Make as card check format

[kieranjmartin]

Resolves #514

Incorportating check_ard_structure into as_card proved to not be straight forwards as as_card is used throughout the code base to coerce objects into a card class that do not pass the checks in check_ard_structure.

For now at least I did the following

1. check_ard_structure can either error or warn based on user input. I made a function to switch between these (I couldn't see any function which existed to do this already). Part of me wonders about allowing the user a bit more control over what it considers an error or a warning; they can switch off caring about the column order, and whether there is a methods entry, but maybe we could allow more in future. That said I think it's probably better as a binary switch for a change
2. In as_card, there is now an argument for whether you want to check if a card follows this structure. At the moment this defaults to TRUE, but if you think this is too strong I could switch it to default FALSE. Doing that would be the most low impact change.

[github-actions]

✅ All contributors have signed the CLA
_{Posted by the CLA Assistant Lite bot.}

[kieranjmartin]

I have read the CLA Document and I hereby sign the CLA

[bzkrouse]

looks great to me!

[ddsjoberg]

Thanks @kieranjmartin ! Can you add a news entry and we'll merge?

[kieranjmartin]

Thanks @kieranjmartin ! Can you add a news entry and we'll merge?

Done :)

[ddsjoberg]

@kieranjmartin thank you for this PR! It's going to be super useful!

Before we merge, can you assess if we need to make updates in cardx as well, e.g. those as_card(check = FALSE) changes? If so, can we make those changes before we merge? Thank you!

[kieranjmartin]

@ddsjoberg I was thinking about this a little; I had a look at reverse dependencies and I can see that both gtsummary and cardx are using as_card; the other packages which takes cards as a dependency don't appear to use it.

I can go through both and add a PR to update this, but I am wondering if maybe I should make check=FALSE the default instead? To an extent I think the way it is currently coded is more principled, but I'm aware that this is a backwards incompatible change; if anyone is using this in a non package context they may get stung by the change too. So putting FALSE as the default would be a friendlier way to approach this.

[ddsjoberg]

Thanks @kieranjmartin !

I agree and like that the default is check = TRUE. For most users, I think this is what they will want when they call the function. There was recently a user who reported an issue in their code, where a check = TRUE would have caught it.

If you can check cardx, I can make the updates in gtsummary?