Reducers

Initially introduced in Clojure, reducers are a model for processing collections which transforms the procedures rather than the collections themselves.

Its advantages include lazy evaluation, few allocations, efficiency, and the potential for convenient parallelization via the closely-related folders.

This framework is an implementation of reducers for Cocoa and Cocoa Touch.

What would I use it for?

Mapping, filtering, and folding Cocoa collections lazily, with low overhead.

Examples

First, import Reducers/Reducers.h:

#import <Reducers/Reducers.h>

Let’s say we have some numbers, but we only want to operate on the even ones:

id<REDReducible> evenNumbers = REDFilter(numbers, ^bool (NSNumber *each) {
    return each.integerValue % 2;
});

Now let’s make a text field for each:

id<REDReducible> views = REDMap(evenNumbers, ^(NSNumber *each) {
    NSTextField *textField = [NSTextField new];
    textField.editable = NO;
    textField.stringValue = each.description;
    return textField;
});

Reducers are evaluated lazily, so we’ve specified a way of turning numbers into text fields with even numbers, but we haven’t yet got a collection of the text fields. We can produce one by appending onto an array:

NSArray *viewsArray = [@[] red_append:views];

Finally, we can place these views into the view hierarchy:

self.view = [NSStackView stackViewWithViews:viewsArray];

Now let’s sum the numbers we produced:

NSNumber *sum = [evenNumbers red_reduce:@0 usingBlock:^(NSNumber *a, NSNumber *b) {
    return @(a.integerValue + b.integerValue);
}];

Notes

Since reducers are evaluated lazily, you need to reduce them to produce a Cocoa collection. -red_append:, which is implemented on NSArray, NSSet, and NSDictionary reduces its argument, and is a convenient way to produce a concrete, evaluated collection using a reducer.

API-level documentation is provided in the headers.