feat: add negative_context support to reduce false positives in context-aware PII detection by TheSabari07 · Pull Request #1969 · data-privacy-stack/presidio

TheSabari07 · 2026-04-12T14:33:42Z

Change Description

Added support for negative_context in context-aware PII detection to reduce false positives.

Added support for negative_context in context-aware PII detection
Applied score penalty when negative context words appear near detected entities
Updated context enhancer to handle both positive and negative signals independently
Ensured compatibility with predefined recognizers by filtering unsupported arguments
Added tests to validate behavior, edge cases, and backward compatibility

Issue reference

Fixes #1686

Checklist

I have reviewed the contribution guidelines
I have signed the CLA (if required)
My code includes unit tests
All unit tests and lint checks pass locally
My PR contains documentation updates / additions if required

TheSabari07 · 2026-04-12T14:35:44Z

Hi @omri374

I worked on this as part of the discussion in #1686.

This PR focuses specifically on adding negative context support to reduce false positives in rule-based detection. Would appreciate your thoughts on the approach

Also, I had a couple of quick questions:

Can I include the additional test file I used to verify the negative context behavior?
Is it okay if I start experimenting with this in the presidio-research repo to further validate the approach?

omri374 · 2026-04-13T14:20:07Z

                f"Got: {context_matching_mode}"
            )
        self.context_matching_mode = context_matching_mode
+        self.negative_context_penalty = negative_context_penalty


Can we add this to the parent ContextAwareEnhancer? it could serve other context enhancers too

Okay @omri374

that makes sense.

I’ll move negative_context_penalty to the base ContextAwareEnhancer so it can be reused across other context enhancers as well.

omri374

This is a great start. Please add tests + update a recognizer to include negative context.

Copilot

Pull request overview

Adds negative_context support to Presidio Analyzer’s context-aware scoring to reduce false positives by penalizing matches when “negative” keywords appear near an entity.

Changes:

Extend recognizer configuration loading to read and pass negative_context (including per-language config handling).
Add negative_context plumbing to EntityRecognizer/PatternRecognizer (init + serialization).
Update LemmaContextAwareEnhancer to apply a configurable score penalty when negative context words appear near detected entities.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.

File	Description
presidio-analyzer/presidio_analyzer/recognizer_registry/recognizers_loader_utils.py	Loads `negative_context` from config and filters it when recognizers don’t accept the argument.
presidio-analyzer/presidio_analyzer/pattern_recognizer.py	Adds `negative_context` to PatternRecognizer init and (de)serialization.
presidio-analyzer/presidio_analyzer/entity_recognizer.py	Adds `negative_context` to the base recognizer API and stores it on the instance.
presidio-analyzer/presidio_analyzer/context_aware_enhancers/lemma_context_aware_enhancer.py	Implements negative-context score penalty logic in the default context enhancer.

Copilot · 2026-04-13T14:37:24Z

+                if negative_context_word != "":
+                    result.score -= self.negative_context_penalty
+                    result.score = max(result.score, ContextAwareEnhancer.MIN_SCORE)
+                    logger.debug("Applied negative context penalty for word '%s'", negative_context_word)


After applying the negative_context penalty, the RecognizerResult.analysis_explanation isn't updated (set_improved_score is only called after the positive boost). This makes the explainability fields (score/score_context_improvement) inconsistent with the final result.score; update the AnalysisExplanation to reflect the post-penalty score as well.

Suggested change

logger.debug("Applied negative context penalty for word '%s'", negative_context_word)

result.analysis_explanation.set_improved_score(result.score)

logger.debug(

"Applied negative context penalty for word '%s'",

negative_context_word,

)

@TheSabari07 please make sure you update the analysis explanation fields as well.

Sure @omri374

Copilot · 2026-04-13T14:37:24Z

                    ),
+                    "negative_context": RecognizerListLoader._get_recognizer_negative_context(
+                        recognizer=recognizer_conf
+                    ),


The added negative_context extraction call is on lines that exceed the configured Ruff line-length (88) (e.g., the 'negative_context' assignment in the returned dict). Please wrap these function calls to avoid E501 lint failures.

