⚡️ Speed up function merge_detections by 10%

[codeflash-ai]

📄 10% (0.10x) speedup for merge_detections in inference/core/workflows/core_steps/fusion/detections_consensus/v1.py

⏱️ Runtime : 2.15 milliseconds → 1.96 milliseconds (best of 20 runs)

📝 Explanation and details

The optimized code achieves a 9% speedup through several key micro-optimizations that reduce attribute lookups, memory allocations, and unnecessary conversions:

What optimizations were applied:

1. Eliminated dictionary lookup redundancy: Cached detections.data as ddata and used dict.get() for scaling keys instead of checking membership with in followed by array access - this reduces multiple hash table lookups to single operations.

2. Reduced function call overhead: Pre-cached aggregator functions (class_selector, boxes_aggregator, masks_aggregator) instead of looking them up from dictionaries multiple times within conditionals.

3. Optimized mask array creation: Replaced np.array([aggregated_mask]) with np.expand_dims(aggregated_mask, axis=0) to avoid intermediate list allocation when creating the mask array.

4. Streamlined confidence aggregation: Moved confidence aggregation outside the return statement and stored result in a variable to avoid nested function calls during object construction.

5. Improved dtype handling in aggregate_field_values: Added fast-path check for arrays already in floating-point format to skip unnecessary astype(float) conversions, reducing array copying overhead.

Why this leads to speedup:

• Dictionary .get() is faster than in checks followed by key access
• Avoiding intermediate list allocations reduces memory pressure and GC overhead
• Pre-caching function references eliminates repeated dictionary lookups
• Dtype checks prevent redundant array conversions for already-float data

Test case performance:
These optimizations are particularly effective for the large detection scenarios (500+ detections) in the test suite, where the cumulative effect of reduced lookups and allocations becomes significant. The optimizations maintain identical functionality while reducing computational overhead in the hot path.

✅ Correctness verification report:

Test	Status
⚙️ Existing Unit Tests	✅ 73 Passed
🌀 Generated Regression Tests	🔘 None Found
⏪ Replay Tests	🔘 None Found
🔎 Concolic Coverage Tests	🔘 None Found
📊 Tests Coverage	100.0%

⚙️ Existing Unit Tests and Runtime

Test File::Test Function	Original ⏱️	Optimized ⏱️	Speedup
workflows/unit_tests/core_steps/fusion/test_detections_consensus.py::test_merge_detections_no_mask	175μs	178μs	-1.18%⚠️
workflows/unit_tests/core_steps/fusion/test_detections_consensus.py::test_merge_detections_with_masks	1.97ms	1.78ms	10.7%✅

To edit these changes git checkout codeflash/optimize-merge_detections-mhbuv5co and push.