Update minimal python version up to 3.10

[d-a-bunin]

Description

Currently (2.7.1) the library supports python versions: (3.8, 3.9, 3.10).

• We can't update to higher versions due to dependency conflict (pytorch_forecasting)
• We use typing module extensively, but on new versions of python we could use expressions like list[int] instead of typing.List[int]
• Old python versions could prohibit us from adding/updating some of the dependencies, because e.g. they could require python > 3.8.

Thoughts

I'm not sure if it is necessary to support the old python versions like 3.8 and 3.9 after the major update.

Related

• TODO: link discussion about pytorch_forecasting after creation

[xeonvs]

Latest versions(1.0+) of pytorch_forecasting now support python 3.1[12] and pandas <3.0

[d-a-bunin]

Currently we have a dependency pytorch_forecasting=^0.9.0" (which means >= 0.9.0 and <1). After updating to v1 some breaking changes in a dependencies were made (e.g. optuna changed from v2 to v3), so it is difficult for us to support both <1 and >=1 for pytorch_forecasting