Move sanitize_value to be common

Fix Indian Passport Number regex

presidio-analyzer/presidio_analyzer/__init__.py

@@ -15,7 +15,6 @@
 from presidio_analyzer.analyzer_engine import AnalyzerEngine
 from presidio_analyzer.batch_analyzer_engine import BatchAnalyzerEngine
 from presidio_analyzer.analyzer_request import AnalyzerRequest
-from presidio_analyzer.analyzer_utils import PresidioAnalyzerUtils
 from presidio_analyzer.context_aware_enhancers import ContextAwareEnhancer
 from presidio_analyzer.context_aware_enhancers import LemmaContextAwareEnhancer
 from presidio_analyzer.analyzer_engine_provider import AnalyzerEngineProvider
@@ -51,6 +50,5 @@
     "ContextAwareEnhancer",
     "LemmaContextAwareEnhancer",
     "BatchAnalyzerEngine",
-    "PresidioAnalyzerUtils",
     "AnalyzerEngineProvider",
 ]

presidio-analyzer/presidio_analyzer/predefined_recognizers/aba_routing_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -59,7 +61,9 @@ def __init__(
         )
 
     def validate_result(self, pattern_text: str) -> bool:  # noqa D102
-        sanitized_value = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        sanitized_value = ValidationUtils.sanitize_value(
+            pattern_text, self.replacement_pairs
+        )
         return self.__checksum(sanitized_value)
 
     @staticmethod
@@ -68,9 +72,3 @@ def __checksum(sanitized_value: str) -> bool:
         for idx, m in enumerate([3, 7, 1, 3, 7, 1, 3, 7, 1]):
             s += int(sanitized_value[idx]) * m
         return s % 10 == 0
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/au_abn_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -72,7 +74,7 @@ def validate_result(self, pattern_text: str) -> bool:
         :return: A bool indicating whether the validation was successful.
         """
         # Pre-processing before validation checks
-        text = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        text = ValidationUtils.sanitize_value(pattern_text, self.replacement_pairs)
         abn_list = [int(digit) for digit in text if not digit.isspace()]
 
         # Set weights based on digit position
@@ -85,9 +87,3 @@ def validate_result(self, pattern_text: str) -> bool:
             sum_product += abn_list[i] * weight[i]
         remainder = sum_product % 89
         return remainder == 0
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/au_acn_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -69,7 +71,7 @@ def validate_result(self, pattern_text: str) -> bool:
         :return: A bool indicating whether the validation was successful.
         """
         # Pre-processing before validation checks
-        text = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        text = ValidationUtils.sanitize_value(pattern_text, self.replacement_pairs)
         acn_list = [int(digit) for digit in text if not digit.isspace()]
 
         # Set weights based on digit position
@@ -82,9 +84,3 @@ def validate_result(self, pattern_text: str) -> bool:
         remainder = sum_product % 10
         complement = 10 - remainder
         return complement == acn_list[-1]
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/au_medicare_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -69,7 +71,7 @@ def validate_result(self, pattern_text: str) -> bool:
         :return: A bool indicating whether the validation was successful.
         """
         # Pre-processing before validation checks
-        text = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        text = ValidationUtils.sanitize_value(pattern_text, self.replacement_pairs)
         medicare_list = [int(digit) for digit in text if not digit.isspace()]
 
         # Set weights based on digit position
@@ -81,9 +83,3 @@ def validate_result(self, pattern_text: str) -> bool:
             sum_product += medicare_list[i] * weight[i]
         remainder = sum_product % 10
         return remainder == medicare_list[8]
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/au_tfn_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -75,7 +77,7 @@ def validate_result(self, pattern_text: str) -> bool:
         :return: A bool indicating whether the validation was successful.
         """
         # Pre-processing before validation checks
-        text = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        text = ValidationUtils.sanitize_value(pattern_text, self.replacement_pairs)
         tfn_list = [int(digit) for digit in text if not digit.isspace()]
 