Copilot · 2026-04-13T14:37:25Z

+            # Apply negative context penalty if recognizer has negative_context defined
+            if recognizer.negative_context:
+                negative_context_word = self._find_supportive_word_in_context(
+                    surrounding_words, recognizer.negative_context, self.context_matching_mode
+                )
+                if negative_context_word != "":
+                    result.score -= self.negative_context_penalty
+                    result.score = max(result.score, ContextAwareEnhancer.MIN_SCORE)
+                    logger.debug("Applied negative context penalty for word '%s'", negative_context_word)


negative_context introduces new scoring behavior (penalty application and clamping) but there are currently no unit tests covering negative_context in the test suite (no references found under presidio-analyzer/tests). Please add tests to validate: (1) penalty is applied when negative context appears in the window, (2) score is clamped at 0, (3) interaction with positive context (boost then penalty), and (4) backward compatibility when negative_context is unset.

TheSabari07 · 2026-04-14T08:39:19Z

This is a great start. Please add tests + update a recognizer to include negative context.

Thank you @omri374

I’ll add unit tests to cover negative_context (penalty, edge cases, and backward compatibility) and also update an existing recognizer to explicitly include negative_context for validation.

…xt enhancer

TheSabari07 · 2026-04-19T17:02:12Z

Hi @omri374

I’ve made the changes suggested in the review:

Updated negative_context handling to behave consistently with positive context (as discussed)
Moved negative_context support into EntityRecognizer level for cleaner design
Removed unnecessary filtering logic in recognizers_loader_utils.py as suggested
Ensured YAML/simple language configs don’t incorrectly extract context/negative_context
Aligned enhancer logic to avoid double boosting and correctly apply negative penalty independently
Updated tests to fully cover these changes and ensure backward compatibility

All tests are passing locally, and I’ve verified no regressions in existing analyzer tests.

Please verify and let me know if any changes are needed

TheSabari07 · 2026-04-25T09:52:52Z

Hi @omri374,
just a quick check on this PR. Happy to make any changes if needed.

omri374 · 2026-04-25T13:16:06Z

Apologies, will review shortly.

TheSabari07 · 2026-04-25T16:40:06Z

Apologies, will review shortly.

No issues @omri374, thanks for the update

Copilot

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 12 comments.

Comments suppressed due to low confidence (2)

presidio-analyzer/presidio_analyzer/analyzer_engine.py:165

negative_context is added to AnalyzerEngine.analyze, but the analyzer service (/analyze) builds its arguments from AnalyzerRequest (app.py), which currently doesn't parse/forward a negative_context field. As-is, REST clients can't use this feature; either wire it through the request object/endpoint or clarify that it's Python-only.

    def analyze(
        self,
        text: str,
        language: str,
        entities: Optional[List[str]] = None,
        correlation_id: Optional[str] = None,
        score_threshold: Optional[float] = None,
        return_decision_process: Optional[bool] = False,
        ad_hoc_recognizers: Optional[List[EntityRecognizer]] = None,
        context: Optional[List[str]] = None,
        negative_context: Optional[List[str]] = None,
        allow_list: Optional[List[str]] = None,
        allow_list_match: Optional[str] = "exact",
        regex_flags: Optional[int] = re.DOTALL | re.MULTILINE | re.IGNORECASE,
        nlp_artifacts: Optional[NlpArtifacts] = None,
    ) -> List[RecognizerResult]:

presidio-analyzer/presidio_analyzer/analyzer_engine.py:165

There’s no unit test exercising the new AnalyzerEngine.analyze(..., negative_context=...) parameter end-to-end (engine -> context enhancer). Adding a focused test would guard the public API behavior and ensure request-level negative context is applied correctly.

    def analyze(
        self,
        text: str,
        language: str,
        entities: Optional[List[str]] = None,
        correlation_id: Optional[str] = None,
        score_threshold: Optional[float] = None,
        return_decision_process: Optional[bool] = False,
        ad_hoc_recognizers: Optional[List[EntityRecognizer]] = None,
        context: Optional[List[str]] = None,
        negative_context: Optional[List[str]] = None,
        allow_list: Optional[List[str]] = None,
        allow_list_match: Optional[str] = "exact",
        regex_flags: Optional[int] = re.DOTALL | re.MULTILINE | re.IGNORECASE,
        nlp_artifacts: Optional[NlpArtifacts] = None,
    ) -> List[RecognizerResult]:

omri374

Thanks! Looks great, there are a few minor things here and there but it's mostly done.
Please confider adding this to the documentation, for example here: https://microsoft.github.io/presidio/tutorial/06_context/ or here

omri374 · 2026-04-27T09:01:02Z

        patterns = patterns if patterns else self.PATTERNS
        context = context if context else self.CONTEXT
+        negative_context = (
+            negative_context if negative_context else self.NEGATIVE_CONTEXT


@TheSabari07 please change this to make sure a user can disable negative context by passing an empty list.

