
return domain
# noinspection PyMissingTypeHints
def similar_answer(post):
if not post.parent:
return False, False, False, ""
question = post.parent
sanitized_body = strip_urls_and_tags(post.body)
for other_answer in question.answers:
if other_answer.post_id != post.post_id:
sanitized_answer = strip_urls_and_tags(other_answer.body)
ratio = similar_ratio(sanitized_body, sanitized_answer)
if ratio >= SIMILAR_ANSWER_THRESHOLD:
return False, False, True, \
u"Answer similar to answer {}, ratio {}".format(other_answer.post_id, ratio)
return False, False, False, ""
# noinspection PyMissingTypeHints
def strip_urls_and_tags(s):
return regex.sub(URL_REGEX, "", regex.sub(r"</?[^>]+>|\w+://", "", s))
# noinspection PyUnusedLocal,PyMissingTypeHints
def mostly_dots(s, site):
if not s:
return False, ""
# Strip code blocks here rather than with `stripcodeblocks` so we get the length of the whole post
body = regex.sub(r"(?s)<pre([\w=\" -]*)?>.*?</pre>", "", s)
body = regex.sub(r"(?s)<code>.*?</code>", "", body)
body = strip_urls_and_tags(body)
s = strip_urls_and_tags(s)
if not s:
return False, ""
dot_count = body.count(".")
if dot_count / len(s) >= 0.4:
return True, u"Post contains {} dots out of {} characters".format(dot_count, len(s))
else:
return False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def mostly_punctuations(s, site):
body = regex.sub(r"(?s)<pre([\w=\" -]*)?>.*?</pre>", "", s)
body = regex.sub(r"(?s)<code>.*?</code>", "", body)
body = strip_urls_and_tags(body)
s = strip_urls_and_tags(s)
if len(s) < 15:
return False, ""
punct_re = regex.compile(r"[[:punct:]]")
all_punc = punct_re.findall(body.replace(".", ""))
count = max([all_punc.count(punc) for punc in set(all_punc)]) if all_punc else 0
frequency = count / len(s)
if frequency >= PUNCTUATION_RATIO:
return True, u"Post contains {} punctuation marks out of {} characters".format(count, len(s))
else:
return False, ""
def toxic_check(post):
s = strip_urls_and_tags(post.body)[:3000]
if not s:
return False, False, False, ""
try:
response = requests.post(PERSPECTIVE, json={
"comment": {
"text": s
},
"requestedAttributes": {
"TOXICITY": {
"scoreType": "PROBABILITY"
}
}
}).json()
except requests.exceptions.ConnectionError:
return False, False, False, ""
if "error" in response:
err_msg = response["error"]["message"]
if not err_msg.startswith("Attribute TOXICITY does not support request languages:"):
log("debug", "Perspective error: {} for string {} (original body {})".format(err_msg, s, post.body))
else:
probability = response["attributeScores"]["TOXICITY"]["summaryScore"]["value"]
if probability > PERSPECTIVE_THRESHOLD:
return False, False, True, "Perspective scored {}".format(probability)
return False, False, False, ""
# noinspection PyUnusedLocal,PyMissingTypeHints
def body_starts_with_title(post):
t = post.title.strip().replace(" ", "")
# Safeguard for answers, should never hit
if len(t) <= 10:
log('warning', "Length of post title is 10 characters or less. This is highly abnormal")
return False, False, False, ""
end_in_url, ending_url = link_at_end(post.body, None)
if not end_in_url:
return False, False, False, ""
ending_url = ending_url.replace("Link at end: ", "")
s = strip_urls_and_tags(post.body).replace(" ", "").replace("\n", "")
if similar_ratio(s[:len(t)], t) >= BODY_TITLE_SIMILAR_RATIO:
return False, False, True, "Body starts with title and ends in URL: " + ending_url
# Strip links and link text
s = regex.sub(r"<a[^>]+>[^<>]*</a>", "", regex.sub(r">>+", "", post.body))
s = strip_urls_and_tags(s).replace(" ", "").replace("\n", "")
if similar_ratio(s[:len(t)], t) >= BODY_TITLE_SIMILAR_RATIO:
return False, False, True, "Body starts with title and ends in URL: " + ending_url
# Final check: Body contains title verbatim
if t in strip_urls_and_tags(post.body).replace(" ", "").replace("\n", ""):
return False, False, True, "Body contains title and ends in URL: " + ending_url
return False, False, False, ""
def luncheon_meat(s, site): # Random "signature" like asdfghjkl
s = regex.search(r"<p>\s*?(\S{8,})\s*?</p>$", s.lower())
if not s:
return False, ""
has_letter = regex.search(r"[A-Za-z]", s[1])
if not has_letter:
return False, ""
p1 = ENGLISH_PRIOR
p2 = UNIFORM_PRIOR
for symbol in s[1]:
if symbol in ENGLISH:
p1 += ENGLISH[symbol]
p2 += UNIFORM
return p2 > p1, "match: {}, p1: {}, p2: {}".format(s[1], p1, p2)
def turkey2(post):
if regex.search("([01]{8}|zoe)", post.body):
pingable = chatcommunicate._clients["stackexchange.com"].get_room(11540).get_pingable_user_names()
if not pingable or not isinstance(pingable, list):
return False, False, False, ""
if post.user_name in pingable:
return False, False, True, "Himalayan pink salt detected"
return False, False, False, ""