         # Set weights based on digit position
@@ -87,9 +89,3 @@ def validate_result(self, pattern_text: str) -> bool:
             sum_product += tfn_list[i] * weight[i]
         remainder = sum_product % 11
         return remainder == 0
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/credit_card_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -59,7 +61,9 @@ def __init__(
         )
 
     def validate_result(self, pattern_text: str) -> bool:  # noqa D102
-        sanitized_value = self.__sanitize_value(pattern_text, self.replacement_pairs)
+        sanitized_value = ValidationUtils.sanitize_value(
+            pattern_text, self.replacement_pairs
+        )
         checksum = self.__luhn_checksum(sanitized_value)
 
         return checksum
@@ -76,9 +80,3 @@ def digits_of(n: str) -> List[int]:
         for d in even_digits:
             checksum += sum(digits_of(str(d * 2)))
         return checksum % 10 == 0
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text

presidio-analyzer/presidio_analyzer/predefined_recognizers/es_nie_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -39,6 +41,9 @@ def __init__(
         supported_entity: str = "ES_NIE",
         replacement_pairs: Optional[List[Tuple[str, str]]] = None,
     ):
+        self.replacement_pairs = (
+            replacement_pairs if replacement_pairs else [("-", ""), (" ", "")]
+        )
         patterns = patterns if patterns else self.PATTERNS
         context = context if context else self.CONTEXT
         super().__init__(
@@ -51,7 +56,9 @@ def __init__(
     def validate_result(self, pattern_text: str) -> bool:
         """Validate the pattern by using the control character."""
 
-        pattern_text = EsNieRecognizer.__sanitize_value(pattern_text)
+        pattern_text = ValidationUtils.sanitize_value(
+            pattern_text, self.replacement_pairs
+        )
 
         letters = "TRWAGMYFPDXBNJZSQVHLCKE"
         letter = pattern_text[-1]
@@ -66,7 +73,3 @@ def validate_result(self, pattern_text: str) -> bool:
         # replace XYZ with 012, and check the mod 23
         number = int(str("XYZ".index(pattern_text[0])) + pattern_text[1:-1])
         return letter == letters[number % 23]
-
-    @staticmethod
-    def __sanitize_value(text: str) -> str:
-        return text.replace("-", "").replace(" ", "")

presidio-analyzer/presidio_analyzer/predefined_recognizers/es_nif_recognizer.py

@@ -1,5 +1,7 @@
 from typing import List, Optional, Tuple
 
+from validation.validation_utils import ValidationUtils
+
 from presidio_analyzer import Pattern, PatternRecognizer
 
 
@@ -47,12 +49,10 @@ def __init__(
         )
 
     def validate_result(self, pattern_text: str) -> bool:  # noqa D102
-        pattern_text = EsNifRecognizer.__sanitize_value(pattern_text)
+        pattern_text = ValidationUtils.sanitize_value(
+            pattern_text, self.replacement_pairs
+        )
         letter = pattern_text[-1]
         number = int("".join(filter(str.isdigit, pattern_text)))
         letters = "TRWAGMYFPDXBNJZSQVHLCKE"
         return letter == letters[number % 23]
-
-    @staticmethod
-    def __sanitize_value(text: str) -> str:
-        return text.replace("-", "").replace(" ", "")

presidio-analyzer/presidio_analyzer/predefined_recognizers/iban_recognizer.py

@@ -3,6 +3,7 @@
 from typing import Dict, List, Optional, Tuple
 
 import regex as re
+from validation.validation_utils import ValidationUtils
 
 from presidio_analyzer import (
     EntityRecognizer,
@@ -79,7 +80,9 @@ def __init__(
 
     def validate_result(self, pattern_text: str):  # noqa D102
         try:
-            pattern_text = self.__sanitize_value(pattern_text, self.replacement_pairs)
+            pattern_text = ValidationUtils.sanitize_value(
+                pattern_text, self.replacement_pairs
+            )
             is_valid_checksum = (
                 self.__generate_iban_check_digits(pattern_text, self.LETTERS)
                 == pattern_text[2:4]
@@ -204,9 +207,3 @@ def __is_valid_format(
             return country_regex and re.match(country_regex, iban, flags=flags)
 
         return False
-
-    @staticmethod
-    def __sanitize_value(text: str, replacement_pairs: List[Tuple[str, str]]) -> str:
-        for search_string, replacement_string in replacement_pairs:
-            text = text.replace(search_string, replacement_string)
-        return text