Collections utilities including combining lists and weighted tries.
A few convenience collection types that I was using in my RPTools project, but decided to surface separately.
These collection types can combine elements as they are added instead of replacing or duplicating them. The elements that are being combined must implement the interface Combine, so that the collections know what to do when it finds it must combine elements.
CombiningList surfaces a new add method add(E element, Predicate matcher) that allows you to provide a predicate to match against elements already in the list. You must supply a backing list to the constructor that it will forward method calls to accordingly. The method returns the number of elements of the list that were modified, or 1 if the element was added as normal (no elements matched the Predicate).
CombiningMaps must use a value type that implements Combine. When calling CombiningMap.put(k, v), if the key k already has a value associated with it, the existing value is combined with the new value instead of replacing it.
Weighted collections allow you to specify weights along with the elements. The weights are then used to retrieve the top X elements in the collection. The collections also support getting a pseudo-random element, where the randomness is determined by the elements' relative weights.
In addition to the above, WeightedList also provides the method findFirst(Predicate matcher), which is a quicker way of writing list.stream().filter(matcher).findFirst().
WeightedTrie is a Trie implementation which allows you to quickly add chains of elements down the trie, combining equal elements and their weights as it does so. For example:
trie.addChain(1.0, new String[] { "A", "B", "C" });
trie.addChain(2.0, new String[] { "A", "D", "C" });
trie.addChain(3.0, new String[] { "A", "D" });
In this example, the top level of the Trie will contain a single element, "A", which has a weight of 6 (1 + 2 + 3). "A" will have two children, "B" with a weight of 1, and "D" with a weight of 5 (2 + 3). "B"s single child "C" will also have a weight of 1, while "D"s single child "C" will have a weight of 2.