Reduce memory usage of Analyzer test suite (#1429)

presidio-analyzer/tests/test_analyzer_engine.py

@@ -1,5 +1,6 @@
 import copy
 from abc import ABC
+from contextlib import nullcontext
 from typing import List, Optional
 import re
 
@@ -58,11 +59,6 @@ def unit_test_guid():
     return "00000000-0000-0000-0000-000000000000"
 
 
-@pytest.fixture(scope="module")
-def nlp_engine(nlp_engines):
-    return nlp_engines["spacy_en"]
-
-
 def test_simple():
     dic = {
         "text": "John Smith drivers license is AC432223",
@@ -91,39 +87,25 @@ def test_when_analyze_with_predefined_recognizers_then_return_results(
     assert len(results) == 1
     assert_result(results[0], "CREDIT_CARD", 14, 33, max_score)
 
-def test_when_analyze_with_unsupported_language_must_match(
-):
-    with pytest.raises(ValueError):
-        registry = RecognizerRegistryProvider(registry_configuration={"supported_languages": ["en"]}).create_recognizer_registry()
-        AnalyzerEngine(
-            registry=registry,
-            supported_languages=["es", "de"]
-        )
-
-    with pytest.raises(ValueError):
-        registry = RecognizerRegistryProvider().create_recognizer_registry()
+@pytest.mark.parametrize(
+    "registry_config,analyzer_lang,expectation",
+    [
+        ({"supported_languages": ["en"]}, ["es", "de"], pytest.raises(ValueError)),
+        (None, ["es", "de"], pytest.raises(ValueError)),
+        ({"supported_languages": ["es", "de"]}, None, pytest.raises(ValueError)),
+        ({"supported_languages": ["es", "de"]}, ["de", "es"], nullcontext()),
+        (None, None, nullcontext()),
+    ]
+)
+def test_when_analyze_with_unsupported_language_must_match(registry_config, analyzer_lang, expectation):
+    with expectation:
+        registry = RecognizerRegistryProvider(registry_configuration=registry_config).create_recognizer_registry()
         AnalyzerEngine(
             registry=registry,
-            supported_languages=["es", "de"]
-        )
-
-    with pytest.raises(ValueError):
-        registry = RecognizerRegistryProvider(registry_configuration={"supported_languages": ["es", "de"]}).create_recognizer_registry()
-        AnalyzerEngine(
-            registry=registry
+            supported_languages=analyzer_lang,
+            nlp_engine=NlpEngineMock(),
         )
 
-    registry = RecognizerRegistryProvider(registry_configuration={"supported_languages": ["es", "de"]}).create_recognizer_registry()
-    AnalyzerEngine(
-        registry=registry,
-        supported_languages=["de", "es"]
-    )
-
-    registry = RecognizerRegistryProvider().create_recognizer_registry()
-    AnalyzerEngine(
-        registry=registry
-    )
-
 def test_when_analyze_with_defaults_success(
 ):
     registry = RecognizerRegistryProvider().create_recognizer_registry()

presidio-analyzer/tests/test_nlp_engine_provider.py

@@ -47,17 +47,6 @@ def nlp_configuration_dict() -> Dict:
     return nlp_configuration
 
 
-@pytest.fixture(scope="session")
-def ner_model_configuration_dict() -> Dict:
-    ner_model_configuration = {
-        "nlp_engine_name": "transformers",
-        "aggregation_strategy": "simple",
-        "alignment_mode": "strict",
-        "low_score_entity_names": ["O"],
-    }
-    return ner_model_configuration
-
-
 def test_when_create_nlp_engine__then_return_default_configuration():
     provider = NlpEngineProvider()
     engine = provider.create_engine()

presidio-analyzer/tests/test_recognizer_registry.py

@@ -13,11 +13,6 @@
 from presidio_analyzer.predefined_recognizers import SpacyRecognizer
 
 
-@pytest.fixture(scope="module")
-def request_id():
-    return "UT"
-
-
 def create_mock_pattern_recognizer(lang, entity, name):
     return PatternRecognizer(
         supported_entity=entity,

presidio-analyzer/tests/test_stanza_recognizer.py

@@ -23,10 +23,9 @@ def nlp_recognizer(nlp_recognizers):
     return nlp_recognizers.get("stanza", None)
 
 
-def prepare_and_analyze(nlp, recognizer, text, ents):
-    nlp.load()
+def prepare_and_analyze(nlp, recognizer, text, entities):
     nlp_artifacts = nlp.process_text(text, "en")
-    results = recognizer.analyze(text, ents, nlp_artifacts)
+    results = recognizer.analyze(text, entities, nlp_artifacts)
     return results
 
 

presidio-analyzer/tests/test_transformers_recognizer.py

@@ -24,7 +24,6 @@ def nlp_engine(nlp_engines):
 
 
 def prepare_and_analyze(nlp, recognizer, text, entities):
-    nlp.load()
     nlp_artifacts = nlp.process_text(text, "en")
     results = recognizer.analyze(text, entities, nlp_artifacts)
     return results