add is_quitte()

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

is_quitte() tests the minimal quitte "standard", that is:

• data.frame class and
• existence and class of model, scenario, region, variable, unit, period, and value columns.

[orichters]

• What is the difference between is_quitte and the old is.quitte? https://github.com/pik-piam/quitte/blob/master/R/is.quitte.R
• I find that a bit confusing…

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

What is the difference between is_quitte and the old is.quitte? https://github.com/pik-piam/quitte/blob/master/R/is.quitte.R

is.quitte() requires a 'quitte' class attribute, and a bunch of columns to be of type factor, both of which are irrelevant for (most) quitte functions.

[fbenke-pik]

So i assume both helpers are needed in different contexts and should exist alongside?
If not: I'd expect this to replace is.quitte. If yes: You should reference the other function and briefly point out differences between the two in the method signatures of both functions, so that a user of library can easily tell whether to use is.quitte or is_quitte.

[orichters]

What is the difference between is_quitte and the old is.quitte? https://github.com/pik-piam/quitte/blob/master/R/is.quitte.R

is.quitte() requires a 'quitte' class attribute, and a bunch of columns to be of type factor, both of which are irrelevant for (most) quitte functions.

In piamInterfaces, we certainly have lots of of functions using levels that expect them to come as factors, but that is ok as the data goes through as.quitte before. I don't fully understand why a second "is quitte" function is useful and would probably have preferred something like is.quitte(strict = FALSE) or so to avoid having these 2 functions with basically the same name, but different behavior.

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

In piamInterfaces, we certainly have lots of of functions using levels that expect them to come as factors

I was referring to functions within the quitte package. Other packages are free to do whatever they want. If you are bored some day and want to make that code more robust, have a look into quitte::unique_or_levels().

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

In the end I did not need that function now, so I will close this merge request. I might open it again should I actually need it.

[orichters]

I was referring to functions within the quitte package. Other packages are free to do whatever they want. If you are bored some day and want to make that code more robust, have a look into quitte::unique_or_levels().

Good to know. But as.quitte guarantees factor columns, I think the code is as robust as needed ;)