feature: address robbievanleeuwen#29 using vectorization on get_resul…

…ts_lists() in results

tests/test_get_result_list.py

@@ -0,0 +1,124 @@
+import pytest
+import time
+import numpy as np
+from numpy import random, ndarray as NdArray, testing
+from dataclasses import dataclass, field
+from colorama import Fore, Style
+
+
+@dataclass(order=True)
+class UltimateBendingResults:
+    # bending angle
+    # theta: float
+
+    # ultimate neutral axis depth
+    # d_n: float = 0
+    # k_u: float = 0
+
+    # resultant actions
+    n: float = 0
+    m_x: float = 0
+    m_y: float = 0
+    m_xy: float = 0
+
+    # label
+    # label: str | None = field(default=None, compare=False)
+
+
+@dataclass
+class BiaxialBendingResults:
+    n: float
+    results: list[UltimateBendingResults] = field(default_factory=list)
+
+    def get_results_lists(self) -> tuple[list[float], list[float]]:
+        # build list of results
+        m_x_list = []
+        m_y_list = []
+
+        for result in self.results:
+            m_x_list.append(result.m_x)
+            m_y_list.append(result.m_y)
+
+        return m_x_list, m_y_list
+
+
+@dataclass
+class NewBiaxialBendingResults:
+    n: float
+    # results: list[UltimateBendingResults] = field(default_factory=list)
+    results: NdArray  # swap the list out for an NdArray
+
+    @classmethod
+    def from_list(
+        cls,
+        n: float,
+        results: list[UltimateBendingResults]
+        ):
+
+        dtype = [('m_x', float), ('m_y', float)]
+        # ... specifies the dtype for both m_x/y
+
+        np_results: NdArray = np.array(
+            [
+            (result_.m_x, result_.m_y) 
+            for result_ in results
+            ], dtype=dtype )
+
+        return cls(n, np_results)
+
+    def get_results_lists(self) -> tuple[np.ndarray, np.ndarray]:
+        return self.results['m_x'], self.results['m_y']
+
+
+def generate_test_data(num_results: int) -> list[UltimateBendingResults]:
+    return [UltimateBendingResults(
+        m_x=random.rand(), m_y=random.rand()
+        # examined operation is extracting the m_x/y from the results, only this is needed
+        ) for _ in range(num_results)]
+
+# --- ⏲️
+def time_execution(func, *args, **kwargs) -> tuple:
+    start_time = time.time()
+    result = func(*args, **kwargs)
+    end_time = time.time()
+
+    return (result, end_time - start_time)
+
+# --- 🧪
+@pytest.mark.parametrize("num_results", [10, 100, 1_000, 10_000, 100_000, 1_000_000])
+def test_get_results_lists(num_results) -> None:
+    random.seed(42)  # seed set for reproducible test outcome
+
+    test_data: list[UltimateBendingResults] = generate_test_data(num_results)
+
+    # --- Extract m_x/y from the same random test_data
+    original_results = BiaxialBendingResults(n=1_000, results=test_data) # same n_value for both, unlike inrweaxrion
+    new_results = NewBiaxialBendingResults.from_list(n=1_000, results=test_data)
+
+    # --- Time the execution of both operations
+    (original_m_x, original_m_y), original_time = time_execution(original_results.get_results_lists) # Original
+    (new_m_x, new_m_y), new_time = time_execution(new_results.get_results_lists)  # New cl
+
+    # --- Check correctness between operations
+    assert len(original_m_x) == len(new_m_x) == num_results  # check no. of results
+    assert len(original_m_y) == len(new_m_y) == num_results
+
+    testing.assert_allclose(original_m_x, new_m_x) # check same values of m_x/y 
+    testing.assert_allclose(original_m_y, new_m_y)
+
+    # --- Calc. speedup 
+    speedup = original_time / new_time if new_time > 0 else float('inf')
+    assert new_time < original_time # 
+
+    # --- 📝 Compare
+    print(f"\nFor {num_results:,} results:")
+
+    # print(f"🐌 Original method: {original_time:.6f} seconds")
+    # print(f"🐎 New method: {new_time:.6f} seconds")
+    print(f"🐌 Original method: {original_time:.3e} seconds]")
+    print(f"🐎 New method: {new_time:.3e} secons]")
+    print(f"⚡ SPEEDUP: {speedup:.2f}x")
+
+if __name__ == "__main__":
+    pytest.main([__file__])
+    test_get_results_lists(num_results=555_555)