Correctly parse comparison constraints with multiple parts

[TimScheckenbach]

Changed the Comparison Class, such that chsaining of comparisons is now possible, and can be evaluated.
Fix for #721

[github-actions]

Bencher Report

Branch	fix/multiple-comparison
Testbed	ubuntu-latest

Click to view all benchmark results

Benchmark	Latency	Benchmark Result milliseconds (ms) (Result Δ%)	Upper Boundary milliseconds (ms) (Limit %)
tests/test_benchmarks.py::test_generate_with_single_hard_constraint	📈 view plot 🚷 view threshold	400.00 ms (+66.14%) Baseline: 240.75 ms	466.14 ms (85.81%)
tests/test_benchmarks.py::test_generate_with_single_soft_constraint	📈 view plot 🚷 view threshold	10,751.45 ms (-27.88%) Baseline: 14,908.17 ms	26,417.96 ms (40.70%)
tests/test_benchmarks.py::test_init_fandango	📈 view plot 🚷 view threshold	150.64 ms (+0.30%) Baseline: 150.20 ms	156.20 ms (96.44%)
tests/test_benchmarks.py::test_parse_spec	📈 view plot 🚷 view threshold	149.98 ms (-0.39%) Baseline: 150.57 ms	155.05 ms (96.73%)

🐰 View full continuous benchmarking report in Bencher

[riesentoaster]

Is there a clean way to not have to evaluate the middle parts more than once?

If not: would it make sense to parse this into two separate constraints? Just to remove this additional logic from the constraint class?

[TimScheckenbach]

Is there a clean way to not have to evaluate the middle parts more than once?

If not: would it make sense to parse this into two separate constraints? Just to remove this additional logic from the constraint class?

I changed it, such that each comparator is evaluated only once.