TheSabari07 · 2026-05-16T05:52:14Z

Hi @omri374,

Sorry for the late update, I was a bit overloaded recently.

I’ve now addressed the review comments and pushed the fixes:

removed redundant context assignment
added context and negative_context to to_dict()
added backward-compatible enhancer handling
refined SSN-specific negative context words

I also re-ran the related tests and verified everything passes.

Please verify and let me know if any further changes are needed.

TheSabari07 · 2026-05-27T03:39:34Z

Hi @omri374,

Just a gentle follow-up on the latest updates for this PR. I’ve addressed the review comments and re-ran the related tests successfully.

Whenever you get a chance, please review and let me know if any further changes are needed from my side.

omri374 · 2026-05-27T14:40:14Z

-            recognizers=recognizers,
-            context=context,
+        # Check if the enhancer supports negative_context parameter for backward compatibility
+        enhancer_signature = inspect.signature(


I would first check if the user passed any context or negative context. If not, we shouldn't run this inspection

omri374 · 2026-05-27T17:06:23Z

Hi @TheSabari07, thanks for the reminder, apologies for the delay! I added one small change request, as we're touching the core of the package. Other than that, please run ruff format to make sure that the linting is correct.

TheSabari07 · 2026-05-29T03:27:00Z

Hi @TheSabari07, thanks for the reminder, apologies for the delay! I added one small change request, as we're touching the core of the package. Other than that, please run ruff format to make sure that the linting is correct.

Hi @omri374,

No issues, and thanks for the review. I’ll add the conditional check for the context parameters, run ruff format, and push the updates shortly.

Also, if possible, could you please share your email/contact for communication? I have a few doubts regarding a project idea and would really appreciate your guidance

Copilot

Pull request overview

Copilot reviewed 11 out of 11 changed files in this pull request and generated 5 comments.

omri374 · 2026-06-01T05:38:54Z

+        # Create empty list in None or lowercase all negative context words in the list
+        if not negative_context:
+            negative_context = []
+        else:
+            negative_context = [word.lower() for word in negative_context]


Please resolve

+            effective_negative_context = []
+            if recognizer.negative_context:
+                effective_negative_context.extend(recognizer.negative_context)
+            if negative_context:
+                effective_negative_context.extend(negative_context)


+    assert enhanced_results[0].score < original_score
+    expected_score = max(original_score - 0.3, 0)
+    assert enhanced_results[0].score == expected_score


TheSabari07 · 2026-05-31T19:54:00Z

Hi @omri374,

Sorry for the delay,

I addressed the optimization request by only running the enhancer signature inspection when negative_context is provided, while keeping backward compatibility for custom enhancers.
I also ran ruff formatting and pushed the updates.

Please review at your convenience and let me know if any further changes are needed.

Thank you.

omri374

Thanks @TheSabari07, appreciate all the work on this! Copilot surfaced a few points to fix.

omri374 · 2026-06-01T05:38:54Z

+        # Create empty list in None or lowercase all negative context words in the list
+        if not negative_context:
+            negative_context = []
+        else:
+            negative_context = [word.lower() for word in negative_context]


Please resolve

TheSabari07 · 2026-06-01T06:25:47Z

Hi @TheSabari07, thanks for the reminder, apologies for the delay! I added one small change request, as we're touching the core of the package. Other than that, please run ruff format to make sure that the linting is correct.

Hi @omri374,

No issues, and thanks for the review. I’ll add the conditional check for the context parameters, run ruff format, and push the updates shortly.

Also, if possible, could you please share your email/contact for communication? I have a few doubts regarding a project idea and would really appreciate your guidance

Hi @omri374,

Thanks for the review.
I will work on the changes suggested and will push the updates shortly

Also, if possible, could you please share your email/contact for communication? I have a few doubts regarding a project idea and would really appreciate your guidance

Edit :

Fixed the negative_context handling to properly distinguish between None and an explicitly provided empty list.
Moved the enhancer capability validation to AnalyzerEngine initialization and removed the runtime signature inspection.
Updated the related test assertions to use pytest.approx() where appropriate.
Ran formatting and verified the relevant test suite passes.

