perf: share one ExpressionRuntime across DataFactoryElement evaluations

src/data_factory_testing_framework/models/_data_factory_element.py

@@ -1,4 +1,5 @@
 import json
+from functools import lru_cache
 from typing import Any
 
 from data_factory_testing_framework._expression_runtime.expression_runtime import ExpressionRuntime
@@ -10,6 +11,13 @@
 from data_factory_testing_framework.state import RunState
 
 
+@lru_cache(maxsize=None)
+def _get_expression_runtime() -> ExpressionRuntime:
+    # Building an ExpressionRuntime compiles the Lark grammar, which costs more than evaluating
+    # an expression. It holds no per-evaluation state, so one instance serves every element.
+    return ExpressionRuntime()
+
+
 class DataFactoryElement:
     expression: str
     result: DataFactoryObjectType
@@ -26,8 +34,7 @@ def __init__(self, expression: str) -> None:
     def evaluate(self, state: RunState) -> DataFactoryObjectType:
         """Evaluate the expression."""
         try:
-            expression_runtime = ExpressionRuntime()
-            self.result = expression_runtime.evaluate(self.expression, state)
+            self.result = _get_expression_runtime().evaluate(self.expression, state)
             return self.result
         except UserError as e:
             raise e from e

tests/functional/datafactory_element/test_evaluate_datafactory_element.py

@@ -1,10 +1,14 @@
+from unittest.mock import patch
+
 import pytest
+from data_factory_testing_framework._expression_runtime.expression_runtime import ExpressionRuntime
 from data_factory_testing_framework.exceptions import (
     DataFactoryElementEvaluationError,
     ParameterNotFoundError,
 )
 from data_factory_testing_framework.exceptions._user_error import UserError
-from data_factory_testing_framework.models import DataFactoryElement
+from data_factory_testing_framework.models import DataFactoryElement, _data_factory_element
+from data_factory_testing_framework.models._data_factory_element import _get_expression_runtime
 from data_factory_testing_framework.state import PipelineRunState
 
 
@@ -44,3 +48,19 @@ def test_evaluate_datafactory_element_raises_technical_errors() -> None:
     # Act
     with pytest.raises(DataFactoryElementEvaluationError):
         data_factory_element.evaluate(state)
+
+
+def test_evaluate_datafactory_element_reuses_one_expression_runtime() -> None:
+    # Arrange
+    _get_expression_runtime.cache_clear()
+    state = PipelineRunState()
+    elements = [DataFactoryElement("@mul(2, 3)"), DataFactoryElement("@add(2, 3)")]
+
+    # Act
+    with patch.object(_data_factory_element, "ExpressionRuntime", wraps=ExpressionRuntime) as runtime_class:
+        results = [element.evaluate(state) for element in elements]
+        results.append(elements[0].evaluate(state))
+
+    # Assert
+    assert results == [6, 5, 6]
+    assert runtime_class.call_count == 1