⚡️ Speed up method Cube.add_aux_factory by 5,758%

[codeflash-ai]

📄 5,758% (57.58x) speedup for Cube.add_aux_factory in lib/iris/cube.py

⏱️ Runtime : 43.1 milliseconds → 735 microseconds (best of 12 runs)

📝 Explanation and details

The key optimization targets the performance bottleneck in the add_aux_factory method. The original code had an expensive O(N) membership check ref_coord not in cube_coords inside a loop, creating O(N*M) time complexity where N is the number of coordinates and M is the number of factory dependencies.

What was optimized:

1. Replaced list-based membership checks with set-based identity checks: Instead of using ref_coord not in cube_coords (which requires linear scanning), the code now creates a set of coordinate IDs using cube_coords_id_set = {id(coord) for coord in cube_coords} and checks id(ref_coord) not in cube_coords_id_set.

2. Eliminated the nested helper function: The original coordsonly helper function created temporary lists and was called twice. This is replaced with direct list building in a single pass.

Why this leads to speedup:

• Set membership testing is O(1) average case vs O(N) for list membership
• Using id() for object identity comparison is faster than Python's default equality checks
• Single-pass coordinate collection eliminates redundant iterations
• Reduces memory allocations from temporary list comprehensions

Performance characteristics:
The line profiler shows the critical line if ref_coord is not None and ref_coord not in cube_coords: took 98.5% of execution time in the original (167ms), but the optimized version with id(ref_coord) not in cube_coords_id_set takes only 8.7% (254μs) - a ~650x improvement on this specific operation. The overall method speedup of 5758% demonstrates how a single algorithmic improvement can dramatically impact performance when the bottleneck is properly identified.

This optimization is particularly effective for test cases with many coordinates or multiple factory dependencies, where the quadratic behavior of the original code becomes pronounced.

✅ Correctness verification report:

Test	Status
⚙️ Existing Unit Tests	✅ 1205 Passed
🌀 Generated Regression Tests	🔘 None Found
⏪ Replay Tests	✅ 150 Passed
🔎 Concolic Coverage Tests	🔘 None Found
📊 Tests Coverage	90.0%

⚙️ Existing Unit Tests and Runtime

Test File::Test Function	Original ⏱️	Optimized ⏱️	Speedup
integration/test_pp.py::TestVertical.test_hybrid_height_with_non_standard_coords	68.9μs	5.46μs	1162%✅
integration/test_pp.py::TestVertical.test_hybrid_pressure_with_non_standard_coords	68.7μs	5.40μs	1173%✅
unit/cube/test_Cube.py::Test_add_metadata.test_add_valid_aux_factory	40.5μs	4.99μs	712%✅
unit/cube/test_Cube.py::Test_add_metadata.test_error_for_add_invalid_aux_factory	48.6μs	11.4μs	326%✅

⏪ Replay Tests and Runtime

Test File::Test Function	Original ⏱️	Optimized ⏱️	Speedup
test_pytest_libiristestsintegrationtest_netcdf__loadsaveattrs_py_libiristestsunitlazy_datatest_non_lazy_p__replay_test_0.py::test_iris_cube_Cube_add_aux_factory	42.8ms	707μs	5952%✅

To edit these changes git checkout codeflash/optimize-Cube.add_aux_factory-mh52lg9j and push.