refactor!: temporary deprecate positional arguments of pl$Series()

[eitsupi]

Part of #903

As of version 0.16, positional arguments are deprecated and a warning is occurred.

[etiennebacher]

I'd like to clarify a bit when this warning appears and when it doesn't before reviewing the rest

[etiennebacher]

Thanks, some comments. I'm also wondering if we should made the change in argument position in 0.18.0 since we mention in #903 that

This would be a very huge breaking change, so we will probably have to continue warnings for several months and then make the change.

Given the rapid pace of development, I think 0.18.0 would be released in a couple of months. What I have in mind is that it's possible rust-polars 0.39.0 (and therefore r-polars 0.16.0) are released in a week or so, and 0.17.0 around the end of April. What do you think about keeping this deprecation warning for 2 full releases?

[eitsupi]

Thanks for your review.

What do you think about keeping this deprecation warning for 2 full releases?

This is a difficult question.

My concern is that the mechanism to fall back to the positional arguments is complicated, and I think it would be less confusing to remove this logic quickly.
That is, ideally pl$Series(name = "foo", "bar") should work fine in 0.15 and future versions, but this version will warning for this.
(Since there are originally three positional arguments, it is quite cumbersome to absorb all the patterns of positional and named arguments.)

So I think it would be better to replace the positional argument in 0.17 instead of 0.18 and remove this logic early to avoid the warning.

Also, as like I have substituted tests here, it should be easy to fix even with this breaking change, since in most cases it should be possible to fix it by simply substituting pl$Series( for as_polars_series(.

[etiennebacher]

What I'm worried about is that we have no control over releases in rust-polars, and the time span between 0.39.0 and 0.40.0 will correspond to the time of our deprecation warning.

What I propose is to keep what we have for now, i.e make the switch in arguments in 0.17.0, but keep the possibility of extending the warning by one release if we consider that it was too short.

[eitsupi]

What I propose is to keep what we have for now, i.e make the switch in arguments in 0.17.0, but keep the possibility of extending the warning by one release if we consider that it was too short.

I think it is worth doing so too.