diff --git a/moptipy/evaluation/selector.py b/moptipy/evaluation/selector.py index c3f3069e0..afab84ddd 100644 --- a/moptipy/evaluation/selector.py +++ b/moptipy/evaluation/selector.py @@ -9,6 +9,13 @@ offered by this module provides the functionality to make such a selection. It may be a bit slow, but hopefully it will pick the largest possible consistent sub-selection or, at least, get close to it. + +The current method to select the data is rather heuristic. It is based on +iteratively deleting those data elements whose configuration elements appear +the least often. This is computed in absolute terms. I actually think +normalizing this score should be better, but seemingly using the absolute raw +score creates larger consistent datasets. So for now, this is how we will do +it. """ from collections import Counter @@ -247,7 +254,7 @@ def select_consistent(data: Iterable[T], log: bool = True) -> list[T]: source.sort(key=lambda x: x[0]) count: int = list.__len__(source) if log: - logger(f"Found {source} records of data.") + logger(f"Found {count} records of data.") set_l: Final[int] = set.__len__({x[0] for x in source}) if set_l != count: diff --git a/moptipy/version.py b/moptipy/version.py index 2cf6864c3..c037144bb 100644 --- a/moptipy/version.py +++ b/moptipy/version.py @@ -2,4 +2,4 @@ from typing import Final #: the version string of `moptipy` -__version__: Final[str] = "0.9.125" +__version__: Final[str] = "0.9.126"