feat(analyzer): add country-specific predefined recognizer for Taiwan (TW) National ID

presidio-analyzer/app.py

@@ -8,13 +8,14 @@
 from typing import Tuple
 
 from flask import Flask, Response, jsonify, request
+from werkzeug.exceptions import HTTPException
+
 from presidio_analyzer import (
     AnalyzerEngine,
     AnalyzerEngineProvider,
     AnalyzerRequest,
     BatchAnalyzerEngine,
 )
-from werkzeug.exceptions import HTTPException
 
 DEFAULT_PORT = "3000"
 DEFAULT_BATCH_SIZE = "500"

presidio-analyzer/presidio_analyzer/predefined_recognizers/country_specific/tw/__init__.py

@@ -0,0 +1,5 @@
+"""Taiwan specific recognizers."""
+
+from .tw_national_id_recognizer import TwNationalIdRecognizer
+
+__all__ = ["TwNationalIdRecognizer"]

presidio-analyzer/presidio_analyzer/predefined_recognizers/country_specific/tw/tw_national_id_recognizer.py

@@ -0,0 +1,78 @@
+"""Taiwan National ID Recognizer."""
+
+from typing import List, Optional
+
+from presidio_analyzer import Pattern, PatternRecognizer
+
+
+class TwNationalIdRecognizer(PatternRecognizer):
+    """Recognize Taiwan National ID using patterns and checksums."""
+
+    PATTERNS = [
+        Pattern(
+            "TW National ID (Strict)",
+            r"(?<![A-Za-z0-9])[A-Z][1289][0-9]{8}(?![A-Za-z0-9])",
+            0.3,
+        )
+    ]
+
+    CONTEXT = [
+        "身分證",
+        "統一證號",
+        "國民身分證",
+        "tw id",
+        "taiwan id",
+    ]
+
+    def __init__(
+        self,
+        patterns: Optional[List[Pattern]] = None,
+        context: Optional[List[str]] = None,
+        supported_language: str = "zh",
+        supported_entity: str = "TW_NATIONAL_ID",
+    ):
+        """Initialize Taiwan National ID Recognizer."""
+        patterns = patterns if patterns is not None else self.PATTERNS
+        context = context if context is not None else self.CONTEXT
+        super().__init__(
+            supported_entity=supported_entity,
+            patterns=patterns,
+            context=context,
+            supported_language=supported_language,
+        )
+
+    def invalidate_result(self, pattern_text: str) -> bool:
+        """Reject invalid Taiwan ID structures via Modulus-10 checksum validation."""
+        if len(pattern_text) != 10:
+            return True
+
+        # Presidio uses IGNORECASE by default; explicitly reject lowercase initial letters
+        first_char = pattern_text[0]
+        if not first_char.isupper():
+            return True
+
+        letter_codes = {
+            'A': 10, 'B': 11, 'C': 12, 'D': 13, 'E': 14, 'F': 15, 'G': 16, 'H': 17,
+            'I': 34, 'J': 18, 'K': 19, 'L': 20, 'M': 21, 'N': 22, 'O': 35, 'P': 23,
+            'Q': 24, 'R': 25, 'S': 26, 'T': 27, 'U': 28, 'V': 29, 'W': 32, 'X': 30,
+            'Y': 31, 'Z': 33
+        }
+
+        if first_char not in letter_codes:
+            return True
+
+        if pattern_text[1] not in ('1', '2', '8', '9'):
+            return True
+
+        try:
+            code = letter_codes[first_char]
+            n1 = code // 10
+            n2 = code % 10
+
+            digits = [n1, n2] + [int(char) for char in pattern_text[1:]]
+            weights = [1, 9, 8, 7, 6, 5, 4, 3, 2, 1, 1]
+            total = sum(d * w for d, w in zip(digits, weights))
+
+            return total % 10 != 0
+        except ValueError:
+            return True

