Class ZFastTrie<T>
- All Implemented Interfaces:
ObjectBidirectionalIterable<T>
,ObjectCollection<T>
,ObjectIterable<T>
,ObjectSet<T>
,ObjectSortedSet<T>
,Serializable
,Cloneable
,Iterable<T>
,Collection<T>
,Set<T>
,SortedSet<T>
In rough terms, the z-fast trie uses time |x|/w (which is optimal) to actually look at the string content,
and log(max{|x|, |x-|, |x+|}) to perform the search. This is known to be (essentially) optimal.
String lengths are up to Integer.MAX_VALUE
, and not limited to be a constant multiple of w for the bounds to hold.
The linear overhead of a z-fast trie is very low. For n keys we allocate 2n − 1 nodes containing six references and two longs, plus a dictionary containing n − 1 nodes (thus using around 2n references and 2n longs).
- See Also:
-
Nested Class Summary
Modifier and TypeClassDescriptionprotected static final class
protected static final class
A linear-probing hash map that compares keys using signatures as a first try.protected static final class
A internal node.protected static final class
An external node, a.k.a. leaf.protected static class
A node of the trie.protected static final class
-
Field Summary
Modifier and TypeFieldDescriptionA dictionary mapping handles to the corresponding internal nodes.static final long
-
Constructor Summary
ConstructorDescriptionZFastTrie
(TransformationStrategy<? super T> transform) Creates a new z-fast trie using the given transformation strategy.ZFastTrie
(Iterable<? extends T> elements, TransformationStrategy<? super T> transform) Creates a new z-fast trie using the given elements and transformation strategy.ZFastTrie
(Iterator<? extends T> elements, TransformationStrategy<? super T> transform) Creates a new z-fast trie using the given elements and transformation strategy. -
Method Summary
Modifier and TypeMethodDescriptionboolean
Returns the first element in the trie that is greater than or equal to the provided bound.static final long
checkMask
(long b) Returns the mask used for check for 2-fattest numbers when the left extreme of the interval is −1.static final long
checkMask
(long a, long b) Returns the mask used for check for 2-fattest numbers when the left extreme of the interval is not -1.Comparator<? super T>
protected void
completeFatBinarySearchStack
(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long a, long b) Completes the stack of a previous successful fat binary search.boolean
protected ZFastTrie.InternalNode<T>
fatBinarySearch
(LongArrayBitVector v, long[] state, long b) Performs a non-exact fat binary search.protected ZFastTrie.InternalNode<T>
fatBinarySearchExact
(LongArrayBitVector v, long[] state, long b) Performs an exact fat binary search.protected void
fatBinarySearchStack
(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long b) Performs a non-exact fat binary search with stack.protected void
fatBinarySearchStackExact
(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long b) Performs an exact fat binary search with stack.first()
Returns the first element in the trie that is smaller than or equal to the provided bound.void
getGrandParentExitNode
(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack) Returns the grandparent of the exit node of a given bit vector.getParentExitNode
(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack) Returns the parent of the exit node of a given bit vector.Returns the first element in the trie that is greater than the provided bound.boolean
isNonempty
(T lowerBound, T upperBound) Returns whether there is an element between the given bounds.iterator()
last()
Returns the first element in the trie that is smaller than the provided bound.static void
predecessor
(T upperBound) Returns the first element in the trie that is smaller than the provided bound.boolean
int
size()
strictSuccessor
(T lowerBound) Returns the first element in the trie that is greater than the provided bound.Returns the first element in the trie that is greater than or equal to the provided bound.static final long
twoFattest
(long a, long b) Returns the 2-fattest number in an interval.weakPredecessor
(T upperBound) Returns the first element in the trie that is smaller than or equal to the provided bound.Methods inherited from class it.unimi.dsi.fastutil.objects.AbstractObjectSet
equals, hashCode
Methods inherited from class it.unimi.dsi.fastutil.objects.AbstractObjectCollection
toString
Methods inherited from class java.util.AbstractCollection
addAll, clear, containsAll, isEmpty, removeAll, retainAll, toArray, toArray
Methods inherited from class java.lang.Object
clone, finalize, getClass, notify, notifyAll, wait, wait, wait
Methods inherited from interface java.util.Collection
parallelStream, removeIf, stream, toArray
Methods inherited from interface it.unimi.dsi.fastutil.objects.ObjectSortedSet
spliterator
-
Field Details
-
serialVersionUID
public static final long serialVersionUID- See Also:
-
handle2Node
A dictionary mapping handles to the corresponding internal nodes.
