From 72144c0ca6d88e097d2fb05c171964d4c44f528c Mon Sep 17 00:00:00 2001
From: Plamen Valentinov Kolev
 <41479552+pvk-developer@users.noreply.github.com>
Date: Tue, 2 May 2023 19:00:48 +0200
Subject: [PATCH] Use cached version of faker rather than instance every time
 (#1406)

* Use cached version of faker rather than instance every time

* instance lru_cache
---
 sdv/metadata/anonymization.py             |  9 ++++++++-
 tests/unit/metadata/test_anonymization.py | 21 +++++++++++++++++----
 2 files changed, 25 insertions(+), 5 deletions(-)

diff --git a/sdv/metadata/anonymization.py b/sdv/metadata/anonymization.py
index 56d465885..7e599e8e1 100644
--- a/sdv/metadata/anonymization.py
+++ b/sdv/metadata/anonymization.py
@@ -2,6 +2,7 @@
 
 import inspect
 import warnings
+from functools import lru_cache
 
 from faker import Faker
 from faker.config import AVAILABLE_LOCALES
@@ -47,6 +48,12 @@
 }
 
 
+@lru_cache()
+def get_faker_instance():
+    """Return a ``faker.Faker`` instance with all the locales."""
+    return Faker(AVAILABLE_LOCALES)
+
+
 def is_faker_function(function_name):
     """Return whether or not the function name is a valid Faker function.
 
@@ -60,7 +67,7 @@ def is_faker_function(function_name):
     try:
         with warnings.catch_warnings():
             warnings.filterwarnings('ignore', module='faker')
-            getattr(Faker(AVAILABLE_LOCALES), function_name)
+            getattr(get_faker_instance(), function_name)
     except AttributeError:
         return False
 
diff --git a/tests/unit/metadata/test_anonymization.py b/tests/unit/metadata/test_anonymization.py
index a854da5a1..b2983cfde 100644
--- a/tests/unit/metadata/test_anonymization.py
+++ b/tests/unit/metadata/test_anonymization.py
@@ -1,7 +1,7 @@
 from unittest.mock import Mock, patch
 
 from sdv.metadata.anonymization import (
-    _detect_provider_name, get_anonymized_transformer, is_faker_function)
+    _detect_provider_name, get_anonymized_transformer, get_faker_instance, is_faker_function)
 
 
 class TestAnonimization:
@@ -110,18 +110,31 @@ def test_is_faker_function(self, faker_mock):
         # Assert
         assert result is True
 
-    @patch('sdv.metadata.anonymization.Faker')
-    def test_is_faker_function_error(self, faker_mock):
+    @patch('sdv.metadata.anonymization.get_faker_instance')
+    def test_is_faker_function_error(self, mock_get_faker_instance):
         """Test that the method returns False if ``function_name`` is not a valid faker function.
 
         If the ``function_name`` is not an attribute of ``Faker()`` then we should return false.
         This test mocks ``Faker`` to not have the attribute that is passed as ``function_name``.
         """
         # Setup
-        faker_mock.return_value = Mock(spec=[])
+        mock_get_faker_instance.return_value = Mock(spec=[])
 
         # Run
         result = is_faker_function('blah')
 
         # Assert
         assert result is False
+        mock_get_faker_instance.assert_called_once()
+
+    @patch('sdv.metadata.anonymization.Faker')
+    def test_get_faker_instance(self, mock_faker):
+        """Test that ``get_faker_instance`` returns the same object."""
+        # Setup
+        first_instance = get_faker_instance()
+
+        # Run
+        second_instance = get_faker_instance()
+
+        # Assert
+        assert id(first_instance) == id(second_instance)