Please review at your convenience and let me know if any further changes are needed.
Thank you.

omri374 · 2026-06-01T14:57:38Z

Hi, feel free to reach me on LinkedIn (Omri Mendels)

TheSabari07 · 2026-06-01T16:13:38Z

Hi, feel free to reach me on LinkedIn (Omri Mendels)

Hi @omri374,

Thank you for sharing your LinkedIn. I sent you a connection request a few weeks ago. My name is Sabari Doss R.

TheSabari07 · 2026-06-22T08:03:29Z

Hi @omri374

Please review at your convenience and let me know if any further changes are needed.
Thank you.

TheSabari07 · 2026-07-01T01:06:04Z

Hi @omri374

Please review at your convenience and let me know if any further changes are needed. Thank you.

Hi @omri374, just checking in to see if there are any updates when you have a moment. Please let me know if any changes are needed from my side.
Thank you.

omri374

Hi @TheSabari07, really sorry for the delay on this! It's been busy times. I left more comments and will fix the conflict with main myself. Thanks!

omri374 · 2026-07-16T10:50:42Z

        "ssid",
    ]

+    NEGATIVE_CONTEXT = [


Please remove any changes to the us_ssn_recognizer. Users might be negatively influenced by this. For example:

PENALIZED by 'test' <- I attest that my SSN is 123-45-6789
PENALIZED by 'demo' <- Patient demographic form, SSN 123-45-6789
PENALIZED by 'test' <- Testimony record: SSN 123-45-6789
PENALIZED by 'test' <- Contested claim, SSN 123-45-6789
PENALIZED by 'demo' <- Democratic party member SSN 123-45-6789

omri374 · 2026-07-16T10:52:32Z

-                {"supported_language": language, "context": None}
+                {
+                    "supported_language": language,
+                    "context": RecognizerListLoader._get_recognizer_context(


this used to be {"supported_language": language, "context": None} but now it's {"supported_language": language, "context": CONTEXT}

omri374 · 2026-07-16T11:06:16Z

            f"Detected by `{recognizer_name}` using pattern `{pattern_name}`"
        )

        explanation = AnalysisExplanation(


Please make sure that negative_context also appears in the explainability feature (AnalysisExplanation)

omri374 · 2026-07-16T11:07:00Z

            "supported_language": self.supported_language,
            "name": self.name,
            "version": self.version,
+            "context": self.context,


please remove both, might not be backward compatible, or add a check if it exists

omri374 · 2026-07-16T11:08:58Z

+        :return: True if enhancer supports negative_context parameter, False otherwise
+        """
+        enhancer_signature = inspect.signature(enhancer.enhance_using_context)
+        return "negative_context" in enhancer_signature.parameters


looks great. Let's also add a warning if it's missing. Users might want to know + if they do send negative_context and for some reason it doesn't work, they wouldn't know.

omri374 · 2026-07-16T11:28:42Z

        self.version = version
        self.is_loaded = False
-        self.context = context if context else []
+        self.context = context if context is not None else []


This is a behavior change. The line in PatternRecognizer said self.context = context, which would be None if context=None, but now it would be []. Not a major change but could create unexpected side effects.

omri374 · 2026-07-16T11:29:43Z

@@ -3,6 +3,7 @@
 import os


Consider wiring this capability into app.py and analyzer_request.py so that it appears in the REST API too.

Copilot

Pull request overview

Copilot reviewed 11 out of 11 changed files in this pull request and generated 5 comments.

+            if not (recognizer.context or recognizer.negative_context):
                logger.debug(
                    "recognizer '%s' does not support context enhancement",
                    recognizer.name,
                )
                continue


        allow_list: Optional[List[str]] = None,
        allow_list_match: Optional[str] = "exact",
        regex_flags: Optional[int] = re.DOTALL | re.MULTILINE | re.IGNORECASE,
+        negative_context: Optional[List[str]] = None,
        nlp_artifacts: Optional[NlpArtifacts] = None,


+    print("Entity filtered out by negative context!")
+```
+
+The score is now reduced from 0.9 to approximately 0.6 (base score 0.9 minus default penalty of 0.3), which is typically below the confidence threshold and gets filtered out.


