SourceType applied to unify usage 'doc' and 'src'

credsweeper/__main__.py

@@ -8,7 +8,7 @@
 
 from credsweeper import __version__
 from credsweeper.app import APP_PATH, CredSweeper
-from credsweeper.common.constants import ThresholdPreset, Severity, RuleType, DiffRowType
+from credsweeper.common.constants import ThresholdPreset, Severity, RuleType, DiffRowType, SourceType
 from credsweeper.file_handler.files_provider import FilesProvider
 from credsweeper.file_handler.patch_provider import PatchProvider
 from credsweeper.file_handler.text_provider import TextProvider
@@ -158,8 +158,8 @@ def get_arguments() -> Namespace:
                         default=0,
                         required=False,
                         metavar="POSITIVE_INT")
-    parser.add_argument("--src", help="sources-specific scanning", dest="src", action="store_true")
-    parser.add_argument("--doc", help="document-specific scanning", dest="doc", action="store_true")
+    parser.add_argument("--src", help="sources-specific scanning", dest=SourceType.SRC.value, action="store_true")
+    parser.add_argument("--doc", help="document-specific scanning", dest=SourceType.DOC.value, action="store_true")
     parser.add_argument("--ml_threshold",
                         help="setup threshold for the ml model. "
                         "The lower the threshold - the more credentials will be reported. "
@@ -270,9 +270,9 @@ def scan(args: Namespace, content_provider: FilesProvider, json_filename: Option
 
         usage_list: List[str] = []
         if args.src:
-            usage_list.append("src")
+            usage_list.append(SourceType.SRC.value)
         if args.doc:
-            usage_list.append("doc")
+            usage_list.append(SourceType.DOC.value)
 
         credsweeper = CredSweeper(rule_path=args.rule_path,
                                   config_path=args.config_path,

credsweeper/app.py

@@ -11,7 +11,7 @@
 # Directory of credsweeper sources MUST be placed before imports to avoid circular import error
 APP_PATH = Path(__file__).resolve().parent
 
-from credsweeper.common.constants import KeyValidationOption, Severity, ThresholdPreset
+from credsweeper.common.constants import KeyValidationOption, Severity, ThresholdPreset, SourceType
 from credsweeper.config import Config
 from credsweeper.credentials import Candidate, CredentialManager
 from credsweeper.deep_scanner.deep_scanner import DeepScanner
@@ -90,7 +90,7 @@ def __init__(self,
                                             use_filters=use_filters,
                                             find_by_ext=find_by_ext,
                                             depth=depth,
-                                            usage_list=usage_list or ["src"],
+                                            usage_list=usage_list or [SourceType.SRC.value],
                                             severity=severity,
                                             size_limit=size_limit,
                                             exclude_lines=exclude_lines,
@@ -303,7 +303,7 @@ def file_scan(self, content_provider: Union[DiffContentProvider, TextContentProv
             candidates.append(dummy_candidate)
 
         else:
-            if self.config.depth or "doc" in self.config.usage_list:
+            if self.config.depth or SourceType.DOC.value in self.config.usage_list:
                 # deep scan with possible data representation
                 candidates = self.deep_scanner.scan(content_provider, self.config.depth, self.config.size_limit)
             else:

credsweeper/common/constants.py

@@ -52,6 +52,12 @@ def get(severity: Union[str, "Severity"]) -> Optional["Severity"]:
         return None
 
 
+class SourceType(Enum):
+    """Stores types of sources to configure rules and filters"""
+    SRC = "src"
+    DOC = "doc"
+
+
 class Base(Enum):
     """Stores types of character sets in lower case"""
     base36 = "base36"

