Add support for allow_list, allow_list_match, regex_flags in REST API

e2e-tests/tests/test_analyzer.py

@@ -480,3 +480,95 @@ def test_given_ad_hoc_deny_list_recognizer_the_right_entities_are_returned():
     assert equal_json_strings(
         expected_response, response_content, ignore_keys=["recognition_metadata"]
     )
+
+
+@pytest.mark.api
+def test_given_allow_list_then_no_entity_is_returned():
+    request_body = """
+    {
+        "text": "email: [email protected]", 
+        "language": "en", 
+        "allow_list": ["[email protected]"]
+    }
+    """
+
+    response_status, response_content = analyze(request_body)
+
+    expected_response = """
+     []
+    """
+    assert response_status == 200
+    assert equal_json_strings(
+        expected_response, response_content
+    )
+
+
+@pytest.mark.api
+def test_given_allow_list_with_regex_match_then_no_entity_is_returned():
+    request_body = """
+    {
+        "text": "email: [email protected]", 
+        "language": "en", 
+        "allow_list": [".*@github.com"],
+        "allow_list_match": "regex"
+    }
+    """
+
+    response_status, response_content = analyze(request_body)
+
+    expected_response = """
+     []
+    """
+    assert response_status == 200
+    assert equal_json_strings(
+        expected_response, response_content
+    )
+
+
+@pytest.mark.api
+def test_given_allow_list_without_setting_allow_list_match_then_normal_entity_is_returned():
+    request_body = """
+    {
+        "text": "email: [email protected]", 
+        "language": "en", 
+        "allow_list": [".*@github.com"]
+    }
+    """
+
+    response_status, response_content = analyze(request_body)
+
+    expected_response = """
+     [
+         {"entity_type": "EMAIL_ADDRESS", "start": 7, "end": 23, "score": 0.85, "analysis_explanation":null}
+     ]
+    """
+    assert response_status == 200
+    assert equal_json_strings(
+        expected_response, response_content, ignore_keys=["recognition_metadata"]
+    )
+
+
+@pytest.mark.api
+def test_given_regex_flags_and_normal_entities_are_returned():
+    # case sensitive flags are turned off, GitHub != github
+    request_body = """
+    {
+        "text": "email: [email protected]",
+        "language": "en", 
+        "allow_list": [".*@github.com"],
+        "allow_list_match": "regex",
+        "regex_flags": 0
+    }
+    """
+
+    response_status, response_content = analyze(request_body)
+
+    expected_response = """
+     [
+         {"entity_type": "EMAIL_ADDRESS", "start": 7, "end": 23, "score": 0.85, "analysis_explanation":null}
+     ]
+    """
+    assert response_status == 200
+    assert equal_json_strings(
+        expected_response, response_content, ignore_keys=["recognition_metadata"]
+    )

presidio-analyzer/app.py

@@ -74,6 +74,9 @@ def analyze() -> Tuple[str, int]:
                     return_decision_process=req_data.return_decision_process,
                     ad_hoc_recognizers=req_data.ad_hoc_recognizers,
                     context=req_data.context,
+                    allow_list=req_data.allow_list,
+                    allow_list_match=req_data.allow_list_match,
+                    regex_flags=req_data.regex_flags
                 )
 
                 return Response(

presidio-analyzer/presidio_analyzer/analyzer_request.py

@@ -1,3 +1,4 @@
+import re
 from typing import Dict
 
 from presidio_analyzer import PatternRecognizer
@@ -34,3 +35,7 @@ def __init__(self, req_data: Dict):
                 PatternRecognizer.from_dict(rec) for rec in ad_hoc_recognizers
             ]
         self.context = req_data.get("context")
+        self.allow_list = req_data.get("allow_list")
+        self.allow_list_match = req_data.get("allow_list_match", "exact")
+        self.regex_flags = req_data.get("regex_flags",
+                                        re.DOTALL | re.MULTILINE | re.IGNORECASE)