-
-
Constructor Details
-
ZFastTrie
Creates a new z-fast trie using the given transformation strategy.- Parameters:
transform
- a transformation strategy that must turn distinct elements into distinct, prefix-free bit vectors.
-
ZFastTrie
Creates a new z-fast trie using the given elements and transformation strategy.- Parameters:
elements
- an iterator returning the elements to be inserted in the trie.transform
- a transformation strategy that must turn distinct elements into distinct, prefix-free bit vectors.
-
ZFastTrie
Creates a new z-fast trie using the given elements and transformation strategy.- Parameters:
elements
- an iterator returning the elements to be inserted in the trie.transform
- a transformation strategy that must turn distinct elements into distinct, prefix-free bit vectors.
-
-
Method Details
-
size
public int size()- Specified by:
size
in interfaceCollection<T>
- Specified by:
size
in interfaceSet<T>
- Specified by:
size
in classAbstractCollection<T>
-
twoFattest
public static final long twoFattest(long a, long b) Returns the 2-fattest number in an interval.Note that to get the length of the handle of a node you must call this function passing the length of the extent of the parent (one less than the node name) and the length of the extent of the node.
- Parameters:
a
- left extreme, ≥-1 (excluded).b
- right extreme, ≥ 0 (included).- Returns:
- the 2-fattest number in (
a
..b
].
-
checkMask
public static final long checkMask(long a, long b) Returns the mask used for check for 2-fattest numbers when the left extreme of the interval is not -1.Note that to get the length of the handle of a node you must call this function passing the length of the extent of the parent (one less than the node name) and the length of the extent of the node.
- Parameters:
a
- left extreme, ≥-1 (excluded).b
- right extreme, ≥ 0 (included).- Returns:
- −1 ≪ λ(
a
⊕b
), the initial mask for fat binary search in(a..b]
.
-
checkMask
public static final long checkMask(long b) Returns the mask used for check for 2-fattest numbers when the left extreme of the interval is −1.- Parameters:
b
- right extreme, ≥ 0 (included).- Returns:
- −1 ≪ λ
b
+ 1, the initial mask for fat binary search in(-1..b]
.
-
add
- Specified by:
add
in interfaceCollection<T>
- Specified by:
add
in interfaceSet<T>
- Overrides:
add
in classAbstractCollection<T>
-
remove
- Specified by:
remove
in interfaceCollection<T>
- Specified by:
remove
in interfaceSet<T>
- Overrides:
remove
in classAbstractCollection<T>
-
getParentExitNode
public ZFastTrie.ParexData<T> getParentExitNode(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack) Returns the parent of the exit node of a given bit vector.- Parameters:
v
- a bit vector.state
- the hash state ofv
precomputed byHashes.preprocessMurmur(BitVector, long)
.stack
- a stack that will be filled with the 2-fat ancestors.- Returns:
- the parent of the exit node of
v
, ornull
if the exit node is the root.
-
getGrandParentExitNode
public void getGrandParentExitNode(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack) Returns the grandparent of the exit node of a given bit vector.- Parameters:
v
- a bit vector.state
- the hash state ofv
precomputed byHashes.preprocessMurmur(BitVector, long)
.stack
- a nonempty stack as filled bygetParentExitNode(LongArrayBitVector, long[], ObjectArrayList)
; the top of the stack must not be the root.
-
fatBinarySearchStack
protected void fatBinarySearchStack(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long b) Performs a non-exact fat binary search with stack.- Parameters:
v
- the bit vector on which to perform the search.state
- preprocessed MurmurHash state forv
.stack
- a stack where the results of the search will be cumulated.b
- the right extreme of the search interval, ≥ −1 (included).
-
fatBinarySearchStackExact
protected void fatBinarySearchStackExact(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long b) Performs an exact fat binary search with stack.- Parameters:
v
- the bit vector on which to perform the search.state
- preprocessed MurmurHash state forv
.stack
- a stack where the results of the search will be cumulated.b
- the right extreme of the search interval, ≥ −1 (included).
-
completeFatBinarySearchStack
protected void completeFatBinarySearchStack(LongArrayBitVector v, long[] state, ObjectArrayList<ZFastTrie.InternalNode<T>> stack, long a, long b) Completes the stack of a previous successful fat binary search.- Parameters:
v
- the bit vector on which to perform the search.state
- preprocessed MurmurHash state forv
.stack
- a stack where the results of the completion will be cumulated.a
- the left extreme of the completion interval, ≥ −1 (excluded)b
- the right extreme of the completion interval, ≥a
(included).