credsweeper/filters/group/group.py

@@ -1,7 +1,7 @@
 from abc import ABC
 from typing import List
 
-from credsweeper.common.constants import GroupType
+from credsweeper.common.constants import GroupType, SourceType
 from credsweeper.config import Config
 from credsweeper.filters import (Filter, LineSpecificKeyCheck, SeparatorUnusualCheck, ValueAllowlistCheck,
                                  ValueArrayDictionaryCheck, ValueBlocklistCheck, ValueCamelCaseCheck,
@@ -51,9 +51,9 @@ def get_keyword_base_filters(config: Config) -> List[Filter]:
             ValueTokenCheck(),
             VariableNotAllowedPatternCheck(),
         ]
-        if "src" in config.usage_list:
+        if SourceType.SRC.value in config.usage_list:
             filters.extend([ValuePatternCheck(config), ValueNotAllowedPatternCheck()])
-        if "doc" in config.usage_list:
+        if SourceType.DOC.value in config.usage_list:
             filters.append(ValueMaskedCheck())
         return filters
 

tests/conftest.py

@@ -5,8 +5,8 @@
 import pytest
 
 from credsweeper.app import APP_PATH
+from credsweeper.common.constants import Severity, SourceType
 from credsweeper.config import Config
-from credsweeper.common.constants import Severity
 from credsweeper.rules import Rule
 from credsweeper.scanner import Scanner
 from credsweeper.utils import Util
@@ -39,7 +39,7 @@ def config() -> Config:
     config_dict["use_filters"] = True
     config_dict["find_by_ext"] = False
     config_dict["depth"] = 0
-    config_dict["usage_list"] = ["src"]
+    config_dict["usage_list"] = [SourceType.SRC.value]
     config_dict["find_by_ext_list"] = [".txt", ".inf"]
     config_dict["size_limit"] = None
     config_dict["min_keyword_value_length"] = 4

tests/data/__init__.py

@@ -1,5 +1,6 @@
 from typing import Dict, Any, List
 
+from credsweeper.common.constants import SourceType
 from tests import SAMPLES_POST_CRED_COUNT, SAMPLES_IN_DEEP_3, SAMPLES_CRED_COUNT, SAMPLES_ML_DOC
 
 DATA_TEST_CFG: List[Dict[str, Any]] = [{
@@ -10,7 +11,7 @@
     "__cred_count": SAMPLES_ML_DOC,
     "sort_output": True,
     "json_filename": "doc.json",
-    "usage_list": ["doc"]
+    "usage_list": [SourceType.DOC.value]
 }, {
     "__cred_count": SAMPLES_IN_DEEP_3,
     "sort_output": True,

tests/filters/test_value_array_dictionary_check.py

@@ -1,6 +1,6 @@
 import pytest
 
-from credsweeper.file_handler.analysis_target import AnalysisTarget
+from credsweeper.common.constants import SourceType
 from credsweeper.filters import ValueArrayDictionaryCheck, VariableNotAllowedPatternCheck
 from credsweeper.rules import Rule
 from tests.filters.conftest import LINE_VALUE_PATTERN, DUMMY_ANALYSIS_TARGET
@@ -19,7 +19,7 @@ def token_rule(self, config) -> Rule:
             "filter_type": [VariableNotAllowedPatternCheck.__name__],
             "use_ml": True,
             "min_line_len": 0,
-            "usage_list": ["src", "doc"],
+            "usage_list": [SourceType.SRC.value, SourceType.DOC.value],
             "validations": []
         }
         rule = Rule(config, token_rule_without_filters)

tests/filters/test_value_similarity_check.py

@@ -1,5 +1,6 @@
 import pytest
 
+from credsweeper.common.constants import SourceType
 from credsweeper.filters import ValueSimilarityCheck, VariableNotAllowedPatternCheck
 from credsweeper.rules import Rule
 from tests.filters.conftest import DUMMY_ANALYSIS_TARGET
@@ -17,7 +18,7 @@ def password_rule(self, config) -> Rule:
             "values": ["password|passwd|pwd"],
             "filter_type": [VariableNotAllowedPatternCheck.__name__],
             "use_ml": True,
-            "usage_list": ["src", "doc"],
+            "usage_list": [SourceType.SRC.value, SourceType.DOC.value],
             "min_line_len": 0,
             "validations": []
         }

tests/ml_model/test_ml_validator.py

@@ -1,5 +1,6 @@
 from credsweeper import ThresholdPreset
 from credsweeper.app import APP_PATH
+from credsweeper.common.constants import SourceType
 
