From 84a33ff84e9e652fc9910648aac0ac47a40234fe Mon Sep 17 00:00:00 2001
From: XIANJUN ZHU <xianjun@us.ibm.com>
Date: Thu, 10 Oct 2019 16:31:44 -0400
Subject: [PATCH] Use RegexBasedDetector.assign_regex_generator (#215)

* feat: use assign regex in cloudant

* feat: use assign regex in db2

* feat: use assign regex in gh

* feat: use assign regex in iam

* feat: use assign regex in sl

* address comments

* address comments

* address comments
---
 cloudant.py      | 27 +++++-----------
 cloudant_test.py | 81 ++++++++++++++++++++++++------------------------
 2 files changed, 47 insertions(+), 61 deletions(-)

diff --git a/cloudant.py b/cloudant.py
index 41a3143dd..2525ab995 100644
--- a/cloudant.py
+++ b/cloudant.py
@@ -69,7 +69,7 @@ class CloudantDetector(RegexBasedDetector):
 
     def verify(self, token, content, potential_secret=None):
 
-        hosts = get_host(content)
+        hosts = find_host(content)
         if not hosts:
             return VerifiedResult.UNVERIFIED
 
@@ -79,28 +79,15 @@ def verify(self, token, content, potential_secret=None):
         return VerifiedResult.VERIFIED_FALSE
 
 
-def get_host(content):
-
-    # opt means optional
-    opt_quote = r'(?:"|\'|)'
-    opt_cl = r'(?:cloudant|cl|)'
-    opt_dash_undrscr = r'(?:_|-|)'
+def find_host(content):
     opt_hostname_keyword = r'(?:hostname|host|username|id|user|userid|user-id|user-name|' \
         'name|user_id|user_name|uname)'
-    opt_space = r'(?: |)'
-    assignment = r'(?:\=|:|:=|=>)+'
     hostname = r'(\w(?:\w|_|-)+)'
-    regex = re.compile(
-        r'{opt_quote}{opt_cl}{opt_dash_undrscr}{opt_hostname_keyword}{opt_space}{opt_quote}'
-        '{assignment}{opt_space}{opt_quote}{hostname}{opt_quote}'.format(
-            opt_quote=opt_quote,
-            opt_cl=opt_cl,
-            opt_dash_undrscr=opt_dash_undrscr,
-            opt_hostname_keyword=opt_hostname_keyword,
-            opt_space=opt_space,
-            hostname=hostname,
-            assignment=assignment,
-        ), flags=re.IGNORECASE,
+
+    regex = RegexBasedDetector.assign_regex_generator(
+        prefix_regex=CloudantDetector.cl,
+        password_keyword_regex=opt_hostname_keyword,
+        password_regex=hostname,
     )
 
     return [
diff --git a/cloudant_test.py b/cloudant_test.py
index d815f71e5..0b88c1584 100644
--- a/cloudant_test.py
+++ b/cloudant_test.py
@@ -8,7 +8,7 @@
 from detect_secrets.core.constants import VerifiedResult
 from detect_secrets.core.potential_secret import PotentialSecret
 from detect_secrets.plugins.cloudant import CloudantDetector
-from detect_secrets.plugins.cloudant import get_host
+from detect_secrets.plugins.cloudant import find_host
 
 CL_HOST = 'testy_test'  # also called user
 # only detecting 64 hex CL generated password
@@ -113,50 +113,49 @@ def test_verify_no_secret(self):
             'no_un={}'.format(CL_HOST),
         ) == VerifiedResult.UNVERIFIED
 
-
-@pytest.mark.parametrize(
-    'content, expected_output',
-    (
+    @pytest.mark.parametrize(
+        'content, expected_output',
         (
-            textwrap.dedent("""
-                --cloudant-hostname = {}
-            """)[1:-1].format(
-                CL_HOST,
+            (
+                textwrap.dedent("""
+                    --cloudant-hostname = {}
+                """)[1:-1].format(
+                    CL_HOST,
+                ),
+                [CL_HOST],
             ),
-            [CL_HOST],
-        ),
 
-        # With quotes
-        (
-            textwrap.dedent("""
-                cl_host = "{}"
-            """)[1:-1].format(
-                CL_HOST,
+            # With quotes
+            (
+                textwrap.dedent("""
+                    cl_host = "{}"
+                """)[1:-1].format(
+                    CL_HOST,
+                ),
+                [CL_HOST],
             ),
-            [CL_HOST],
-        ),
 
-        # multiple candidates
-        (
-            textwrap.dedent("""
-                cloudant_id = '{}'
-                cl-user = '{}'
-                CLOUDANT_USERID = '{}'
-                cloudant-uname: {}
-            """)[1:-1].format(
-                CL_HOST,
-                'test2_testy_test',
-                'test3-testy-testy',
-                'notanemail',
+            # multiple candidates
+            (
+                textwrap.dedent("""
+                    cloudant_id = '{}'
+                    cl-user = '{}'
+                    CLOUDANT_USERID = '{}'
+                    cloudant-uname: {}
+                """)[1:-1].format(
+                    CL_HOST,
+                    'test2_testy_test',
+                    'test3-testy-testy',
+                    'notanemail',
+                ),
+                [
+                    CL_HOST,
+                    'test2_testy_test',
+                    'test3-testy-testy',
+                    'notanemail',
+                ],
             ),
-            [
-                CL_HOST,
-                'test2_testy_test',
-                'test3-testy-testy',
-                'notanemail',
-            ],
         ),
-    ),
-)
-def test_get_host(content, expected_output):
-    assert get_host(content) == expected_output
+    )
+    def test_find_host(self, content, expected_output):
+        assert find_host(content) == expected_output