-
fatBinarySearch
Performs a non-exact fat binary search.- Parameters:
v
- the bit vector on which to perform the search.state
- preprocessed MurmurHash state forv
.b
- the right extreme of the search interval, ≥ −1 (included).- Returns:
- the parent of the exit node or the exit node, in case of success; an arbitrary node otherwise.
-
fatBinarySearchExact
protected ZFastTrie.InternalNode<T> fatBinarySearchExact(LongArrayBitVector v, long[] state, long b) Performs an exact fat binary search.- Parameters:
v
- the bit vector on which to perform the search.state
- preprocessed MurmurHash state forv
.b
- the right extreme of the search interval, ≥ −1 (included).- Returns:
- the parent of the exit node.
-
contains
- Specified by:
contains
in interfaceCollection<T>
- Specified by:
contains
in interfaceSet<T>
- Overrides:
contains
in classAbstractCollection<T>
-
successor
Returns the first element in the trie that is greater than or equal to the provided bound.- Parameters:
lowerBound
- a lower bound on the returned value.- Returns:
- the first element in the trie that is greater than or equal to
lowerBound
, ornull
if no such element exists.
-
ceiling
Returns the first element in the trie that is greater than or equal to the provided bound.- Parameters:
lowerBound
- a lower bound on the returned value.- Returns:
- the first element in the trie that is greater than or equal to
lowerBound
, ornull
if no such element exists. - Implementation Specification:
- This method just delegates to
successor(Object)
.
-
strictSuccessor
Returns the first element in the trie that is greater than the provided bound.- Parameters:
lowerBound
- a strict lower bound on the returned value.- Returns:
- the first element in the trie that is greater than
lowerBound
, ortail
if no such element exists.
-
higher
Returns the first element in the trie that is greater than the provided bound.- Parameters:
lowerBound
- a strict lower bound on the returned value.- Returns:
- the first element in the trie that is greater than
lowerBound
, ortail
if no such element exists. - Implementation Specification:
- This method just delegates to
strictSuccessor(Object)
.
-
predecessor
Returns the first element in the trie that is smaller than the provided bound.- Parameters:
upperBound
- a strict upper bound on the returned value.- Returns:
- the first element in the trie that is smaller than
upperBound
, orhead
if no such element exists.
-
lower
Returns the first element in the trie that is smaller than the provided bound.- Parameters:
upperBound
- a strict upper bound on the returned value.- Returns:
- the first element in the trie that is smaller than
upperBound
, orhead
if no such element exists. - Implementation Specification:
- This method just delegates to
predecessor(Object)
.
-
weakPredecessor
Returns the first element in the trie that is smaller than or equal to the provided bound.- Parameters:
upperBound
- an upper bound on the returned value.- Returns:
- the first element in the trie that is smaller than or equal to
upperBound
, orhead
if no such element exists.
-
floor
Returns the first element in the trie that is smaller than or equal to the provided bound.- Parameters:
upperBound
- an upper bound on the returned value.- Returns:
- the first element in the trie that is smaller than or equal to
upperBound
, orhead
if no such element exists. - Implementation Specification:
- This method just delegates to
weakPredecessor(Object)
.
-
isNonempty
Returns whether there is an element between the given bounds.- Parameters:
lowerBound
- a lower bound.upperBound
- an upper bound.- Returns:
- true if there is an element in the interval
[lowerBound...upperBound)
.
-
iterator
- Specified by:
iterator
in interfaceCollection<T>
- Specified by:
iterator
in interfaceIterable<T>
- Specified by:
iterator
in interfaceObjectBidirectionalIterable<T>
- Specified by:
iterator
in interfaceObjectCollection<T>
- Specified by:
iterator
in interfaceObjectIterable<T>
- Specified by:
iterator
in interfaceObjectSet<T>
- Specified by:
iterator
in interfaceObjectSortedSet<T>
- Specified by:
iterator
in interfaceSet<T>
- Specified by:
iterator
in classAbstractObjectSortedSet<T>
-
iterator
- Specified by:
iterator
in interfaceObjectSortedSet<T>
-
comparator
- Specified by:
comparator
in interfaceSortedSet<T>
-
first
-
last
-
headSet
-
subSet
-
tailSet
-
main
-