# FLEE WHILE YOU STILL CAN.
def religion_troll(s, site):
regexes = [
r'(?:(?:Rubellite\W*(?:Fae|Yaksi)|Sarvabhouma|Rohit|Anisha|Ahmed\W*Bkhaty|Anubhav\W*Jha|Vineet\W*Aggarwal|Josh'
r'\W*K|Doniel\W*F|mbloch)(?:|\b.{1,200}?\b)(?:(?:mother)?[sf]uck|pis+|pus+(?:y|ies)|boo+b|tit|coc*k|dick|ass'
r'(?:hole|lick)?|penis|phallus|vagina|cunt(?:bag)?|genital|rectum|dog+ystyle|blowjob|scum(?:bag)|bitch|bastard'
r'|slut+(?:ish|y)?|harlot|whore|bloody|rotten|diseased|swelling|lesbian|queer|(?:homo|trans|bi)(?:sexual)?|'
r'retard|jew(?:ish)?|nig+er|nic+a|fag(?:g?[eio]t)?|heretic|idiot|sinners|raghead|muslim|hindu)(?:(?:ing|e[dr]'
r'|e?)[sz]?)?|(?:(?:mother)?[sf]uck|pis+|pus+(?:y|ies)|boo+b|tit|coc*k|dick|ass(?:hole|lick)?|penis|phallus|'
r'vagina|cunt(?:bag)?|genital|rectum|dog+ystyle|blowjob|scum(?:bag)|bitch|bastard|slut+(?:ish|y)?|harlot|whore'
r'|bloody|rotten|diseased|swelling|lesbian|queer|(?:homo|trans|bi)(?:sexual)?|retard|jew(?:ish)?|nig+er|nic+a|'
r'fag(?:g?[eio]t)?|heretic|idiot|sinners|raghead|muslim|hindu)(?:(?:ing|e[dr]|e?)[sz]?)?(?:|\b.{1,200}?\b)'
r'(?:Rubellite\W*(?:Fae|Yaksi)|Sarvabhouma|Rohit|Anisha|Ahmed\W*Bkhaty|Anubhav\W*Jha|Vineet\W*Aggarwal|Josh'
r'\W*K|Doniel\W*F|mbloch))',
r'(?:[hl]indu(?:s|ism)?|jew(?:s|ish)?|juda(?:ic|ism)|kike)(?:\W*(?:(?:chew|kiss|bunch|are|is|n|it|is|and|of|an?'