+```
+final_score = max(base_score - negative_context_penalty, 0)
+```
+
+- **base_score**: The original detection score (e.g., 0.9)
+- **negative_context_penalty**: Default is 0.3 (can be configured)
+- **final_score**: Never goes below 0


+        enhancer_signature = inspect.signature(enhancer.enhance_using_context)
+        return "negative_context" in enhancer_signature.parameters


TheSabari07 added 8 commits April 12, 2026 18:44

add optional negative_context support to EntityRecognizer

919c957

Merge branch 'microsoft:main' into feature/negative-context

94b30ec

add negative_context support to PatternRecognizer with serialization

dfa2f6b

apply negative context penalty in LemmaContextAwareEnhancer

27b664b

add YAML parsing support for negative_context

1ae1105

align default handling for context fields in EntityRecognizer

505dd8a

ensure negative_context reduces score correctly in enhancer

51447b0

safely handle negative_context in recognizer loader

0acd355

github-actions Bot added the external label Apr 12, 2026

TheSabari07 mentioned this pull request Apr 13, 2026

Improved context awareness using ML - Embeddings, ML classifiers or other non-rule-based approaches #1686

Open

omri374 reviewed Apr 13, 2026

View reviewed changes

Comment thread presidio-analyzer/presidio_analyzer/recognizer_registry/recognizers_loader_utils.py

omri374 reviewed Apr 13, 2026

View reviewed changes

omri374 requested a review from Copilot April 13, 2026 14:27

Copilot started reviewing on behalf of omri374 April 13, 2026 14:30 View session

Copilot AI reviewed Apr 13, 2026

View reviewed changes

omri374 and others added 3 commits April 17, 2026 16:11

Merge branch 'main' into feature/negative-context

48acd01

feat: introduce negative_context support across recognizers and conte…

f66b40e

…xt enhancer

test: add comprehensive tests for negative_context feature

01d9208

feat: add negative_context usage in US SSN recognizer

53c2a14

omri374 requested a review from Copilot April 26, 2026 20:25

Copilot started reviewing on behalf of omri374 April 26, 2026 20:26 View session

Copilot AI reviewed Apr 26, 2026

View reviewed changes

omri374 reviewed Apr 27, 2026

View reviewed changes

omri374 and others added 5 commits May 14, 2026 07:54

Merge branch 'main' into feature/negative-context

afc4590

refactor: remove redundant context assignment

fcfc741

fix: add context fields to recognizer serialization

f4f44c2

fix: preserve enhancer backward compatibility

02cdbc9

improve: refine ssn negative context words

e2ab7c4

Merge branch 'main' into feature/negative-context

570dca6

Merge branch 'main' into feature/negative-context

79cd70b

omri374 reviewed May 27, 2026

View reviewed changes

perf: avoid unnecessary enhancer signature inspection

627f128

Copilot AI review requested due to automatic review settings May 31, 2026 19:42

Copilot AI reviewed May 31, 2026

View reviewed changes

omri374 reviewed Jun 1, 2026

View reviewed changes

TheSabari07 added 3 commits June 1, 2026 14:32

fix: preserve negative context override semantics

330a4e7

refactor: validate enhancer capabilities at initialization

cb12a2b

test: use approximate float assertions

bd36b96

omri374 reviewed Jul 16, 2026

View reviewed changes

Merge remote-tracking branch 'origin/main' into pr1969-merge

d51fc8f

Copilot AI review requested due to automatic review settings July 16, 2026 12:48

Copilot started reviewing on behalf of omri374 July 16, 2026 12:50 View session

Copilot AI reviewed Jul 16, 2026

View reviewed changes

		enhancer_signature = inspect.signature(enhancer.enhance_using_context)
		return "negative_context" in enhancer_signature.parameters

Uh oh!

Conversation

TheSabari07 commented Apr 12, 2026

Change Description

Issue reference

Checklist

Uh oh!

TheSabari07 commented Apr 12, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

omri374 left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Copilot AI Apr 13, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI Apr 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 13, 2026

Choose a reason for hiding this comment

Uh oh!

TheSabari07 commented Apr 14, 2026

Uh oh!

TheSabari07 commented Apr 19, 2026

Uh oh!

TheSabari07 commented Apr 25, 2026

Uh oh!

omri374 commented Apr 25, 2026

Uh oh!

TheSabari07 commented Apr 25, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

omri374 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

TheSabari07 commented May 16, 2026

Uh oh!

TheSabari07 commented May 27, 2026

Uh oh!

omri374 May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

omri374 May 27, 2026 •

edited

Loading

TheSabari07 commented Jun 1, 2026 •

edited

Loading