 from credsweeper.config import Config
 from credsweeper.credentials import Candidate
@@ -17,7 +18,7 @@ def test_ml_validator_simple_p():
     config_dict["use_filters"] = True
     config_dict["find_by_ext"] = False
     config_dict["depth"] = 0
-    config_dict["usage_list"] = ["src"]
+    config_dict["usage_list"] = [SourceType.SRC.value]
     config_dict["find_by_ext_list"] = []
     config_dict["size_limit"] = None
     config = Config(config_dict)

tests/rules/test_rule.py

@@ -3,7 +3,7 @@
 
 import pytest
 
-from credsweeper.common.constants import Severity
+from credsweeper.common.constants import Severity, SourceType
 from credsweeper.config import Config
 from credsweeper.filters.group import GeneralPattern
 from credsweeper.rules import Rule
@@ -22,7 +22,7 @@ class TestRuleConfigParsing:
             "min_line_len": 32,
             "use_ml": False,
             "validations": [],
-            "usage_list": ["src", "doc"]
+            "usage_list": [SourceType.SRC.value, SourceType.DOC.value]
         },
         # Check proper config with no validations
         {
@@ -33,7 +33,7 @@ class TestRuleConfigParsing:
             "filter_type": GeneralPattern.__name__,
             "min_line_len": 32,
             "use_ml": False,
-            "usage_list": ["src", "doc"]
+            "usage_list": [SourceType.SRC.value, SourceType.DOC.value]
         },
     ])
     def rule_config(self, request: str) -> Any:

tests/scanner/scan_type/test_multipattern.py

@@ -1,9 +1,9 @@
 import random
 import string
 import unittest
-from unittest.mock import patch, MagicMock
+from unittest.mock import MagicMock
 
-from credsweeper.common.constants import MAX_LINE_LENGTH
+from credsweeper.common.constants import MAX_LINE_LENGTH, SourceType
 from credsweeper.config import Config
 from credsweeper.file_handler.analysis_target import AnalysisTarget
 from credsweeper.rules import Rule
@@ -22,7 +22,7 @@ def setUp(self) -> None:
                 "name": "MULTI_PATTERN_RULE",
                 "severity": "info",
                 "type": "pattern",
-                "usage_list": ["src"],
+                "usage_list": [SourceType.SRC.value],
                 "values": ["a", "b"],
                 "filter_type": [],
                 "min_line_len": 0,

tests/test_main.py

@@ -21,7 +21,7 @@
 from credsweeper.__main__ import EXIT_FAILURE, EXIT_SUCCESS
 from credsweeper.app import APP_PATH
 from credsweeper.app import CredSweeper
-from credsweeper.common.constants import ThresholdPreset, Severity
+from credsweeper.common.constants import ThresholdPreset, Severity, SourceType
 from credsweeper.credentials import Candidate
 from credsweeper.file_handler.files_provider import FilesProvider
 from credsweeper.file_handler.text_content_provider import TextContentProvider
@@ -696,7 +696,7 @@ def test_credit_card_number_n(self) -> None:
 
     def test_doc_p(self) -> None:
         content_provider: FilesProvider = TextProvider([SAMPLES_PATH / "test.html"])
-        cred_sweeper = CredSweeper(usage_list=["doc"])
+        cred_sweeper = CredSweeper(usage_list=[SourceType.DOC.value])
         cred_sweeper.run(content_provider=content_provider)
         found_credentials = cred_sweeper.credential_manager.get_credentials()
         expected_credential_lines = [

tests/test_utils/dummy_line_data.py

@@ -1,6 +1,7 @@
 import re
 
 from credsweeper.app import APP_PATH
+from credsweeper.common.constants import SourceType
 from credsweeper.config import Config
 from credsweeper.credentials import LineData
 from credsweeper.utils import Util
@@ -14,7 +15,7 @@ def config() -> Config:
     config_dict["use_filters"] = True
     config_dict["find_by_ext"] = False
     config_dict["depth"] = 0
-    config_dict["usage_list"] = ["src"]
+    config_dict["usage_list"] = [SourceType.SRC.value]
     config_dict["size_limit"] = None
     return Config(config_dict)