From 6fae4b3c3977edcddfeb9e0080f91718ae5c77d6 Mon Sep 17 00:00:00 2001 From: mateo-berri <277851410+mateo-berri@users.noreply.github.com> Date: Wed, 2 Sep 2026 13:32:21 -0700 Subject: [PATCH] fix(guardrails): keep the presidio output masker from unmasking after an in-memory update --- .../proxy/guardrails/guardrail_hooks/presidio.py | 2 ++ .../guardrails/guardrail_hooks/test_presidio.py | 12 ++++++++++++ .../proxy/guardrails/test_guardrail_registry.py | 13 +++++++++++-- 3 files changed, 25 insertions(+), 2 deletions(-) diff --git a/litellm/proxy/guardrails/guardrail_hooks/presidio.py b/litellm/proxy/guardrails/guardrail_hooks/presidio.py index da51a905ae3..70ea21320ee 100644 --- a/litellm/proxy/guardrails/guardrail_hooks/presidio.py +++ b/litellm/proxy/guardrails/guardrail_hooks/presidio.py @@ -1633,6 +1633,8 @@ class _OPTIONAL_PresidioPIIMasking(CustomGuardrail): Update the guardrails litellm params in memory """ super().update_in_memory_litellm_params(litellm_params) + if self.apply_to_output: + self.output_parse_pii = False if litellm_params.pii_entities_config: self.pii_entities_config = litellm_params.pii_entities_config if litellm_params.presidio_score_thresholds: diff --git a/tests/test_litellm/proxy/guardrails/guardrail_hooks/test_presidio.py b/tests/test_litellm/proxy/guardrails/guardrail_hooks/test_presidio.py index fcf940afd0d..84f7611c0c0 100644 --- a/tests/test_litellm/proxy/guardrails/guardrail_hooks/test_presidio.py +++ b/tests/test_litellm/proxy/guardrails/guardrail_hooks/test_presidio.py @@ -3129,6 +3129,18 @@ def test_update_in_memory_applies_analyze_chunk_size(): assert guardrail.presidio_analyze_chunk_size_bytes == 99_000 +def test_update_in_memory_keeps_output_masker_from_unmasking(): + masker = _OPTIONAL_PresidioPIIMasking(mock_testing=True, apply_to_output=True, output_parse_pii=False) + unmasker = _OPTIONAL_PresidioPIIMasking(mock_testing=True, output_parse_pii=True) + params = LitellmParams(guardrail="presidio", mode="pre_call", output_parse_pii=True) + + masker.update_in_memory_litellm_params(params) + unmasker.update_in_memory_litellm_params(params) + + assert (masker.apply_to_output, masker.output_parse_pii) == (True, False) + assert (unmasker.apply_to_output, unmasker.output_parse_pii) == (False, True) + + def test_merge_drops_truncated_same_type_fragment_from_overlap(): """A boundary entity seen truncated by chunk 1 and whole by chunk 2 must merge to the single full span; keeping both overlapping spans corrupts the diff --git a/tests/test_litellm/proxy/guardrails/test_guardrail_registry.py b/tests/test_litellm/proxy/guardrails/test_guardrail_registry.py index 5cbdef5f92f..24742e1bac2 100644 --- a/tests/test_litellm/proxy/guardrails/test_guardrail_registry.py +++ b/tests/test_litellm/proxy/guardrails/test_guardrail_registry.py @@ -566,7 +566,14 @@ def test_update_in_memory_guardrail_reaches_presidio_siblings_and_keeps_their_st try: handler.initialize_guardrail(_presidio_db_guardrail({"EMAIL_ADDRESS": "MASK", "IP_ADDRESS": "MASK"})) tracked = _presidio_callbacks_in(litellm.callbacks) - roles_before = [(callback.apply_to_output, callback.event_hook) for callback in tracked] + roles_before = [ + (callback.apply_to_output, callback.output_parse_pii, callback.event_hook) for callback in tracked + ] + assert roles_before == [ + (False, True, [GuardrailEventHooks.pre_call, GuardrailEventHooks.post_call]), + (False, True, GuardrailEventHooks.post_call), + (True, False, GuardrailEventHooks.post_call), + ] updated = Guardrail( guardrail_id=PRESIDIO_SIBLINGS_GID, @@ -585,7 +592,9 @@ def test_update_in_memory_guardrail_reaches_presidio_siblings_and_keeps_their_st handler.update_in_memory_guardrail(guardrail_id=PRESIDIO_SIBLINGS_GID, guardrail=updated) assert [callback.pii_entities_config for callback in tracked] == [{"EMAIL_ADDRESS": "MASK"}] * 3 - assert [(callback.apply_to_output, callback.event_hook) for callback in tracked] == roles_before + assert [ + (callback.apply_to_output, callback.output_parse_pii, callback.event_hook) for callback in tracked + ] == roles_before assert _presidio_callbacks_in(litellm.callbacks) == tracked finally: for cb_list, snapshot in zip(lists, snapshots):