r'|the|greedy|bag|with|jewdaism)\W*)*(?:(?:worse\W*than(?:(?:\W*and)?\W*(?:aids|cancer|tuberculosis|syphilis|'
r'tumor|internal\W*bleeding)+)+)|(?:damned|going)\W*to\W*(?:rot\W*in\W*)?hell|stinking|filthy|dirty|bloody|'
r'(?:saggy)?\W*ass\W*?(?:hole|licker|e)?s?|w?hores?|bitch(?:es)?|idiots?|gandus?|homo(?:sexual)?s?|(?:fag|mag)'
r'(?:g?[eio]t)?s?|morons?|(circumcised\W*)?bastards?|jhandus?|(?:mother\W*?)?fuck(?:s|ing|ers)?|sucks?|'
r'cockroach(?:es)?|excreta|need\W*to\W*be\W*(?:exterminated|castrated)|puss(?:y|ies)(?:\W*licking)|hairy|'
r'rotten|cock\W*(?:slurping|suck(?:ers?|ing)?)|blood\W*?suck(?:ing|ers?)|parasites?|swines?(?:\W*piss)?|'
r'(?:scum|cunt)\W*bags?|rotten|corpses|slurping|mutilated\W*genitals?|whoremonger|rodents?|pests?|demonic|evil|'
r'uncivilized|barbaric|racist|satanic|savage|dead\W*swines)+)+',
r'(?:(?:stupid|bloody|gandus?|lindus?|w?hore(?:monger(?:s|ing)|s)?|impotent|dumb|(?:mother\W*?)?fuck(?:ing|ers)'
r'?|assholes?(?:\W*?of\W*?)?|bitch(?:es)?|dirty|stinking|filthy|blood\W*?(?:thirsty|sucking)|racist|dumb|pussy)'
r'\W*?(?:and|is|you|flesh|of|a)*\W*?)+(?:[hl]indu(?:s|ism)?|jew(?:s|ish)?|juda(?:ism|ic)|kike)',
r'rama?\W*?(?:(?:was|is|a|an)\W*?)*\W*?(?:bastard|impotent|asshole|(?:mother\W*?)?fuck(?:ing|ers|s)|'
r'(?:fag|mag)(?:g?[eio]t)?s?)+',
r'(?:\b|\w*)(?:o(?:[^A-Za-z]|&#?\w+;)*)(?:d(?:[^A-Za-z]|&#?\w+;)*)(?:u(?:[^A-Za-z]|&#?\w+;)*)(?:d(?:[^A-Za-z]|'
r'&#?\w+;)*)(?:u(?:[^A-Za-z]|&#?\w+;)*)(?:w(?:[^A-Za-z]|&#?\w+;)*)(?:a(?:[^A-Za-z]|&#?\w+;)*)\w*(?:\b|\w*)'
]
offensive = any(regex.search(x, s) for x in regexes)
return offensive, 'Potential religion site troll post' if offensive else ''
load_blacklists()
# noinspection PyClassHasNoInit
class FindSpam:
bad_keywords_nwb = [ # "nwb" == "no word boundary"
u"ಌ", "vashi?k[ae]r[ae]n", "babyli(ss|cious)", "garcinia", "cambogia", "acai ?berr",
"(eye|skin|aging) ?cream", "b ?a ?m ?((w ?o ?w)|(w ?a ?r))", "online ?it ?guru",
"abam26", "watch2live", "cogniq", "(serum|lift) ?eye", "tophealth", "poker[ -]?online",
"caralluma", r"male\Wperf(?!ormer)", "anti[- ]?aging", "lumisse", "(ultra|berry|body)[ -]?ketone",
"(cogni|oro)[ -]?(lift|plex)", "diabazole", "forskolin", "tonaderm", "luma(genex|lift)",
"(skin|face|eye)[- ]?(serum|therapy|hydration|tip|renewal|gel|lotion|cream)",
"(skin|eye)[- ]?lift", "(skin|herbal) ?care", "nuando[ -]?instant", "\\bnutra", "nitro[ -]?slim",
"aimee[ -]?cream", "slimatrex", "cosmitone", "smile[ -]?pro[ -]?direct", "bellavei", "opuderm",
r"contact (me|us)\W*<a ", "follicure", "kidney[ -]?bean[ -]?extract", "ecoflex",
r"\brsgold", "bellavei", "goji ?xtreme", "lumagenex", "ajkobeshoes", "kreatine",
"packers.{0,15}(movers|logistic).{0,25}</a>", "guaranteedprofitinvestment",
"(brain|breast|male|penile|penis)[- ]?(enhance|enlarge|improve|boost|plus|peak)",
"renuva(cell|derm)", " %uh ", " %ah ", "svelme", "tapsi ?sarkar", "viktminskning",
"unique(doc)?producers", "green ?tone ?pro", "troxyphen", "seremolyn", "revolyn",
"(?:networking|cisco|sas|hadoop|mapreduce|oracle|dba|php|sql|javascript|js|java|designing|marketing|"
"salesforce|joomla)( certification)? (courses?|training).{0,25}</a>",
r"(?:design|development|compan(y|ies)|training|courses?|automation)(\b.{1,8}\b)?\L<city>\b",
r"\b\L<city>(\b.{1,8}\b)?(?:tour)", # TODO: Populate this "after city" keyword list
u"Ｃ[Ｏ|0]Ｍ", "ecoflex", "no2factor", "no2blast", "sunergetic", "capilux", "sante ?avis",
"enduros", "dianabol", r"ICQ#?\d{4}-?\d{5}", "3073598075", "lumieres", "viarex", "revimax",
"celluria", "viatropin", "(meg|test)adrox", "nordic ?loan ?firm", r"safflower\Woil",
"(essay|resume|article|dissertation|thesis) ?writing ?service", "satta ?matka", r"b\W?o\W?j\W?i\W?t\W?e\W?r",
r"rams[ey]+\W?dave", "(🐽|🐷){3,}"
]