The greykite package - as released on pypi pins all its dependencies to exact versions or narrow version ranges.
This is not a good idea, as this makes installation together with other packages very difficult, and impossibly for any other package with a comparably narrow, but different version range for one of the dependencies in question.
A number of these are, in addition, very outdated versions of packages, e.g., pandas 1.X (now on 2.2) or scikit-learn 1.3.1 (now on 1.6) - pandas may be sensible due to some breaking changes in v2, but there should not be a need ot pin scikit-learn==1.3.1.
I would suggest to adopt larger possible ranges for the dependencies.
The
greykitepackage - as released onpypipins all its dependencies to exact versions or narrow version ranges.This is not a good idea, as this makes installation together with other packages very difficult, and impossibly for any other package with a comparably narrow, but different version range for one of the dependencies in question.
A number of these are, in addition, very outdated versions of packages, e.g.,
pandas 1.X(now on 2.2) orscikit-learn 1.3.1(now on 1.6) -pandasmay be sensible due to some breaking changes in v2, but there should not be a need ot pinscikit-learn==1.3.1.I would suggest to adopt larger possible ranges for the dependencies.