Removed ML from well known pattern (#448)

* removed extra keys * removed ml for well-known prefixes patterns * tests fixed * benchmark scores fix * benchmark scores fix 2
Samsung · Nov 1, 2023 · f76d16a · f76d16a
1 parent 3467170
commit f76d16a
Show file tree

Hide file tree

Showing 7 changed files with 293 additions and 172 deletions.
diff --git a/cicd/benchmark.txt b/cicd/benchmark.txt
@@ -10,16 +10,16 @@ Predefined Pattern                  326            2          40
 Private Key                        1001            1           3
 Seed, Salt, Nonce                    40            4           4
 TOTAL:                             5307        63688        5644
-Detected Credentials: 5993
-credsweeper result_cnt : 5337, lost_cnt : 0, true_cnt : 4439, false_cnt : 898
+Detected Credentials: 5997
+credsweeper result_cnt : 5339, lost_cnt : 0, true_cnt : 4441, false_cnt : 898
 Category                      TP    FP        TN    FN        FPR        FNR       ACC       PRC       RCL        F1
 --------------------------  ----  ----  --------  ----  ---------  ---------  --------  --------  --------  --------
 Authentication Key & Token    54     4        28    16  0.125      0.228571   0.803922  0.931034  0.771429  0.84375
 Generic Secret               973     3       215    83  0.0137615  0.0785985  0.932496  0.996926  0.921402  0.957677
-Generic Token                287     7       596    46  0.0116086  0.138138   0.943376  0.97619   0.861862  0.91547
+Generic Token                289     7       596    44  0.0116086  0.132132   0.945513  0.976351  0.867868  0.918919
 Other                        818   750     63395   258  0.0116923  0.239777   0.984545  0.521684  0.760223  0.618759
 Password                     995   130      4150   410  0.0303738  0.291815   0.905013  0.884444  0.708185  0.786561
 Predefined Pattern           309     2        40    17  0.0476191  0.0521472  0.94837   0.993569  0.947853  0.970173
 Private Key                  967     0         4    34             0.033966   0.966169  1         0.966034  0.982724
 Seed, Salt, Nonce             36     2         6     4  0.25       0.1        0.875     0.947368  0.9       0.923077
-                            4439   898  19428253   868  4.622e-05  0.163558   0.999909  0.831741  0.836442  0.834085
+                            4441   898  19428253   866  4.622e-05  0.163181   0.999909  0.831804  0.836819  0.834304
diff --git a/credsweeper/rules/config.yaml b/credsweeper/rules/config.yaml
@@ -199,7 +199,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>(ABIA|ACCA|AGPA|AIDA|AIPA|AKIA|ANPA|ANVA|AROA|APKA|ASCA|ASIA)[0-9A-Z]{16,17})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - A
   min_line_len: 20
@@ -212,7 +211,6 @@
     - (^|[^.0-9A-Za-z_/+-])(?P<value>(AKIA|ASIA)[0-9A-Z]{16,17})([^=0-9A-Za-z_/+-]|$)
     - (?P<value>[0-9a-zA-Z/+]{40})
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - AKIA
     - ASIA
@@ -224,7 +222,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>amzn\.mws\.[0-9a-z]{8}-[0-9a-z]{4}-[0-9a-z]{4}-[0-9a-z]{4}-[0-9a-z]{12})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - amzn
   min_line_len: 30
@@ -247,7 +244,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>dt0[a-zA-Z]{1}[0-9]{2}\.[A-Z0-9]{24}\.[A-Z0-9]{64})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - dt0
   min_line_len: 90
@@ -258,7 +254,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>EAAC[0-9A-Za-z]{27,})
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - EAAC
   min_line_len: 31
@@ -282,7 +277,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>AIza[0-9A-Za-z_-]{35})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   validations:
     - GoogleApiKeyValidation
   required_substrings:
@@ -296,7 +290,6 @@
     - (?P<value>[0-9]+\-[0-9A-Za-z_]{32}\.apps\.googleusercontent\.com)
     - (?<![0-9a-zA-Z_-])(?P<value>[0-9a-zA-Z_-]{24})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   validations:
     - GoogleMultiValidation
   required_substrings:
@@ -309,7 +302,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>ya29\.[0-9A-Za-z_-]{22,})
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - ya29.
   min_line_len: 27
@@ -320,7 +312,6 @@
   values:
     - (?i)(?P<value>heroku(.{0,20})?[0-9a-f]{8}(-[0-9a-f]{4})+-[0-9a-f]{12})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - heroku
   min_line_len: 24
@@ -331,7 +322,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>IGQVJ[\w]{100,})
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - IGQVJ
   min_line_len: 105
@@ -353,7 +343,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>[0-9a-zA-Z]{32}-us[0-9]{1,2})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   validations:
     - MailChimpKeyValidation
   required_substrings:
@@ -366,7 +355,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>key-[0-9a-zA-Z]{32})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - key-
   min_line_len: 36
@@ -390,7 +378,6 @@
   values:
     - (?P<value>access_token\$production\$[0-9a-z]{16}\$[0-9a-z]{32})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   required_substrings:
     - access_token$production$
   min_line_len: 72
@@ -410,7 +397,6 @@
   values:
     - (?P<value>sk_live_[0-9a-z]{32})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   required_substrings:
     - sk_live_
   min_line_len: 40
@@ -433,7 +419,6 @@
   values:
     - (?P<value>SG\.[\w_]{16,32}\.[\w_]{16,64})
   filter_type: GeneralPattern
-  use_ml: false
   required_substrings:
     - SG.
   min_line_len: 34
@@ -454,7 +439,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>xox[a|b|p|r|o|s]\-[-a-zA-Z0-9]{10,250})
   filter_type: GeneralPattern
-  use_ml: true
   validations:
     - SlackTokenValidation
   required_substrings:
@@ -467,7 +451,6 @@
   values:
     - (?P<value>hooks\.slack\.com/services/T\w{8}/B\w{8}/\w{24})
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - hooks.slack.com/services/T
   min_line_len: 61
@@ -478,7 +461,6 @@
   values:
     - (?P<value>sk_live_[0-9a-zA-Z]{24})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   validations:
     - StripeApiKeyValidation
   required_substrings:
@@ -491,7 +473,6 @@
   values:
     - (?P<value>rk_live_[0-9a-zA-Z]{24})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - rk_live_
   min_line_len: 32
@@ -502,7 +483,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>EAAA[0-9A-Za-z_-]{60})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   validations:
     - SquareAccessTokenValidation
   required_substrings:
@@ -515,7 +495,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>sq0[a-z]{3}-[0-9A-Za-z_-]{22})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   validations:
     - SquareClientIdValidation
   required_substrings:
@@ -528,7 +507,6 @@
   values:
     - (?P<value>sq0csp-[0-9A-Za-z_-]{43})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: false
   required_substrings:
     - sq0csp
   min_line_len: 50
@@ -551,7 +529,6 @@
   values:
     - (^|[^.0-9A-Za-z_/+-])(?P<value>SK[0-9a-fA-F]{32})([^=0-9A-Za-z_/+-]|$)
   filter_type: GeneralPattern
-  use_ml: true
   required_substrings:
     - SK
   min_line_len: 34

diff --git a/tests/__init__.py b/tests/__init__.py
@@ -8,7 +8,7 @@
 SAMPLES_CRED_LINE_COUNT: int = 402
 
 # credentials count after post-processing
-SAMPLES_POST_CRED_COUNT: int = 293
+SAMPLES_POST_CRED_COUNT: int = 296
 
 # with option --doc
 SAMPLES_IN_DOC = 422