Struct Top 

pub struct Top<A, C: New> { /* private fields */ }

This probabilistic data structure tracks the n top keys given a stream of (key,value) tuples, ordered by the sum of the values for each key (the “aggregated value”). It uses only O(n) space.

Its implementation is two parts:

• a doubly linked hashmap, mapping the top n keys to their aggregated values, and ordered by their aggregated values. This is used to keep a more precise track of the aggregated value of the top n keys, and reduce collisions in the count-min sketch.
• a count-min sketch to track all of the keys outside the top n. This data structure is also known as a counting Bloom filter. It uses conservative updating for increased accuracy.

The algorithm is as follows:

while a key and value from the input stream arrive:
    if H[key] exists
        increment aggregated value associated with H[key]
    elsif number of items in H < k
        put H[key] into map with its associated value
    else
        add C[key] into the count-min sketch with its associated value
        if aggregated value associated with C[key] is > the lowest aggregated value in H
            move the lowest key and value from H into C
            move C[key] and value from C into H
endwhile

See An Improved Data Stream Summary: The Count-Min Sketch and its Applications and New Directions in Traffic Measurement and Accounting for background on the count-min sketch with conservative updating.

Implementations

impl<A: Hash + Eq + Clone, C: Ord + New + for<'a> UnionAssign<&'a C> + Intersect> Top<A, C>

pub fn new( n: usize, probability: f64, tolerance: f64, config: <C as New>::Config, ) -> Self

Create an empty Top data structure with the specified n capacity.

pub fn capacity(&self) -> usize

The n most frequent elements we have capacity to track.

pub fn push<V: ?Sized>(&mut self, item: A, value: &V)

where C: for<'a> AddAssign<&'a V> + IntersectPlusUnionIsPlus,

“Visit” an element.

pub fn clear(&mut self)

Clears the Top data structure, as if it was new.

pub fn iter(&self) -> TopIter<'_, A, C>

An iterator visiting all elements and their counts in descending order of frequency. The iterator element type is (&’a A, usize).

Trait Implementations

impl<A: Hash + Eq + Clone, C: Ord + New + Clone + for<'a> AddAssign<&'a C> + for<'a> UnionAssign<&'a C> + Intersect + IntersectPlusUnionIsPlus> Add for Top<A, C>

type Output = Top<A, C>

The resulting type after applying the + operator.

fn add(self, other: Self) -> Self

Performs the + operation.

impl<A: Hash + Eq + Clone, C: Ord + New + Clone + for<'a> AddAssign<&'a C> + for<'a> UnionAssign<&'a C> + Intersect + IntersectPlusUnionIsPlus> AddAssign for Top<A, C>

fn add_assign(&mut self, other: Self)

Performs the += operation.

impl<A: Clone, C: Clone + New> Clone for Top<A, C>

fn clone(&self) -> Top<A, C>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<A: Hash + Eq + Clone + Debug, C: Ord + New + Clone + for<'a> UnionAssign<&'a C> + Intersect + Debug> Debug for Top<A, C>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de, A, C> Deserialize<'de> for Top<A, C>

where A: Hash + Eq + Deserialize<'de>, C: Deserialize<'de> + New, <C as New>::Config: Deserialize<'de>,

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<A, C> Serialize for Top<A, C>

where A: Hash + Eq + Serialize, C: Serialize + New, <C as New>::Config: Serialize,

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl<A: Hash + Eq + Clone, C: Ord + New + Clone + for<'a> AddAssign<&'a C> + for<'a> UnionAssign<&'a C> + Intersect + IntersectPlusUnionIsPlus> Sum<Top<A, C>> for Option<Top<A, C>>

fn sum<I>(iter: I) -> Self

where I: Iterator<Item = Top<A, C>>,

Takes an iterator and generates Self from the elements by “summing up” the items.