[ENH] un-stale the package and give simple instructions for working install

[veberi-alteo]

Could somebody give a docker image, or detailes of package versions. I tried multiple installations and most of the time had errors when running the tutorials.

As the library becomes stale it is more and more difficult to make the proper environment.

Please share detailes, thanks.

[ari99]

I'm starting to work on getting it running, this allowed me to get the tests passing (see my fork for a code change as well)

pip install poetry
poetry install
pip install numba
pip install cpflows

[veberi-alteo]

Hi, thanks for the answer, do you use the latest available versions from the above list then?

[IlIlllIIllIIlll]

Try using an older version, build with these two files：

poetry.lock pyproject.toml

pip install poetry
poetry install --all-extras

and build whl package

python -m build

[fkiraly]

We will hopefully "un-stale" the library over the next weeks https://github.com/jdb78/pytorch-forecasting/issues/1592

Help with this would be greatly appreciated!

If you would like to, you can get in touch on the sktime discord: https://discord.com/invite/54ACzaFsn7 - mention this when you give your intro on the intro channel.

[fkiraly]

@ari99, the changes in your fork look great, thanks a lot for your effort! Would you be willing to make pull requests out of them - optimally, split up in small, mergeable bits? That is, multiple pull request, each of which fixes one thing at a time?

From a maintenance perspective, we need to adopt a piecewise approach, and we need to ensure no currently running code is going to break.

[fkiraly]

Also apologies, we (the sktime maintainers) had not spotted your fork until now - the usual way to make your contributions seen easily is via pull requests, that's why I'm asking. That also allows you to get the contribution recognition you deserve!