Python Benchmarks¶
These benchmarks are not finished and are preliminary. Even after including many test cases I would still take any benchmark with a huge grain of salt!
The test names are organized as “test_TESTDATA_LIBRARY[_lowquality]”. TESTDATA refers to the test line set, the library is either simplifyline or simplification which is a similar rust library with the implemented RDP algorithm.
Our implementation follows the SimplifyJS library trick which can use a radial distance pre-filter when placed in lowquality mode (high_quality=False).
Note that the simplification library has way more features then this library and is probably a safer bet for use. I didn’t know about it until after I made this library. I also still desire a drop in C++ replacement with CMake library as well.
------------------------------------------------------------------------------------------------------ benchmark: 3 tests -----------------------------------------------------------------------------------------------------
Name (time in us) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_example1_simplifyline_lowquality 605.1000 (1.0) 806.2000 (1.0) 636.5691 (1.0) 16.2028 (1.0) 642.0000 (1.0) 21.8500 (1.0) 404;9 1,570.9213 (1.0) 1575 1
test_example1_simplifyline 3,734.1000 (6.17) 5,241.5000 (6.50) 3,835.5558 (6.03) 144.7665 (8.93) 3,805.7000 (5.93) 96.0000 (4.39) 5;4 260.7184 (0.17) 260 1
test_example1_simplification 9,910.6000 (16.38) 10,614.7000 (13.17) 10,140.0295 (15.93) 93.2924 (5.76) 10,119.3000 (15.76) 106.6250 (4.88) 18;4 98.6190 (0.06) 95 1
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------