S7 Object-Oriented Programming System for R

Data stored within an S7 object is referred to as properties. You use the @ operator to both retrieve and assign property values.

Properties are automatically validated against the types declared in new_class() and against the class's validator function upon assignment.

x@start          # Get property
x@end <- 20     # Set property

S7 is an experimental OOP system designed as a successor to S3 and S4. You can install it from CRAN using install.packages().

install.packages("S7")

When working with the fr package (which utilizes S7 for object validation), you may encounter validation errors stating that an fr_schema object is invalid. The specific error message is:

Error: <fr::fr_schema> object is invalid: - all items in @fields should be fr_field objects

This indicates that the fields component of an fr_schema object contains elements that are not valid fr_field objects. This error commonly surfaces during:

• Running examples (e.g., using as_data_frame() on an fr_tdr object).
• Running tests (e.g., testthat failures).
• Building vignettes (e.g., creating_a_tabular-data-resource.Rmd or read_fr_tdr.Rmd).
• Executing code in vignettes using as_fr_tdr() or read_fr_tdr().

S7 uses a functional OOP model where methods are associated with generic functions.

1. Create a generic: Use new_generic() to define a new generic function. Use the := operator to name it. If the generic requires arguments that are not used for dispatch, provide a body via the fun argument.
2. Define a method: Use the method<- assignment operator to register a specific implementation for a class.

Methods can also be registered for base types on S7 generics, or for S7 classes on existing S3/S4 generics.

# 1. Create the generic
inside := new_generic("x")

# 2. Define a method for a specific class
method(inside, Range) <- function(x, y) {
  y >= x@start & y <= x@end
}

# 3. Call the generic
inside(x, c(0, 5, 10, 15))

S7 classes are formally defined using new_class(). A class definition includes a list of properties (specifying types) and an optional validator function to ensure object integrity.

To name the class, use the := operator during assignment. The object returned by new_class() serves as the constructor for creating instances of that class.

Range := new_class(
  properties = list(
    start = class_double,
    end = class_double
  ),
  validator = function(self) {
    if (length(self@start) != 1) {
      "@start must be length 1"
    } else if (length(self@end) != 1) {
      "@end must be length 1"
    } else if (self@end < self@start) {
      "@end must be greater than or equal to @start"
    }
  }
)

# Create an instance
x <- Range(start = 1, end = 10)