Updating black formatter to latest

[ryan-wolbeck]

Thank you for the PR, I think we will just need to run black locally and push and this should pass

[ryan-wolbeck]

It looks like our build on 3.6 is getting stuck resolving dependencies, can you modify the makefile https://github.com/stanfordmlgroup/ngboost/blob/master/Makefile to be 'pip install poetry==1.2.1 && poetry install' and push one more time?

[robertsont]

Seems poetry is not supported on versions of python older than 3.7

Might be worth considering abandoning 3.6. Also, building for 3.9 and 3.10 is something to think about as well.

[xmnlab]

I think that it makes total sense.

[ryan-wolbeck]

@robertsont yeah I think moving away from 3.6 makes sense especially because Python itself has dropped support. Can you update this pr to remove 3.6 and include 3.9/3.10? I think we will need to modify line 14 in .github/workflows/pythonpackage.yml and line 16 of pyproject.toml

Tagging @alejandroschuler just so you are aware of this change

[ryan-wolbeck]

@robertsont I think we will need to update the version on numpy for 3.10, while you are in there can you also update python to be >= 3.7

[robertsont]

Fixing the tests now. Also taking the suggestion from the scikit-learn maintainers to not use the load_boston() data set and instead use the fetch_california_housing() data set.

[ryan-wolbeck]

Yeah versions look good, thanks for updating that test set.

[ryan-wolbeck]

@robertsont I think we might have to bump the version of numpy up as well

[ryan-wolbeck]

Closed by #301