presidio-analyzer/tests/test_tw_national_id_recognizer.py

@@ -0,0 +1,33 @@
+import pytest
+from presidio_analyzer.predefined_recognizers.country_specific.tw import TwNationalIdRecognizer
+
+
+@pytest.fixture(scope="module")
+def recognizer():
+    return TwNationalIdRecognizer()
+
+
+@pytest.mark.parametrize(
+    "text, expected_len, expected_positions",
+    [
+        # Valid Taiwan IDs (Verified with Modulus-10 checksum)
+        ("My ID is A123456789.", 1, ((9, 19),)),
+        ("B120000008", 1, ((0, 10),)),
+        ("F120000002", 1, ((0, 10),)),
+        ("H120000004", 1, ((0, 10),)),
+        # Adjacent to Chinese characters
+        ("身分證A123456789", 1, ((3, 13),)),
+        # Invalid Formats / Non-Matches / Checksum Failures
+        ("A323456789", 0, ()),  # Invalid gender code (3)
+        ("A12345678", 0, ()),   # Too short
+        ("A1234567890", 0, ()), # Too long
+        ("1123456789", 0, ()),  # Missing letter
+        ("a123456789", 0, ()),  # Lowercase prefix rejected
+        ("A123456780", 0, ()),  # Checksum failure
+    ],
+)
+def test_tw_national_id_recognizer(text, expected_len, expected_positions, recognizer):
+    results = recognizer.analyze(text, entities=["TW_NATIONAL_ID"])
+    assert len(results) == expected_len
+    if expected_len > 0:
+        assert (results[0].start, results[0].end) == expected_positions[0]

presidio-cli/presidio_cli/analyzer.py

@@ -1,7 +1,6 @@
 from typing import Generator, Optional, Union
 
 from presidio_analyzer import RecognizerResult
-
 from presidio_cli.config import PresidioCLIConfig
 
 

presidio-cli/presidio_cli/config.py

@@ -3,6 +3,7 @@
 
 import pathspec
 import yaml
+
 from presidio_analyzer import AnalyzerEngine
 
 

presidio-image-redactor/presidio_image_redactor/dicom_image_pii_verify_engine.py

@@ -4,8 +4,8 @@
 import PIL
 import pydicom
 from PIL import Image
-from presidio_analyzer import PatternRecognizer
 
+from presidio_analyzer import PatternRecognizer
 from presidio_image_redactor import (
     OCR,
     BboxProcessor,

presidio-image-redactor/presidio_image_redactor/dicom_image_redactor_engine.py

@@ -9,10 +9,10 @@
 import pydicom
 from matplotlib import pyplot as plt  # necessary import for PIL typing # noqa: F401
 from PIL import Image, ImageOps
-from presidio_analyzer import PatternRecognizer
 from pydicom.multival import MultiValue
 from pydicom.pixel_data_handlers.util import apply_voi_lut
 
+from presidio_analyzer import PatternRecognizer
 from presidio_image_redactor import (
     ImageAnalyzerEngine,  # noqa: F401
     ImageRedactorEngine,

presidio-image-redactor/presidio_image_redactor/image_analyzer_engine.py

@@ -6,8 +6,8 @@
 import matplotlib.pyplot as plt
 import numpy as np
 from PIL import Image, ImageChops
-from presidio_analyzer import AnalyzerEngine, RecognizerResult
 
+from presidio_analyzer import AnalyzerEngine, RecognizerResult
 from presidio_image_redactor import OCR, ImagePreprocessor, TesseractOCR
 from presidio_image_redactor.entities import ImageRecognizerResult
 

presidio-image-redactor/presidio_image_redactor/image_pii_verify_engine.py

@@ -2,8 +2,8 @@
 from typing import List, Optional
 
 from PIL import Image, ImageChops
-from presidio_analyzer import PatternRecognizer
 
+from presidio_analyzer import PatternRecognizer
 from presidio_image_redactor.image_redactor_engine import ImageRedactorEngine
 
 

presidio-image-redactor/presidio_image_redactor/image_redactor_engine.py

@@ -1,9 +1,9 @@
 from typing import List, Optional, Tuple, Union
 
-import presidio_analyzer  # required for isinstance check which throws an error when trying to specify PatternRecognizer  # noqa: E501
 from PIL import Image, ImageChops, ImageDraw
-from presidio_analyzer import PatternRecognizer
 
+import presidio_analyzer  # required for isinstance check which throws an error when trying to specify PatternRecognizer  # noqa: E501
+from presidio_analyzer import PatternRecognizer
 from presidio_image_redactor import BboxProcessor, ImageAnalyzerEngine
 from presidio_image_redactor.entities import ImageRecognizerResult
 

presidio-structured/presidio_structured/analysis_builder.py

@@ -5,13 +5,13 @@
 from typing import Dict, Iterator, List, Optional, Union
 
 from pandas import DataFrame
+
 from presidio_analyzer import (
     AnalyzerEngine,
     BatchAnalyzerEngine,
     DictAnalyzerResult,
     RecognizerResult,
 )
-
 from presidio_structured.config import StructuredAnalysis
 
 NON_PII_ENTITY_TYPE = "NON_PII"