Enumerations

An enumeration is a datatype that can take a fixed, finite set of values. The Ridge model of scikit-learn for instance allows the user to choose a solver by selecting one of the strings "auto", "svd", "cholesky", "lsqr", "sparse_cg", "sag", "saga". However, the string datatype does not prevent the user from passing another invalid string like "automatic", leading to unspecified behavior. To prevent this, in Simple-ML the solver is an enumeration.

Declaring an Enumeration

The syntax to declare an enumeration is as follows:

• The keyword enum.
• The name of the enumeration.
• A list of the instances, i. e. the valid values of the enumeration enclosed in curly braces and separated by commas.

Coming back to the ridge solver example from the introduction, we would implement this in Simple-ML as follows, so that only the seven specified values are valid instances of the datatype.

enum RidgeSolver {
    AUTO,
    SVD,
    CHOLESKY,
    LSQR,
    SPARSE_CG,
    SAG,
    SAGA
}

Enum Variants

TODO

Constructors

TODO

Using an Enumeration as a Type

To express that the type of a declaration is an enumeration we simply write the name of the enumeration as the type. For example, when declaring a class called "Ridge" for the ridge regression model we can declare a parameter of type "RidgeSolver" like this:

class Ridge(solver: RidgeSolver) {
    // ...
}